Add LLM review command for MusicXML correction

New `sheet-music-ocr review` command that sends MusicXML output to an
LLM (Claude or OpenAI, configurable via --provider flag) for reviewing
and fixing common OCR errors like incorrect pitches, rhythms, key
signatures, and garbled lyrics. Uses httpx for direct API calls.

https://claude.ai/code/session_017GqUbuRDT58toRaxMtfRmf

.gitignore

@@ -169,7 +169,3 @@ test.*
 # Frontend build output
 frontend/dist/
 frontend/node_modules/
-
-# data dir for training, validation, and testing
-data/
-config.toml

.vscode/settings.json

@@ -308,7 +308,6 @@
     "usernamehw",
     "userprefs",
     "vaninventory",
-    "vdev",
     "vfat",
     "victron",
     "virt",

common/optional/brain_substituter.nix

@@ -1,6 +0,0 @@
-{
-  nix.settings = {
-    trusted-substituters = [ "http://192.168.95.35:5000" ];
-    substituters = [ "http://192.168.95.35:5000/?priority=1&want-mass-query=true" ];
-  };
-}

flake.lock

@@ -8,11 +8,11 @@
       },
       "locked": {
         "dir": "pkgs/firefox-addons",
-        "lastModified": 1773979456,
+        "lastModified": 1772824881,
-        "narHash": "sha256-9kBMJ5IvxqNlkkj/swmE8uK1Sc7TL/LIRUI958m7uBM=",
+        "narHash": "sha256-NqX+JCA8hRV3GoYrsqnHB2IWKte1eQ8NK2WVbJkORcw=",
         "owner": "rycee",
         "repo": "nur-expressions",
-        "rev": "81e28f47ac18d9e89513929c77e711e657b64851",
+        "rev": "07e1616c9b13fe4794dad4bcc33cd7088c554465",
         "type": "gitlab"
       },
       "original": {
@@ -29,11 +29,11 @@
         ]
       },
       "locked": {
-        "lastModified": 1774007980,
+        "lastModified": 1772807318,
-        "narHash": "sha256-FOnZjElEI8pqqCvB6K/1JRHTE8o4rer8driivTpq2uo=",
+        "narHash": "sha256-Qjw6ILt8cb2HQQpCmWNLMZZ63wEo1KjTQt+1BcQBr7k=",
         "owner": "nix-community",
         "repo": "home-manager",
-        "rev": "9670de2921812bc4e0452f6e3efd8c859696c183",
+        "rev": "daa2c221320809f5514edde74d0ad0193ad54ed8",
         "type": "github"
       },
       "original": {
@@ -44,11 +44,11 @@
     },
     "nixos-hardware": {
       "locked": {
-        "lastModified": 1774018263,
+        "lastModified": 1771969195,
-        "narHash": "sha256-HHYEwK1A22aSaxv2ibhMMkKvrDGKGlA/qObG4smrSqc=",
+        "narHash": "sha256-qwcDBtrRvJbrrnv1lf/pREQi8t2hWZxVAyeMo7/E9sw=",
         "owner": "nixos",
         "repo": "nixos-hardware",
-        "rev": "2d4b4717b2534fad5c715968c1cece04a172b365",
+        "rev": "41c6b421bdc301b2624486e11905c9af7b8ec68e",
         "type": "github"
       },
       "original": {
@@ -60,11 +60,11 @@
     },
     "nixpkgs": {
       "locked": {
-        "lastModified": 1773821835,
+        "lastModified": 1772624091,
-        "narHash": "sha256-TJ3lSQtW0E2JrznGVm8hOQGVpXjJyXY2guAxku2O9A4=",
+        "narHash": "sha256-QKyJ0QGWBn6r0invrMAK8dmJoBYWoOWy7lN+UHzW1jc=",
         "owner": "nixos",
         "repo": "nixpkgs",
-        "rev": "b40629efe5d6ec48dd1efba650c797ddbd39ace0",
+        "rev": "80bdc1e5ce51f56b19791b52b2901187931f5353",
         "type": "github"
       },
       "original": {
@@ -76,11 +76,11 @@
     },
     "nixpkgs-master": {
       "locked": {
-        "lastModified": 1774051532,
+        "lastModified": 1772842888,
-        "narHash": "sha256-d3CGMweyYIcPuTj5BKq+1Lx4zwlgL31nVtN647tOZKo=",
+        "narHash": "sha256-bQRYIwRb9xuEMHTLd5EzjHhYMKzbUbIo7abFV84iUjM=",
         "owner": "nixos",
         "repo": "nixpkgs",
-        "rev": "8620c0b5cc8fbe76502442181be1d0514bc3a1b7",
+        "rev": "af5157af67f118e13172750f63012f199b61e3a1",
         "type": "github"
       },
       "original": {
@@ -125,11 +125,11 @@
         ]
       },
       "locked": {
-        "lastModified": 1773889674,
+        "lastModified": 1772495394,
-        "narHash": "sha256-+ycaiVAk3MEshJTg35cBTUa0MizGiS+bgpYw/f8ohkg=",
+        "narHash": "sha256-hmIvE/slLKEFKNEJz27IZ8BKlAaZDcjIHmkZ7GCEjfw=",
         "owner": "Mic92",
         "repo": "sops-nix",
-        "rev": "29b6519f3e0780452bca0ac0be4584f04ac16cc5",
+        "rev": "1d9b98a29a45abe9c4d3174bd36de9f28755e3ff",
         "type": "github"
       },
       "original": {

frontend/README.md

@@ -0,0 +1,73 @@
+# React + TypeScript + Vite
+
+This template provides a minimal setup to get React working in Vite with HMR and some ESLint rules.
+
+Currently, two official plugins are available:
+
+- [@vitejs/plugin-react](https://github.com/vitejs/vite-plugin-react/blob/main/packages/plugin-react) uses [Babel](https://babeljs.io/) (or [oxc](https://oxc.rs) when used in [rolldown-vite](https://vite.dev/guide/rolldown)) for Fast Refresh
+- [@vitejs/plugin-react-swc](https://github.com/vitejs/vite-plugin-react/blob/main/packages/plugin-react-swc) uses [SWC](https://swc.rs/) for Fast Refresh
+
+## React Compiler
+
+The React Compiler is not enabled on this template because of its impact on dev & build performances. To add it, see [this documentation](https://react.dev/learn/react-compiler/installation).
+
+## Expanding the ESLint configuration
+
+If you are developing a production application, we recommend updating the configuration to enable type-aware lint rules:
+
+```js
+export default defineConfig([
+  globalIgnores(['dist']),
+  {
+    files: ['**/*.{ts,tsx}'],
+    extends: [
+      // Other configs...
+
+      // Remove tseslint.configs.recommended and replace with this
+      tseslint.configs.recommendedTypeChecked,
+      // Alternatively, use this for stricter rules
+      tseslint.configs.strictTypeChecked,
+      // Optionally, add this for stylistic rules
+      tseslint.configs.stylisticTypeChecked,
+
+      // Other configs...
+    ],
+    languageOptions: {
+      parserOptions: {
+        project: ['./tsconfig.node.json', './tsconfig.app.json'],
+        tsconfigRootDir: import.meta.dirname,
+      },
+      // other options...
+    },
+  },
+])
+```
+
+You can also install [eslint-plugin-react-x](https://github.com/Rel1cx/eslint-react/tree/main/packages/plugins/eslint-plugin-react-x) and [eslint-plugin-react-dom](https://github.com/Rel1cx/eslint-react/tree/main/packages/plugins/eslint-plugin-react-dom) for React-specific lint rules:
+
+```js
+// eslint.config.js
+import reactX from 'eslint-plugin-react-x'
+import reactDom from 'eslint-plugin-react-dom'
+
+export default defineConfig([
+  globalIgnores(['dist']),
+  {
+    files: ['**/*.{ts,tsx}'],
+    extends: [
+      // Other configs...
+      // Enable lint rules for React
+      reactX.configs['recommended-typescript'],
+      // Enable lint rules for React DOM
+      reactDom.configs.recommended,
+    ],
+    languageOptions: {
+      parserOptions: {
+        project: ['./tsconfig.node.json', './tsconfig.app.json'],
+        tsconfigRootDir: import.meta.dirname,
+      },
+      // other options...
+    },
+  },
+])
+```

frontend/eslint.config.js

@@ -0,0 +1,23 @@
+import js from '@eslint/js'
+import globals from 'globals'
+import reactHooks from 'eslint-plugin-react-hooks'
+import reactRefresh from 'eslint-plugin-react-refresh'
+import tseslint from 'typescript-eslint'
+import { defineConfig, globalIgnores } from 'eslint/config'
+
+export default defineConfig([
+  globalIgnores(['dist']),
+  {
+    files: ['**/*.{ts,tsx}'],
+    extends: [
+      js.configs.recommended,
+      tseslint.configs.recommended,
+      reactHooks.configs.flat.recommended,
+      reactRefresh.configs.vite,
+    ],
+    languageOptions: {
+      ecmaVersion: 2020,
+      globals: globals.browser,
+    },
+  },
+])

frontend/index.html

@@ -0,0 +1,13 @@
+<!doctype html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <link rel="icon" type="image/svg+xml" href="/vite.svg" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>frontend</title>
+  </head>
+  <body>
+    <div id="root"></div>
+    <script type="module" src="/src/main.tsx"></script>
+  </body>
+</html>

frontend/package.json

@@ -0,0 +1,31 @@
+{
+  "name": "frontend",
+  "private": true,
+  "version": "0.0.0",
+  "type": "module",
+  "scripts": {
+    "dev": "vite",
+    "build": "tsc -b && vite build",
+    "lint": "eslint .",
+    "preview": "vite preview"
+  },
+  "dependencies": {
+    "react": "^19.2.0",
+    "react-dom": "^19.2.0",
+    "react-router-dom": "^7.12.0"
+  },
+  "devDependencies": {
+    "@eslint/js": "^9.39.1",
+    "@types/node": "^24.10.1",
+    "@types/react": "^19.2.5",
+    "@types/react-dom": "^19.2.3",
+    "@vitejs/plugin-react": "^5.1.1",
+    "eslint": "^9.39.1",
+    "eslint-plugin-react-hooks": "^7.0.1",
+    "eslint-plugin-react-refresh": "^0.4.24",
+    "globals": "^16.5.0",
+    "typescript": "~5.9.3",
+    "typescript-eslint": "^8.46.4",
+    "vite": "^7.2.4"
+  }
+}

frontend/public/vite.svg

@@ -0,0 +1 @@
+<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" class="iconify iconify--logos" width="31.88" height="32" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 257"><defs><linearGradient id="IconifyId1813088fe1fbc01fb466" x1="-.828%" x2="57.636%" y1="7.652%" y2="78.411%"><stop offset="0%" stop-color="#41D1FF"></stop><stop offset="100%" stop-color="#BD34FE"></stop></linearGradient><linearGradient id="IconifyId1813088fe1fbc01fb467" x1="43.376%" x2="50.316%" y1="2.242%" y2="89.03%"><stop offset="0%" stop-color="#FFEA83"></stop><stop offset="8.333%" stop-color="#FFDD35"></stop><stop offset="100%" stop-color="#FFA800"></stop></linearGradient></defs><path fill="url(#IconifyId1813088fe1fbc01fb466)" d="M255.153 37.938L134.897 252.976c-2.483 4.44-8.862 4.466-11.382.048L.875 37.958c-2.746-4.814 1.371-10.646 6.827-9.67l120.385 21.517a6.537 6.537 0 0 0 2.322-.004l117.867-21.483c5.438-.991 9.574 4.796 6.877 9.62Z"></path><path fill="url(#IconifyId1813088fe1fbc01fb467)" d="M185.432.063L96.44 17.501a3.268 3.268 0 0 0-2.634 3.014l-5.474 92.456a3.268 3.268 0 0 0 3.997 3.378l24.777-5.718c2.318-.535 4.413 1.507 3.936 3.838l-7.361 36.047c-.495 2.426 1.782 4.5 4.151 3.78l15.304-4.649c2.372-.72 4.652 1.36 4.15 3.788l-11.698 56.621c-.732 3.542 3.979 5.473 5.943 2.437l1.313-2.028l72.516-144.72c1.215-2.423-.88-5.186-3.54-4.672l-25.505 4.922c-2.396.462-4.435-1.77-3.759-4.114l16.646-57.705c.677-2.35-1.37-4.583-3.769-4.113Z"></path></svg>

frontend/src/App.css

@@ -0,0 +1,654 @@
+* {
+  box-sizing: border-box;
+}
+
+body {
+  margin: 0;
+  font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, sans-serif;
+  background: var(--color-bg);
+  color: var(--color-text);
+}
+
+.app {
+  max-width: 1000px;
+  margin: 0 auto;
+  padding: 20px;
+}
+
+nav {
+  display: flex;
+  align-items: center;
+  gap: 20px;
+  padding: 15px 0;
+  border-bottom: 1px solid var(--color-border);
+  margin-bottom: 20px;
+}
+
+.theme-toggle {
+  margin-left: auto;
+}
+
+nav a {
+  color: var(--color-primary);
+  text-decoration: none;
+  font-weight: 500;
+}
+
+nav a:hover {
+  text-decoration: underline;
+}
+
+main {
+  background: var(--color-bg-card);
+  padding: 20px;
+  border-radius: 8px;
+  box-shadow: var(--shadow);
+}
+
+.header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  margin-bottom: 20px;
+}
+
+.header h1 {
+  margin: 0;
+}
+
+.btn {
+  display: inline-block;
+  padding: 8px 16px;
+  border: 1px solid var(--color-border);
+  border-radius: 4px;
+  background: var(--color-bg-card);
+  color: var(--color-text);
+  text-decoration: none;
+  cursor: pointer;
+  font-size: 14px;
+  margin-left: 8px;
+}
+
+.btn:hover {
+  background: var(--color-bg-hover);
+}
+
+.btn-primary {
+  background: var(--color-primary);
+  border-color: var(--color-primary);
+  color: white;
+}
+
+.btn-primary:hover {
+  background: var(--color-primary-hover);
+}
+
+.btn-danger {
+  background: var(--color-danger);
+  border-color: var(--color-danger);
+  color: white;
+}
+
+.btn-danger:hover {
+  background: var(--color-danger-hover);
+}
+
+.btn-small {
+  padding: 4px 8px;
+  font-size: 12px;
+}
+
+.btn:disabled {
+  opacity: 0.6;
+  cursor: not-allowed;
+}
+
+table {
+  width: 100%;
+  border-collapse: collapse;
+}
+
+th,
+td {
+  padding: 12px;
+  text-align: left;
+  border-bottom: 1px solid var(--color-border-light);
+}
+
+th {
+  font-weight: 600;
+  background: var(--color-bg-muted);
+}
+
+tr:hover {
+  background: var(--color-bg-muted);
+}
+
+.error {
+  background: var(--color-bg-error);
+  color: var(--color-text-error);
+  padding: 10px;
+  border-radius: 4px;
+  margin-bottom: 20px;
+}
+
+.info-grid {
+  display: grid;
+  grid-template-columns: repeat(auto-fill, minmax(200px, 1fr));
+  gap: 10px;
+  margin-bottom: 20px;
+}
+
+.section {
+  margin-top: 30px;
+  padding-top: 20px;
+  border-top: 1px solid var(--color-border-light);
+}
+
+.section h3 {
+  margin-top: 0;
+  margin-bottom: 15px;
+}
+
+.section h4 {
+  margin: 15px 0 10px;
+  font-size: 14px;
+  color: var(--color-text-muted);
+}
+
+.section ul {
+  list-style: none;
+  padding: 0;
+  margin: 0;
+}
+
+.section li {
+  display: flex;
+  align-items: center;
+  gap: 10px;
+  padding: 8px 0;
+  border-bottom: 1px solid var(--color-border-lighter);
+}
+
+.tag {
+  display: inline-block;
+  background: var(--color-tag-bg);
+  padding: 2px 8px;
+  border-radius: 12px;
+  font-size: 12px;
+  color: var(--color-text-muted);
+}
+
+.add-form {
+  display: flex;
+  gap: 10px;
+  margin-top: 15px;
+  flex-wrap: wrap;
+}
+
+.add-form select,
+.add-form input {
+  padding: 8px;
+  border: 1px solid var(--color-border);
+  border-radius: 4px;
+  min-width: 200px;
+  background: var(--color-bg-card);
+  color: var(--color-text);
+}
+
+.form-group {
+  margin-bottom: 20px;
+}
+
+.form-group label {
+  display: block;
+  font-weight: 500;
+  margin-bottom: 5px;
+}
+
+.form-group input,
+.form-group textarea,
+.form-group select {
+  width: 100%;
+  padding: 10px;
+  border: 1px solid var(--color-border);
+  border-radius: 4px;
+  font-size: 14px;
+  background: var(--color-bg-card);
+  color: var(--color-text);
+}
+
+.form-group textarea {
+  resize: vertical;
+}
+
+.form-row {
+  display: grid;
+  grid-template-columns: 1fr 1fr;
+  gap: 20px;
+}
+
+.checkbox-group {
+  display: flex;
+  flex-wrap: wrap;
+  gap: 15px;
+}
+
+.checkbox-label {
+  display: flex;
+  align-items: center;
+  gap: 5px;
+  cursor: pointer;
+}
+
+.form-actions {
+  display: flex;
+  gap: 10px;
+  margin-top: 30px;
+  padding-top: 20px;
+  border-top: 1px solid var(--color-border-light);
+}
+
+.need-list .header {
+  margin-bottom: 20px;
+}
+
+.need-form {
+  background: var(--color-bg-muted);
+  padding: 20px;
+  border-radius: 4px;
+  margin-bottom: 20px;
+}
+
+.need-items {
+  list-style: none;
+  padding: 0;
+}
+
+.need-items li {
+  display: flex;
+  justify-content: space-between;
+  align-items: flex-start;
+  padding: 15px;
+  border: 1px solid var(--color-border-light);
+  border-radius: 4px;
+  margin-bottom: 10px;
+}
+
+.need-info p {
+  margin: 5px 0 0;
+  color: var(--color-text-muted);
+  font-size: 14px;
+}
+
+a {
+  color: var(--color-primary);
+}
+
+a:hover {
+  text-decoration: underline;
+}
+
+/* Graph styles */
+.graph-container {
+  width: 100%;
+}
+
+.graph-hint {
+  color: var(--color-text-muted);
+  font-size: 14px;
+  margin-bottom: 15px;
+}
+
+.selected-info {
+  margin-top: 15px;
+  padding: 15px;
+  background: var(--color-bg-muted);
+  border-radius: 8px;
+}
+
+.selected-info h3 {
+  margin: 0 0 10px;
+}
+
+.selected-info p {
+  margin: 5px 0;
+  color: var(--color-text-muted);
+}
+
+.legend {
+  margin-top: 20px;
+  padding: 15px;
+  background: var(--color-bg-muted);
+  border-radius: 8px;
+}
+
+.legend h4 {
+  margin: 0 0 10px;
+  font-size: 14px;
+}
+
+.legend-items {
+  display: flex;
+  flex-wrap: wrap;
+  gap: 15px;
+}
+
+.legend-item {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  font-size: 12px;
+  color: var(--color-text-muted);
+}
+
+.legend-line {
+  width: 30px;
+  border-radius: 2px;
+}
+
+/* Weight control styles */
+.weight-control {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  font-size: 12px;
+  color: var(--color-text-muted);
+}
+
+.weight-control input[type="range"] {
+  width: 80px;
+  cursor: pointer;
+}
+
+.weight-value {
+  min-width: 20px;
+  text-align: center;
+  font-weight: 600;
+}
+
+.weight-display {
+  font-size: 12px;
+  color: var(--color-text-muted);
+  margin-left: auto;
+}
+
+/* ID Card Styles */
+.id-card {
+  width: 100%;
+}
+
+.id-card-inner {
+  background: linear-gradient(135deg, #0a0a0f 0%, #1a1a2e 50%, #0a0a0f 100%);
+  background-image:
+    radial-gradient(white 1px, transparent 1px),
+    linear-gradient(135deg, #0a0a0f 0%, #1a1a2e 50%, #0a0a0f 100%);
+  background-size: 50px 50px, 100% 100%;
+  background-position: 0 0, 0 0;
+  color: #fff;
+  border-radius: 12px;
+  padding: 25px;
+  min-height: 500px;
+  position: relative;
+  overflow: hidden;
+}
+
+.id-card-header {
+  display: flex;
+  justify-content: space-between;
+  align-items: flex-start;
+  margin-bottom: 15px;
+}
+
+.id-card-header-left {
+  flex: 1;
+}
+
+.id-card-header-right {
+  display: flex;
+  flex-direction: column;
+  align-items: flex-end;
+  gap: 10px;
+}
+
+.id-card-title {
+  font-size: 2.5rem;
+  font-weight: 700;
+  margin: 0;
+  color: #fff;
+  text-shadow: 2px 2px 4px rgba(0,0,0,0.5);
+}
+
+.id-profile-pic {
+  width: 80px;
+  height: 80px;
+  border-radius: 8px;
+  object-fit: cover;
+  border: 2px solid rgba(255,255,255,0.3);
+}
+
+.id-profile-placeholder {
+  width: 80px;
+  height: 80px;
+  border-radius: 8px;
+  background: linear-gradient(135deg, #4ecdc4 0%, #44a8a0 100%);
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  border: 2px solid rgba(255,255,255,0.3);
+}
+
+.id-profile-placeholder span {
+  font-size: 2rem;
+  font-weight: 700;
+  color: #fff;
+  text-shadow: 1px 1px 2px rgba(0,0,0,0.3);
+}
+
+.id-card-actions {
+  display: flex;
+  gap: 8px;
+}
+
+.id-card-actions .btn {
+  background: rgba(255,255,255,0.1);
+  border-color: rgba(255,255,255,0.3);
+  color: #fff;
+}
+
+.id-card-actions .btn:hover {
+  background: rgba(255,255,255,0.2);
+}
+
+.id-card-body {
+  display: grid;
+  grid-template-columns: 1fr 1.5fr;
+  gap: 30px;
+}
+
+.id-card-left {
+  display: flex;
+  flex-direction: column;
+  gap: 8px;
+}
+
+.id-field {
+  font-size: 1rem;
+  line-height: 1.4;
+}
+
+.id-field-block {
+  margin-top: 15px;
+  font-size: 0.95rem;
+  line-height: 1.5;
+}
+
+.id-label {
+  color: #4ecdc4;
+  font-weight: 500;
+}
+
+.id-card-right {
+  display: flex;
+  flex-direction: column;
+  gap: 20px;
+}
+
+.id-bio {
+  font-size: 0.9rem;
+  line-height: 1.6;
+  color: #e0e0e0;
+}
+
+.id-relationships {
+  margin-top: 10px;
+}
+
+.id-section-title {
+  font-size: 1.5rem;
+  margin: 0 0 15px;
+  color: #fff;
+  border-bottom: 1px solid rgba(255,255,255,0.2);
+  padding-bottom: 8px;
+}
+
+.id-rel-group {
+  margin-bottom: 12px;
+  font-size: 0.9rem;
+  line-height: 1.6;
+}
+
+.id-rel-label {
+  color: #a0a0a0;
+}
+
+.id-rel-group a {
+  color: #4ecdc4;
+  text-decoration: none;
+}
+
+.id-rel-group a:hover {
+  text-decoration: underline;
+}
+
+.id-rel-type {
+  color: #888;
+  font-size: 0.85em;
+}
+
+.id-card-warnings {
+  margin-top: 30px;
+  padding-top: 20px;
+  border-top: 1px solid rgba(255,255,255,0.2);
+  display: flex;
+  flex-wrap: wrap;
+  gap: 20px;
+}
+
+.id-warning {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  font-size: 0.9rem;
+  color: #ff6b6b;
+}
+
+.warning-dot {
+  width: 8px;
+  height: 8px;
+  background: #ff6b6b;
+  border-radius: 50%;
+  flex-shrink: 0;
+}
+
+.warning-desc {
+  color: #ccc;
+}
+
+/* Management section */
+.id-card-manage {
+  margin-top: 20px;
+  background: var(--color-bg-muted);
+  border-radius: 8px;
+  padding: 15px;
+}
+
+.id-card-manage summary {
+  cursor: pointer;
+  font-weight: 600;
+  font-size: 1.1rem;
+  padding: 5px 0;
+}
+
+.id-card-manage[open] summary {
+  margin-bottom: 15px;
+  border-bottom: 1px solid var(--color-border-light);
+  padding-bottom: 10px;
+}
+
+.manage-section {
+  margin-bottom: 25px;
+}
+
+.manage-section h3 {
+  margin: 0 0 15px;
+  font-size: 1rem;
+}
+
+.manage-relationships {
+  display: flex;
+  flex-direction: column;
+  gap: 10px;
+  margin-bottom: 15px;
+}
+
+.manage-rel-item {
+  display: flex;
+  align-items: center;
+  gap: 12px;
+  padding: 10px;
+  background: var(--color-bg-card);
+  border-radius: 6px;
+  flex-wrap: wrap;
+}
+
+.manage-rel-item a {
+  font-weight: 500;
+  min-width: 120px;
+}
+
+.manage-needs-list {
+  list-style: none;
+  padding: 0;
+  margin: 0 0 15px;
+}
+
+.manage-needs-list li {
+  display: flex;
+  align-items: center;
+  gap: 12px;
+  padding: 10px;
+  background: var(--color-bg-card);
+  border-radius: 6px;
+  margin-bottom: 8px;
+}
+
+.manage-needs-list li .btn {
+  margin-left: auto;
+}
+
+/* Responsive adjustments */
+@media (max-width: 768px) {
+  .id-card-body {
+    grid-template-columns: 1fr;
+  }
+
+  .id-card-title {
+    font-size: 1.8rem;
+  }
+
+  .id-card-header {
+    flex-direction: column;
+    gap: 15px;
+  }
+}

frontend/src/App.tsx

@@ -0,0 +1,50 @@
+import { useEffect, useState } from "react";
+import { Link, Route, Routes } from "react-router-dom";
+import { ContactDetail } from "./components/ContactDetail";
+import { ContactForm } from "./components/ContactForm";
+import { ContactList } from "./components/ContactList";
+import { NeedList } from "./components/NeedList";
+import { RelationshipGraph } from "./components/RelationshipGraph";
+import "./App.css";
+
+function App() {
+  const [theme, setTheme] = useState<"light" | "dark">(() => {
+    return (localStorage.getItem("theme") as "light" | "dark") || "light";
+  });
+
+  useEffect(() => {
+    document.documentElement.setAttribute("data-theme", theme);
+    localStorage.setItem("theme", theme);
+  }, [theme]);
+
+  const toggleTheme = () => {
+    setTheme((prev) => (prev === "light" ? "dark" : "light"));
+  };
+
+  return (
+    <div className="app">
+      <nav>
+        <Link to="/contacts">Contacts</Link>
+        <Link to="/graph">Graph</Link>
+        <Link to="/needs">Needs</Link>
+        <button className="btn btn-small theme-toggle" onClick={toggleTheme}>
+          {theme === "light" ? "Dark" : "Light"}
+        </button>
+      </nav>
+
+      <main>
+        <Routes>
+          <Route path="/" element={<ContactList />} />
+          <Route path="/contacts" element={<ContactList />} />
+          <Route path="/contacts/new" element={<ContactForm />} />
+          <Route path="/contacts/:id" element={<ContactDetail />} />
+          <Route path="/contacts/:id/edit" element={<ContactForm />} />
+          <Route path="/graph" element={<RelationshipGraph />} />
+          <Route path="/needs" element={<NeedList />} />
+        </Routes>
+      </main>
+    </div>
+  );
+}
+
+export default App;

frontend/src/api/client.ts

@@ -0,0 +1,105 @@
+import type {
+  Contact,
+  ContactCreate,
+  ContactListItem,
+  ContactRelationship,
+  ContactRelationshipCreate,
+  ContactRelationshipUpdate,
+  ContactUpdate,
+  GraphData,
+  Need,
+  NeedCreate,
+} from "../types";
+
+const API_BASE = "";
+
+async function request<T>(
+  endpoint: string,
+  options?: RequestInit
+): Promise<T> {
+  const response = await fetch(`${API_BASE}${endpoint}`, {
+    ...options,
+    headers: {
+      "Content-Type": "application/json",
+      ...options?.headers,
+    },
+  });
+
+  if (!response.ok) {
+    const error = await response.json().catch(() => ({}));
+    throw new Error(error.detail || `HTTP ${response.status}`);
+  }
+
+  return response.json();
+}
+
+export const api = {
+  // Needs
+  needs: {
+    list: () => request<Need[]>("/api/needs"),
+    get: (id: number) => request<Need>(`/api/needs/${id}`),
+    create: (data: NeedCreate) =>
+      request<Need>("/api/needs", {
+        method: "POST",
+        body: JSON.stringify(data),
+      }),
+    delete: (id: number) =>
+      request<{ deleted: boolean }>(`/api/needs/${id}`, { method: "DELETE" }),
+  },
+
+  // Contacts
+  contacts: {
+    list: (skip = 0, limit = 100) =>
+      request<ContactListItem[]>(`/api/contacts?skip=${skip}&limit=${limit}`),
+    get: (id: number) => request<Contact>(`/api/contacts/${id}`),
+    create: (data: ContactCreate) =>
+      request<Contact>("/api/contacts", {
+        method: "POST",
+        body: JSON.stringify(data),
+      }),
+    update: (id: number, data: ContactUpdate) =>
+      request<Contact>(`/api/contacts/${id}`, {
+        method: "PATCH",
+        body: JSON.stringify(data),
+      }),
+    delete: (id: number) =>
+      request<{ deleted: boolean }>(`/api/contacts/${id}`, { method: "DELETE" }),
+
+    // Contact-Need relationships
+    addNeed: (contactId: number, needId: number) =>
+      request<{ added: boolean }>(`/api/contacts/${contactId}/needs/${needId}`, {
+        method: "POST",
+      }),
+    removeNeed: (contactId: number, needId: number) =>
+      request<{ removed: boolean }>(`/api/contacts/${contactId}/needs/${needId}`, {
+        method: "DELETE",
+      }),
+
+    // Contact-Contact relationships
+    getRelationships: (contactId: number) =>
+      request<ContactRelationship[]>(`/api/contacts/${contactId}/relationships`),
+    addRelationship: (contactId: number, data: ContactRelationshipCreate) =>
+      request<ContactRelationship>(`/api/contacts/${contactId}/relationships`, {
+        method: "POST",
+        body: JSON.stringify(data),
+      }),
+    updateRelationship: (contactId: number, relatedContactId: number, data: ContactRelationshipUpdate) =>
+      request<ContactRelationship>(
+        `/api/contacts/${contactId}/relationships/${relatedContactId}`,
+        {
+          method: "PATCH",
+          body: JSON.stringify(data),
+        }
+      ),
+    removeRelationship: (contactId: number, relatedContactId: number) =>
+      request<{ deleted: boolean }>(
+        `/api/contacts/${contactId}/relationships/${relatedContactId}`,
+        { method: "DELETE" }
+      ),
+  },
+
+  // Graph
+  graph: {
+    get: () => request<GraphData>("/api/graph"),
+  },
+};

frontend/src/assets/react.svg

@@ -0,0 +1 @@
+<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" class="iconify iconify--logos" width="35.93" height="32" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 228"><path fill="#00D8FF" d="M210.483 73.824a171.49 171.49 0 0 0-8.24-2.597c.465-1.9.893-3.777 1.273-5.621c6.238-30.281 2.16-54.676-11.769-62.708c-13.355-7.7-35.196.329-57.254 19.526a171.23 171.23 0 0 0-6.375 5.848a155.866 155.866 0 0 0-4.241-3.917C100.759 3.829 77.587-4.822 63.673 3.233C50.33 10.957 46.379 33.89 51.995 62.588a170.974 170.974 0 0 0 1.892 8.48c-3.28.932-6.445 1.924-9.474 2.98C17.309 83.498 0 98.307 0 113.668c0 15.865 18.582 31.778 46.812 41.427a145.52 145.52 0 0 0 6.921 2.165a167.467 167.467 0 0 0-2.01 9.138c-5.354 28.2-1.173 50.591 12.134 58.266c13.744 7.926 36.812-.22 59.273-19.855a145.567 145.567 0 0 0 5.342-4.923a168.064 168.064 0 0 0 6.92 6.314c21.758 18.722 43.246 26.282 56.54 18.586c13.731-7.949 18.194-32.003 12.4-61.268a145.016 145.016 0 0 0-1.535-6.842c1.62-.48 3.21-.974 4.76-1.488c29.348-9.723 48.443-25.443 48.443-41.52c0-15.417-17.868-30.326-45.517-39.844Zm-6.365 70.984c-1.4.463-2.836.91-4.3 1.345c-3.24-10.257-7.612-21.163-12.963-32.432c5.106-11 9.31-21.767 12.459-31.957c2.619.758 5.16 1.557 7.61 2.4c23.69 8.156 38.14 20.213 38.14 29.504c0 9.896-15.606 22.743-40.946 31.14Zm-10.514 20.834c2.562 12.94 2.927 24.64 1.23 33.787c-1.524 8.219-4.59 13.698-8.382 15.893c-8.067 4.67-25.32-1.4-43.927-17.412a156.726 156.726 0 0 1-6.437-5.87c7.214-7.889 14.423-17.06 21.459-27.246c12.376-1.098 24.068-2.894 34.671-5.345a134.17 134.17 0 0 1 1.386 6.193ZM87.276 214.515c-7.882 2.783-14.16 2.863-17.955.675c-8.075-4.657-11.432-22.636-6.853-46.752a156.923 156.923 0 0 1 1.869-8.499c10.486 2.32 22.093 3.988 34.498 4.994c7.084 9.967 14.501 19.128 21.976 27.15a134.668 134.668 0 0 1-4.877 4.492c-9.933 8.682-19.886 14.842-28.658 17.94ZM50.35 144.747c-12.483-4.267-22.792-9.812-29.858-15.863c-6.35-5.437-9.555-10.836-9.555-15.216c0-9.322 13.897-21.212 37.076-29.293c2.813-.98 5.757-1.905 8.812-2.773c3.204 10.42 7.406 21.315 12.477 32.332c-5.137 11.18-9.399 22.249-12.634 32.792a134.718 134.718 0 0 1-6.318-1.979Zm12.378-84.26c-4.811-24.587-1.616-43.134 6.425-47.789c8.564-4.958 27.502 2.111 47.463 19.835a144.318 144.318 0 0 1 3.841 3.545c-7.438 7.987-14.787 17.08-21.808 26.988c-12.04 1.116-23.565 2.908-34.161 5.309a160.342 160.342 0 0 1-1.76-7.887Zm110.427 27.268a347.8 347.8 0 0 0-7.785-12.803c8.168 1.033 15.994 2.404 23.343 4.08c-2.206 7.072-4.956 14.465-8.193 22.045a381.151 381.151 0 0 0-7.365-13.322Zm-45.032-43.861c5.044 5.465 10.096 11.566 15.065 18.186a322.04 322.04 0 0 0-30.257-.006c4.974-6.559 10.069-12.652 15.192-18.18ZM82.802 87.83a323.167 323.167 0 0 0-7.227 13.238c-3.184-7.553-5.909-14.98-8.134-22.152c7.304-1.634 15.093-2.97 23.209-3.984a321.524 321.524 0 0 0-7.848 12.897Zm8.081 65.352c-8.385-.936-16.291-2.203-23.593-3.793c2.26-7.3 5.045-14.885 8.298-22.6a321.187 321.187 0 0 0 7.257 13.246c2.594 4.48 5.28 8.868 8.038 13.147Zm37.542 31.03c-5.184-5.592-10.354-11.779-15.403-18.433c4.902.192 9.899.29 14.978.29c5.218 0 10.376-.117 15.453-.343c-4.985 6.774-10.018 12.97-15.028 18.486Zm52.198-57.817c3.422 7.8 6.306 15.345 8.596 22.52c-7.422 1.694-15.436 3.058-23.88 4.071a382.417 382.417 0 0 0 7.859-13.026a347.403 347.403 0 0 0 7.425-13.565Zm-16.898 8.101a358.557 358.557 0 0 1-12.281 19.815a329.4 329.4 0 0 1-23.444.823c-7.967 0-15.716-.248-23.178-.732a310.202 310.202 0 0 1-12.513-19.846h.001a307.41 307.41 0 0 1-10.923-20.627a310.278 310.278 0 0 1 10.89-20.637l-.001.001a307.318 307.318 0 0 1 12.413-19.761c7.613-.576 15.42-.876 23.31-.876H128c7.926 0 15.743.303 23.354.883a329.357 329.357 0 0 1 12.335 19.695a358.489 358.489 0 0 1 11.036 20.54a329.472 329.472 0 0 1-11 20.722Zm22.56-122.124c8.572 4.944 11.906 24.881 6.52 51.026c-.344 1.668-.73 3.367-1.15 5.09c-10.622-2.452-22.155-4.275-34.23-5.408c-7.034-10.017-14.323-19.124-21.64-27.008a160.789 160.789 0 0 1 5.888-5.4c18.9-16.447 36.564-22.941 44.612-18.3ZM128 90.808c12.625 0 22.86 10.235 22.86 22.86s-10.235 22.86-22.86 22.86s-22.86-10.235-22.86-22.86s10.235-22.86 22.86-22.86Z"></path></svg>

frontend/src/components/ContactDetail.tsx

@@ -0,0 +1,456 @@
+import { useEffect, useState } from "react";
+import { Link, useParams } from "react-router-dom";
+import { api } from "../api/client";
+import type { Contact, ContactListItem, Need, RelationshipTypeValue } from "../types";
+import { RELATIONSHIP_TYPES } from "../types";
+
+export function ContactDetail() {
+  const { id } = useParams<{ id: string }>();
+  const [contact, setContact] = useState<Contact | null>(null);
+  const [allNeeds, setAllNeeds] = useState<Need[]>([]);
+  const [allContacts, setAllContacts] = useState<ContactListItem[]>([]);
+  const [loading, setLoading] = useState(true);
+  const [error, setError] = useState<string | null>(null);
+
+  const [newNeedId, setNewNeedId] = useState<number | "">("");
+  const [newRelContactId, setNewRelContactId] = useState<number | "">("");
+  const [newRelType, setNewRelType] = useState<RelationshipTypeValue | "">("");
+
+  useEffect(() => {
+    if (!id) return;
+    Promise.all([
+      api.contacts.get(Number(id)),
+      api.needs.list(),
+      api.contacts.list(),
+    ])
+      .then(([c, n, contacts]) => {
+        setContact(c);
+        setAllNeeds(n);
+        setAllContacts(contacts.filter((ct) => ct.id !== Number(id)));
+      })
+      .catch((err) => setError(err.message))
+      .finally(() => setLoading(false));
+  }, [id]);
+
+  const handleAddNeed = async () => {
+    if (!contact || newNeedId === "") return;
+    try {
+      await api.contacts.addNeed(contact.id, Number(newNeedId));
+      const updated = await api.contacts.get(contact.id);
+      setContact(updated);
+      setNewNeedId("");
+    } catch (err) {
+      setError(err instanceof Error ? err.message : "Failed to add need");
+    }
+  };
+
+  const handleRemoveNeed = async (needId: number) => {
+    if (!contact) return;
+    try {
+      await api.contacts.removeNeed(contact.id, needId);
+      const updated = await api.contacts.get(contact.id);
+      setContact(updated);
+    } catch (err) {
+      setError(err instanceof Error ? err.message : "Failed to remove need");
+    }
+  };
+
+  const handleAddRelationship = async () => {
+    if (!contact || newRelContactId === "" || newRelType === "") return;
+    try {
+      await api.contacts.addRelationship(contact.id, {
+        related_contact_id: Number(newRelContactId),
+        relationship_type: newRelType,
+      });
+      const updated = await api.contacts.get(contact.id);
+      setContact(updated);
+      setNewRelContactId("");
+      setNewRelType("");
+    } catch (err) {
+      setError(
+        err instanceof Error ? err.message : "Failed to add relationship"
+      );
+    }
+  };
+
+  const handleRemoveRelationship = async (relatedContactId: number) => {
+    if (!contact) return;
+    try {
+      await api.contacts.removeRelationship(contact.id, relatedContactId);
+      const updated = await api.contacts.get(contact.id);
+      setContact(updated);
+    } catch (err) {
+      setError(
+        err instanceof Error ? err.message : "Failed to remove relationship"
+      );
+    }
+  };
+
+  const handleUpdateWeight = async (relatedContactId: number, newWeight: number) => {
+    if (!contact) return;
+    try {
+      await api.contacts.updateRelationship(contact.id, relatedContactId, {
+        closeness_weight: newWeight,
+      });
+      const updated = await api.contacts.get(contact.id);
+      setContact(updated);
+    } catch (err) {
+      setError(
+        err instanceof Error ? err.message : "Failed to update weight"
+      );
+    }
+  };
+
+  if (loading) return <div>Loading...</div>;
+  if (error) return <div className="error">Error: {error}</div>;
+  if (!contact) return <div>Contact not found</div>;
+
+  const availableNeeds = allNeeds.filter(
+    (n) => !contact.needs.some((cn) => cn.id === n.id)
+  );
+
+  const getContactName = (contactId: number) => {
+    const c = allContacts.find((ct) => ct.id === contactId);
+    return c?.name || `Contact #${contactId}`;
+  };
+
+  const getRelationshipDisplayName = (type: string) => {
+    const rt = RELATIONSHIP_TYPES.find((r) => r.value === type);
+    return rt?.displayName || type;
+  };
+
+  // Group relationships by category for display
+  const groupRelationships = () => {
+    const familial: typeof contact.related_to = [];
+    const friends: typeof contact.related_to = [];
+    const partners: typeof contact.related_to = [];
+    const professional: typeof contact.related_to = [];
+    const other: typeof contact.related_to = [];
+
+    const familialTypes = ['parent', 'child', 'sibling', 'grandparent', 'grandchild', 'aunt_uncle', 'niece_nephew', 'cousin', 'in_law'];
+    const friendTypes = ['best_friend', 'close_friend', 'friend', 'acquaintance', 'neighbor'];
+    const partnerTypes = ['spouse', 'partner'];
+    const professionalTypes = ['mentor', 'mentee', 'business_partner', 'colleague', 'manager', 'direct_report', 'client'];
+
+    for (const rel of contact.related_to) {
+      if (familialTypes.includes(rel.relationship_type)) {
+        familial.push(rel);
+      } else if (friendTypes.includes(rel.relationship_type)) {
+        friends.push(rel);
+      } else if (partnerTypes.includes(rel.relationship_type)) {
+        partners.push(rel);
+      } else if (professionalTypes.includes(rel.relationship_type)) {
+        professional.push(rel);
+      } else {
+        other.push(rel);
+      }
+    }
+
+    return { familial, friends, partners, professional, other };
+  };
+
+  const relationshipGroups = groupRelationships();
+
+  return (
+    <div className="id-card">
+      <div className="id-card-inner">
+        {/* Header with name and profile pic */}
+        <div className="id-card-header">
+          <div className="id-card-header-left">
+            <h1 className="id-card-title">I.D.: {contact.name}</h1>
+          </div>
+          <div className="id-card-header-right">
+            {contact.profile_pic ? (
+              <img
+                src={contact.profile_pic}
+                alt={`${contact.name}'s profile`}
+                className="id-profile-pic"
+              />
+            ) : (
+              <div className="id-profile-placeholder">
+                <span>{contact.name.charAt(0).toUpperCase()}</span>
+              </div>
+            )}
+            <div className="id-card-actions">
+              <Link to={`/contacts/${contact.id}/edit`} className="btn btn-small">
+                Edit
+              </Link>
+              <Link to="/contacts" className="btn btn-small">
+                Back
+              </Link>
+            </div>
+          </div>
+        </div>
+
+        <div className="id-card-body">
+          {/* Left column - Basic info */}
+          <div className="id-card-left">
+            {contact.legal_name && (
+              <div className="id-field">Legal name: {contact.legal_name}</div>
+            )}
+            {contact.suffix && (
+              <div className="id-field">Suffix: {contact.suffix}</div>
+            )}
+            {contact.gender && (
+              <div className="id-field">Gender: {contact.gender}</div>
+            )}
+            {contact.age && (
+              <div className="id-field">Age: {contact.age}</div>
+            )}
+            {contact.current_job && (
+              <div className="id-field">Job: {contact.current_job}</div>
+            )}
+            {contact.social_structure_style && (
+              <div className="id-field">Social style: {contact.social_structure_style}</div>
+            )}
+            {contact.self_sufficiency_score !== null && (
+              <div className="id-field">Self-Sufficiency: {contact.self_sufficiency_score}</div>
+            )}
+            {contact.timezone && (
+              <div className="id-field">Timezone: {contact.timezone}</div>
+            )}
+
+            {contact.safe_conversation_starters && (
+              <div className="id-field-block">
+                <span className="id-label">Safe con starters:</span> {contact.safe_conversation_starters}
+              </div>
+            )}
+
+            {contact.topics_to_avoid && (
+              <div className="id-field-block">
+                <span className="id-label">Topics to avoid:</span> {contact.topics_to_avoid}
+              </div>
+            )}
+
+            {contact.goals && (
+              <div className="id-field-block">
+                <span className="id-label">Goals:</span> {contact.goals}
+              </div>
+            )}
+          </div>
+
+          {/* Right column - Bio and Relationships */}
+          <div className="id-card-right">
+            {contact.bio && (
+              <div className="id-bio">
+                <span className="id-label">Bio:</span> {contact.bio}
+              </div>
+            )}
+
+            <div className="id-relationships">
+              <h2 className="id-section-title">Relationships</h2>
+
+              {relationshipGroups.familial.length > 0 && (
+                <div className="id-rel-group">
+                  <span className="id-rel-label">Familial:</span>{" "}
+                  {relationshipGroups.familial.map((rel, i) => (
+                    <span key={rel.related_contact_id}>
+                      <Link to={`/contacts/${rel.related_contact_id}`}>
+                        {getContactName(rel.related_contact_id)}
+                      </Link>
+                      <span className="id-rel-type">({getRelationshipDisplayName(rel.relationship_type)})</span>
+                      {i < relationshipGroups.familial.length - 1 && ", "}
+                    </span>
+                  ))}
+                </div>
+              )}
+
+              {relationshipGroups.partners.length > 0 && (
+                <div className="id-rel-group">
+                  <span className="id-rel-label">Partners:</span>{" "}
+                  {relationshipGroups.partners.map((rel, i) => (
+                    <span key={rel.related_contact_id}>
+                      <Link to={`/contacts/${rel.related_contact_id}`}>
+                        {getContactName(rel.related_contact_id)}
+                      </Link>
+                      {i < relationshipGroups.partners.length - 1 && ", "}
+                    </span>
+                  ))}
+                </div>
+              )}
+
+              {relationshipGroups.friends.length > 0 && (
+                <div className="id-rel-group">
+                  <span className="id-rel-label">Friends:</span>{" "}
+                  {relationshipGroups.friends.map((rel, i) => (
+                    <span key={rel.related_contact_id}>
+                      <Link to={`/contacts/${rel.related_contact_id}`}>
+                        {getContactName(rel.related_contact_id)}
+                      </Link>
+                      {i < relationshipGroups.friends.length - 1 && ", "}
+                    </span>
+                  ))}
+                </div>
+              )}
+
+              {relationshipGroups.professional.length > 0 && (
+                <div className="id-rel-group">
+                  <span className="id-rel-label">Professional:</span>{" "}
+                  {relationshipGroups.professional.map((rel, i) => (
+                    <span key={rel.related_contact_id}>
+                      <Link to={`/contacts/${rel.related_contact_id}`}>
+                        {getContactName(rel.related_contact_id)}
+                      </Link>
+                      <span className="id-rel-type">({getRelationshipDisplayName(rel.relationship_type)})</span>
+                      {i < relationshipGroups.professional.length - 1 && ", "}
+                    </span>
+                  ))}
+                </div>
+              )}
+
+              {relationshipGroups.other.length > 0 && (
+                <div className="id-rel-group">
+                  <span className="id-rel-label">Other:</span>{" "}
+                  {relationshipGroups.other.map((rel, i) => (
+                    <span key={rel.related_contact_id}>
+                      <Link to={`/contacts/${rel.related_contact_id}`}>
+                        {getContactName(rel.related_contact_id)}
+                      </Link>
+                      <span className="id-rel-type">({getRelationshipDisplayName(rel.relationship_type)})</span>
+                      {i < relationshipGroups.other.length - 1 && ", "}
+                    </span>
+                  ))}
+                </div>
+              )}
+
+              {contact.related_from.length > 0 && (
+                <div className="id-rel-group">
+                  <span className="id-rel-label">Known by:</span>{" "}
+                  {contact.related_from.map((rel, i) => (
+                    <span key={rel.contact_id}>
+                      <Link to={`/contacts/${rel.contact_id}`}>
+                        {getContactName(rel.contact_id)}
+                      </Link>
+                      {i < contact.related_from.length - 1 && ", "}
+                    </span>
+                  ))}
+                </div>
+              )}
+            </div>
+          </div>
+        </div>
+
+        {/* Needs/Warnings at bottom */}
+        {contact.needs.length > 0 && (
+          <div className="id-card-warnings">
+            {contact.needs.map((need) => (
+              <div key={need.id} className="id-warning">
+                <span className="warning-dot"></span>
+                Warning: {need.name}
+                {need.description && <span className="warning-desc"> - {need.description}</span>}
+              </div>
+            ))}
+          </div>
+        )}
+      </div>
+
+      {/* Management section (expandable) */}
+      <details className="id-card-manage">
+        <summary>Manage Contact</summary>
+
+        <div className="manage-section">
+          <h3>Manage Relationships</h3>
+          <div className="manage-relationships">
+            {contact.related_to.map((rel) => (
+              <div key={rel.related_contact_id} className="manage-rel-item">
+                <Link to={`/contacts/${rel.related_contact_id}`}>
+                  {getContactName(rel.related_contact_id)}
+                </Link>
+                <span className="tag">{getRelationshipDisplayName(rel.relationship_type)}</span>
+                <label className="weight-control">
+                  <span>Closeness:</span>
+                  <input
+                    type="range"
+                    min="1"
+                    max="10"
+                    value={rel.closeness_weight}
+                    onChange={(e) => handleUpdateWeight(rel.related_contact_id, Number(e.target.value))}
+                  />
+                  <span className="weight-value">{rel.closeness_weight}</span>
+                </label>
+                <button
+                  onClick={() => handleRemoveRelationship(rel.related_contact_id)}
+                  className="btn btn-small btn-danger"
+                >
+                  Remove
+                </button>
+              </div>
+            ))}
+          </div>
+
+          {allContacts.length > 0 && (
+            <div className="add-form">
+              <select
+                value={newRelContactId}
+                onChange={(e) =>
+                  setNewRelContactId(
+                    e.target.value ? Number(e.target.value) : ""
+                  )
+                }
+              >
+                <option value="">Select contact...</option>
+                {allContacts.map((c) => (
+                  <option key={c.id} value={c.id}>
+                    {c.name}
+                  </option>
+                ))}
+              </select>
+              <select
+                value={newRelType}
+                onChange={(e) => setNewRelType(e.target.value as RelationshipTypeValue | "")}
+              >
+                <option value="">Select relationship type...</option>
+                {RELATIONSHIP_TYPES.map((rt) => (
+                  <option key={rt.value} value={rt.value}>
+                    {rt.displayName}
+                  </option>
+                ))}
+              </select>
+              <button onClick={handleAddRelationship} className="btn btn-primary">
+                Add Relationship
+              </button>
+            </div>
+          )}
+        </div>
+
+        <div className="manage-section">
+          <h3>Manage Needs/Warnings</h3>
+          <ul className="manage-needs-list">
+            {contact.needs.map((need) => (
+              <li key={need.id}>
+                <strong>{need.name}</strong>
+                {need.description && <span> - {need.description}</span>}
+                <button
+                  onClick={() => handleRemoveNeed(need.id)}
+                  className="btn btn-small btn-danger"
+                >
+                  Remove
+                </button>
+              </li>
+            ))}
+          </ul>
+          {availableNeeds.length > 0 && (
+            <div className="add-form">
+              <select
+                value={newNeedId}
+                onChange={(e) =>
+                  setNewNeedId(e.target.value ? Number(e.target.value) : "")
+                }
+              >
+                <option value="">Select a need...</option>
+                {availableNeeds.map((n) => (
+                  <option key={n.id} value={n.id}>
+                    {n.name}
+                  </option>
+                ))}
+              </select>
+              <button onClick={handleAddNeed} className="btn btn-primary">
+                Add Need
+              </button>
+            </div>
+          )}
+        </div>
+      </details>
+    </div>
+  );
+}

frontend/src/components/ContactForm.tsx

@@ -0,0 +1,325 @@
+import { useEffect, useState } from "react";
+import { useNavigate, useParams } from "react-router-dom";
+import { api } from "../api/client";
+import type { ContactCreate, Need } from "../types";
+
+export function ContactForm() {
+  const { id } = useParams<{ id: string }>();
+  const navigate = useNavigate();
+  const isEdit = Boolean(id);
+
+  const [allNeeds, setAllNeeds] = useState<Need[]>([]);
+  const [loading, setLoading] = useState(isEdit);
+  const [error, setError] = useState<string | null>(null);
+  const [submitting, setSubmitting] = useState(false);
+
+  const [form, setForm] = useState<ContactCreate>({
+    name: "",
+    age: null,
+    bio: null,
+    current_job: null,
+    gender: null,
+    goals: null,
+    legal_name: null,
+    profile_pic: null,
+    safe_conversation_starters: null,
+    self_sufficiency_score: null,
+    social_structure_style: null,
+    ssn: null,
+    suffix: null,
+    timezone: null,
+    topics_to_avoid: null,
+    need_ids: [],
+  });
+
+  useEffect(() => {
+    const loadData = async () => {
+      try {
+        const needs = await api.needs.list();
+        setAllNeeds(needs);
+
+        if (id) {
+          const contact = await api.contacts.get(Number(id));
+          setForm({
+            name: contact.name,
+            age: contact.age,
+            bio: contact.bio,
+            current_job: contact.current_job,
+            gender: contact.gender,
+            goals: contact.goals,
+            legal_name: contact.legal_name,
+            profile_pic: contact.profile_pic,
+            safe_conversation_starters: contact.safe_conversation_starters,
+            self_sufficiency_score: contact.self_sufficiency_score,
+            social_structure_style: contact.social_structure_style,
+            ssn: contact.ssn,
+            suffix: contact.suffix,
+            timezone: contact.timezone,
+            topics_to_avoid: contact.topics_to_avoid,
+            need_ids: contact.needs.map((n) => n.id),
+          });
+        }
+      } catch (err) {
+        setError(err instanceof Error ? err.message : "Failed to load data");
+      } finally {
+        setLoading(false);
+      }
+    };
+    loadData();
+  }, [id]);
+
+  const handleSubmit = async (e: React.FormEvent) => {
+    e.preventDefault();
+    setSubmitting(true);
+    setError(null);
+
+    try {
+      if (isEdit) {
+        await api.contacts.update(Number(id), form);
+        navigate(`/contacts/${id}`);
+      } else {
+        const created = await api.contacts.create(form);
+        navigate(`/contacts/${created.id}`);
+      }
+    } catch (err) {
+      setError(err instanceof Error ? err.message : "Save failed");
+      setSubmitting(false);
+    }
+  };
+
+  const updateField = <K extends keyof ContactCreate>(
+    field: K,
+    value: ContactCreate[K]
+  ) => {
+    setForm((prev) => ({ ...prev, [field]: value }));
+  };
+
+  const toggleNeed = (needId: number) => {
+    setForm((prev) => ({
+      ...prev,
+      need_ids: prev.need_ids?.includes(needId)
+        ? prev.need_ids.filter((id) => id !== needId)
+        : [...(prev.need_ids || []), needId],
+    }));
+  };
+
+  if (loading) return <div>Loading...</div>;
+
+  return (
+    <div className="contact-form">
+      <h1>{isEdit ? "Edit Contact" : "New Contact"}</h1>
+
+      {error && <div className="error">{error}</div>}
+
+      <form onSubmit={handleSubmit}>
+        <div className="form-group">
+          <label htmlFor="name">Name *</label>
+          <input
+            id="name"
+            type="text"
+            value={form.name}
+            onChange={(e) => updateField("name", e.target.value)}
+            required
+          />
+        </div>
+
+        <div className="form-row">
+          <div className="form-group">
+            <label htmlFor="legal_name">Legal Name</label>
+            <input
+              id="legal_name"
+              type="text"
+              value={form.legal_name || ""}
+              onChange={(e) =>
+                updateField("legal_name", e.target.value || null)
+              }
+            />
+          </div>
+          <div className="form-group">
+            <label htmlFor="suffix">Suffix</label>
+            <input
+              id="suffix"
+              type="text"
+              value={form.suffix || ""}
+              onChange={(e) => updateField("suffix", e.target.value || null)}
+            />
+          </div>
+        </div>
+
+        <div className="form-row">
+          <div className="form-group">
+            <label htmlFor="age">Age</label>
+            <input
+              id="age"
+              type="number"
+              value={form.age ?? ""}
+              onChange={(e) =>
+                updateField("age", e.target.value ? Number(e.target.value) : null)
+              }
+            />
+          </div>
+          <div className="form-group">
+            <label htmlFor="gender">Gender</label>
+            <input
+              id="gender"
+              type="text"
+              value={form.gender || ""}
+              onChange={(e) => updateField("gender", e.target.value || null)}
+            />
+          </div>
+        </div>
+
+        <div className="form-group">
+          <label htmlFor="current_job">Current Job</label>
+          <input
+            id="current_job"
+            type="text"
+            value={form.current_job || ""}
+            onChange={(e) =>
+              updateField("current_job", e.target.value || null)
+            }
+          />
+        </div>
+
+        <div className="form-group">
+          <label htmlFor="timezone">Timezone</label>
+          <input
+            id="timezone"
+            type="text"
+            value={form.timezone || ""}
+            onChange={(e) => updateField("timezone", e.target.value || null)}
+          />
+        </div>
+
+        <div className="form-group">
+          <label htmlFor="profile_pic">Profile Picture URL</label>
+          <input
+            id="profile_pic"
+            type="url"
+            placeholder="https://example.com/photo.jpg"
+            value={form.profile_pic || ""}
+            onChange={(e) => updateField("profile_pic", e.target.value || null)}
+          />
+        </div>
+
+        <div className="form-group">
+          <label htmlFor="bio">Bio</label>
+          <textarea
+            id="bio"
+            value={form.bio || ""}
+            onChange={(e) => updateField("bio", e.target.value || null)}
+            rows={3}
+          />
+        </div>
+
+        <div className="form-group">
+          <label htmlFor="goals">Goals</label>
+          <textarea
+            id="goals"
+            value={form.goals || ""}
+            onChange={(e) => updateField("goals", e.target.value || null)}
+            rows={3}
+          />
+        </div>
+
+        <div className="form-group">
+          <label htmlFor="social_structure_style">Social Structure Style</label>
+          <input
+            id="social_structure_style"
+            type="text"
+            value={form.social_structure_style || ""}
+            onChange={(e) =>
+              updateField("social_structure_style", e.target.value || null)
+            }
+          />
+        </div>
+
+        <div className="form-group">
+          <label htmlFor="self_sufficiency_score">
+            Self-Sufficiency Score (1-10)
+          </label>
+          <input
+            id="self_sufficiency_score"
+            type="number"
+            min="1"
+            max="10"
+            value={form.self_sufficiency_score ?? ""}
+            onChange={(e) =>
+              updateField(
+                "self_sufficiency_score",
+                e.target.value ? Number(e.target.value) : null
+              )
+            }
+          />
+        </div>
+
+        <div className="form-group">
+          <label htmlFor="safe_conversation_starters">
+            Safe Conversation Starters
+          </label>
+          <textarea
+            id="safe_conversation_starters"
+            value={form.safe_conversation_starters || ""}
+            onChange={(e) =>
+              updateField("safe_conversation_starters", e.target.value || null)
+            }
+            rows={2}
+          />
+        </div>
+
+        <div className="form-group">
+          <label htmlFor="topics_to_avoid">Topics to Avoid</label>
+          <textarea
+            id="topics_to_avoid"
+            value={form.topics_to_avoid || ""}
+            onChange={(e) =>
+              updateField("topics_to_avoid", e.target.value || null)
+            }
+            rows={2}
+          />
+        </div>
+
+        <div className="form-group">
+          <label htmlFor="ssn">SSN</label>
+          <input
+            id="ssn"
+            type="text"
+            value={form.ssn || ""}
+            onChange={(e) => updateField("ssn", e.target.value || null)}
+          />
+        </div>
+
+        {allNeeds.length > 0 && (
+          <div className="form-group">
+            <label>Needs/Accommodations</label>
+            <div className="checkbox-group">
+              {allNeeds.map((need) => (
+                <label key={need.id} className="checkbox-label">
+                  <input
+                    type="checkbox"
+                    checked={form.need_ids?.includes(need.id) || false}
+                    onChange={() => toggleNeed(need.id)}
+                  />
+                  {need.name}
+                </label>
+              ))}
+            </div>
+          </div>
+        )}
+
+        <div className="form-actions">
+          <button type="submit" className="btn btn-primary" disabled={submitting}>
+            {submitting ? "Saving..." : "Save"}
+          </button>
+          <button
+            type="button"
+            className="btn"
+            onClick={() => navigate(isEdit ? `/contacts/${id}` : "/contacts")}
+          >
+            Cancel
+          </button>
+        </div>
+      </form>
+    </div>
+  );
+}

frontend/src/components/ContactList.tsx

@@ -0,0 +1,79 @@
+import { useEffect, useState } from "react";
+import { Link } from "react-router-dom";
+import { api } from "../api/client";
+import type { ContactListItem } from "../types";
+
+export function ContactList() {
+  const [contacts, setContacts] = useState<ContactListItem[]>([]);
+  const [loading, setLoading] = useState(true);
+  const [error, setError] = useState<string | null>(null);
+
+  useEffect(() => {
+    api.contacts
+      .list()
+      .then(setContacts)
+      .catch((err) => setError(err.message))
+      .finally(() => setLoading(false));
+  }, []);
+
+  const handleDelete = async (id: number) => {
+    if (!confirm("Delete this contact?")) return;
+    try {
+      await api.contacts.delete(id);
+      setContacts((prev) => prev.filter((c) => c.id !== id));
+    } catch (err) {
+      setError(err instanceof Error ? err.message : "Delete failed");
+    }
+  };
+
+  if (loading) return <div>Loading...</div>;
+  if (error) return <div className="error">Error: {error}</div>;
+
+  return (
+    <div className="contact-list">
+      <div className="header">
+        <h1>Contacts</h1>
+        <Link to="/contacts/new" className="btn btn-primary">
+          Add Contact
+        </Link>
+      </div>
+
+      {contacts.length === 0 ? (
+        <p>No contacts yet.</p>
+      ) : (
+        <table>
+          <thead>
+            <tr>
+              <th>Name</th>
+              <th>Job</th>
+              <th>Timezone</th>
+              <th>Actions</th>
+            </tr>
+          </thead>
+          <tbody>
+            {contacts.map((contact) => (
+              <tr key={contact.id}>
+                <td>
+                  <Link to={`/contacts/${contact.id}`}>{contact.name}</Link>
+                </td>
+                <td>{contact.current_job || "-"}</td>
+                <td>{contact.timezone || "-"}</td>
+                <td>
+                  <Link to={`/contacts/${contact.id}/edit`} className="btn">
+                    Edit
+                  </Link>
+                  <button
+                    onClick={() => handleDelete(contact.id)}
+                    className="btn btn-danger"
+                  >
+                    Delete
+                  </button>
+                </td>
+              </tr>
+            ))}
+          </tbody>
+        </table>
+      )}
+    </div>
+  );
+}

frontend/src/components/NeedList.tsx

@@ -0,0 +1,117 @@
+import { useEffect, useState } from "react";
+import { api } from "../api/client";
+import type { Need, NeedCreate } from "../types";
+
+export function NeedList() {
+  const [needs, setNeeds] = useState<Need[]>([]);
+  const [loading, setLoading] = useState(true);
+  const [error, setError] = useState<string | null>(null);
+  const [showForm, setShowForm] = useState(false);
+  const [form, setForm] = useState<NeedCreate>({ name: "", description: null });
+  const [submitting, setSubmitting] = useState(false);
+
+  useEffect(() => {
+    api.needs
+      .list()
+      .then(setNeeds)
+      .catch((err) => setError(err.message))
+      .finally(() => setLoading(false));
+  }, []);
+
+  const handleSubmit = async (e: React.FormEvent) => {
+    e.preventDefault();
+    if (!form.name.trim()) return;
+
+    setSubmitting(true);
+    try {
+      const created = await api.needs.create(form);
+      setNeeds((prev) => [...prev, created]);
+      setForm({ name: "", description: null });
+      setShowForm(false);
+    } catch (err) {
+      setError(err instanceof Error ? err.message : "Create failed");
+    } finally {
+      setSubmitting(false);
+    }
+  };
+
+  const handleDelete = async (id: number) => {
+    if (!confirm("Delete this need?")) return;
+    try {
+      await api.needs.delete(id);
+      setNeeds((prev) => prev.filter((n) => n.id !== id));
+    } catch (err) {
+      setError(err instanceof Error ? err.message : "Delete failed");
+    }
+  };
+
+  if (loading) return <div>Loading...</div>;
+
+  return (
+    <div className="need-list">
+      <div className="header">
+        <h1>Needs / Accommodations</h1>
+        <button
+          onClick={() => setShowForm(!showForm)}
+          className="btn btn-primary"
+        >
+          {showForm ? "Cancel" : "Add Need"}
+        </button>
+      </div>
+
+      {error && <div className="error">{error}</div>}
+
+      {showForm && (
+        <form onSubmit={handleSubmit} className="need-form">
+          <div className="form-group">
+            <label htmlFor="name">Name *</label>
+            <input
+              id="name"
+              type="text"
+              value={form.name}
+              onChange={(e) => setForm({ ...form, name: e.target.value })}
+              placeholder="e.g., Light Sensitive, ADHD"
+              required
+            />
+          </div>
+          <div className="form-group">
+            <label htmlFor="description">Description</label>
+            <textarea
+              id="description"
+              value={form.description || ""}
+              onChange={(e) =>
+                setForm({ ...form, description: e.target.value || null })
+              }
+              placeholder="Optional description..."
+              rows={2}
+            />
+          </div>
+          <button type="submit" className="btn btn-primary" disabled={submitting}>
+            {submitting ? "Creating..." : "Create"}
+          </button>
+        </form>
+      )}
+
+      {needs.length === 0 ? (
+        <p>No needs defined yet.</p>
+      ) : (
+        <ul className="need-items">
+          {needs.map((need) => (
+            <li key={need.id}>
+              <div className="need-info">
+                <strong>{need.name}</strong>
+                {need.description && <p>{need.description}</p>}
+              </div>
+              <button
+                onClick={() => handleDelete(need.id)}
+                className="btn btn-danger"
+              >
+                Delete
+              </button>
+            </li>
+          ))}
+        </ul>
+      )}
+    </div>
+  );
+}

frontend/src/components/RelationshipGraph.tsx

@@ -0,0 +1,330 @@
+import { useEffect, useRef, useState } from "react";
+import { api } from "../api/client";
+import type { GraphData, GraphEdge, GraphNode } from "../types";
+import { RELATIONSHIP_TYPES } from "../types";
+
+interface SimNode extends GraphNode {
+  x: number;
+  y: number;
+  vx: number;
+  vy: number;
+}
+
+interface SimEdge extends GraphEdge {
+  sourceNode: SimNode;
+  targetNode: SimNode;
+}
+
+export function RelationshipGraph() {
+  const canvasRef = useRef<HTMLCanvasElement>(null);
+  const [data, setData] = useState<GraphData | null>(null);
+  const [error, setError] = useState<string | null>(null);
+  const [loading, setLoading] = useState(true);
+  const [selectedNode, setSelectedNode] = useState<SimNode | null>(null);
+  const nodesRef = useRef<SimNode[]>([]);
+  const edgesRef = useRef<SimEdge[]>([]);
+  const dragNodeRef = useRef<SimNode | null>(null);
+  const animationRef = useRef<number>(0);
+
+  useEffect(() => {
+    api.graph.get()
+      .then(setData)
+      .catch((err) => setError(err.message))
+      .finally(() => setLoading(false));
+  }, []);
+
+  useEffect(() => {
+    if (!data || !canvasRef.current) return;
+
+    const canvas = canvasRef.current;
+    const maybeCtx = canvas.getContext("2d");
+    if (!maybeCtx) return;
+    const ctx: CanvasRenderingContext2D = maybeCtx;
+
+    const width = canvas.width;
+    const height = canvas.height;
+    const centerX = width / 2;
+    const centerY = height / 2;
+
+    // Initialize nodes with random positions
+    const nodes: SimNode[] = data.nodes.map((node) => ({
+      ...node,
+      x: centerX + (Math.random() - 0.5) * 300,
+      y: centerY + (Math.random() - 0.5) * 300,
+      vx: 0,
+      vy: 0,
+    }));
+    nodesRef.current = nodes;
+
+    const nodeMap = new Map(nodes.map((n) => [n.id, n]));
+
+    // Create edges with node references
+    const edges: SimEdge[] = data.edges
+      .map((edge) => {
+        const sourceNode = nodeMap.get(edge.source);
+        const targetNode = nodeMap.get(edge.target);
+        if (!sourceNode || !targetNode) return null;
+        return { ...edge, sourceNode, targetNode };
+      })
+      .filter((e): e is SimEdge => e !== null);
+    edgesRef.current = edges;
+
+    // Force simulation parameters
+    const repulsion = 5000;
+    const springStrength = 0.05;
+    const baseSpringLength = 150;
+    const damping = 0.9;
+    const centerPull = 0.01;
+
+    function simulate() {
+      const nodes = nodesRef.current;
+      const edges = edgesRef.current;
+
+      // Reset forces
+      for (const node of nodes) {
+        node.vx = 0;
+        node.vy = 0;
+      }
+
+      // Repulsion between all nodes
+      for (let i = 0; i < nodes.length; i++) {
+        for (let j = i + 1; j < nodes.length; j++) {
+          const dx = nodes[j].x - nodes[i].x;
+          const dy = nodes[j].y - nodes[i].y;
+          const dist = Math.sqrt(dx * dx + dy * dy) || 1;
+          const force = repulsion / (dist * dist);
+          const fx = (dx / dist) * force;
+          const fy = (dy / dist) * force;
+          nodes[i].vx -= fx;
+          nodes[i].vy -= fy;
+          nodes[j].vx += fx;
+          nodes[j].vy += fy;
+        }
+      }
+
+      // Spring forces for edges - closer relationships = shorter springs
+      // Weight is 1-10, normalize to 0-1 for calculations
+      for (const edge of edges) {
+        const dx = edge.targetNode.x - edge.sourceNode.x;
+        const dy = edge.targetNode.y - edge.sourceNode.y;
+        const dist = Math.sqrt(dx * dx + dy * dy) || 1;
+        // Higher weight (1-10) = shorter ideal length
+        // Normalize: weight 10 -> 0.5x length, weight 1 -> 1.4x length
+        const normalizedWeight = edge.closeness_weight / 10;
+        const idealLength = baseSpringLength * (1.5 - normalizedWeight);
+        const displacement = dist - idealLength;
+        const force = springStrength * displacement;
+        const fx = (dx / dist) * force;
+        const fy = (dy / dist) * force;
+        edge.sourceNode.vx += fx;
+        edge.sourceNode.vy += fy;
+        edge.targetNode.vx -= fx;
+        edge.targetNode.vy -= fy;
+      }
+
+      // Pull toward center
+      for (const node of nodes) {
+        node.vx += (centerX - node.x) * centerPull;
+        node.vy += (centerY - node.y) * centerPull;
+      }
+
+      // Apply velocities with damping (skip dragged node)
+      for (const node of nodes) {
+        if (node === dragNodeRef.current) continue;
+        node.x += node.vx * damping;
+        node.y += node.vy * damping;
+        // Keep within bounds
+        node.x = Math.max(30, Math.min(width - 30, node.x));
+        node.y = Math.max(30, Math.min(height - 30, node.y));
+      }
+    }
+
+    function getEdgeColor(weight: number): string {
+      // Interpolate from light gray (distant) to dark blue (close)
+      // weight is 1-10, normalize to 0-1
+      const normalized = weight / 10;
+      const hue = 220;
+      const saturation = 70;
+      const lightness = 80 - normalized * 40;
+      return `hsl(${hue}, ${saturation}%, ${lightness}%)`;
+    }
+
+    function draw(context: CanvasRenderingContext2D) {
+      const nodes = nodesRef.current;
+      const edges = edgesRef.current;
+
+      context.clearRect(0, 0, width, height);
+
+      // Draw edges
+      for (const edge of edges) {
+        // Weight is 1-10, scale line width accordingly
+        const lineWidth = 1 + (edge.closeness_weight / 10) * 3;
+        context.strokeStyle = getEdgeColor(edge.closeness_weight);
+        context.lineWidth = lineWidth;
+        context.beginPath();
+        context.moveTo(edge.sourceNode.x, edge.sourceNode.y);
+        context.lineTo(edge.targetNode.x, edge.targetNode.y);
+        context.stroke();
+
+        // Draw relationship type label at midpoint
+        const midX = (edge.sourceNode.x + edge.targetNode.x) / 2;
+        const midY = (edge.sourceNode.y + edge.targetNode.y) / 2;
+        context.fillStyle = "#666";
+        context.font = "10px sans-serif";
+        context.textAlign = "center";
+        const typeInfo = RELATIONSHIP_TYPES.find(t => t.value === edge.relationship_type);
+        const label = typeInfo?.displayName || edge.relationship_type;
+        context.fillText(label, midX, midY - 5);
+      }
+
+      // Draw nodes
+      for (const node of nodes) {
+        const isSelected = node === selectedNode;
+        const radius = isSelected ? 25 : 20;
+
+        // Node circle
+        context.beginPath();
+        context.arc(node.x, node.y, radius, 0, Math.PI * 2);
+        context.fillStyle = isSelected ? "#0066cc" : "#fff";
+        context.fill();
+        context.strokeStyle = "#0066cc";
+        context.lineWidth = 2;
+        context.stroke();
+
+        // Node label
+        context.fillStyle = isSelected ? "#fff" : "#333";
+        context.font = "12px sans-serif";
+        context.textAlign = "center";
+        context.textBaseline = "middle";
+        const name = node.name.length > 10 ? node.name.slice(0, 9) + "…" : node.name;
+        context.fillText(name, node.x, node.y);
+      }
+    }
+
+    function animate() {
+      simulate();
+      draw(ctx);
+      animationRef.current = requestAnimationFrame(animate);
+    }
+
+    animate();
+
+    return () => {
+      cancelAnimationFrame(animationRef.current);
+    };
+  }, [data, selectedNode]);
+
+  // Mouse interaction handlers
+  useEffect(() => {
+    const canvas = canvasRef.current;
+    if (!canvas) return;
+
+    function getNodeAtPosition(x: number, y: number): SimNode | null {
+      for (const node of nodesRef.current) {
+        const dx = x - node.x;
+        const dy = y - node.y;
+        if (dx * dx + dy * dy < 400) {
+          return node;
+        }
+      }
+      return null;
+    }
+
+    function handleMouseDown(e: MouseEvent) {
+      const rect = canvas!.getBoundingClientRect();
+      const x = e.clientX - rect.left;
+      const y = e.clientY - rect.top;
+      const node = getNodeAtPosition(x, y);
+      if (node) {
+        dragNodeRef.current = node;
+        setSelectedNode(node);
+      }
+    }
+
+    function handleMouseMove(e: MouseEvent) {
+      if (!dragNodeRef.current) return;
+      const rect = canvas!.getBoundingClientRect();
+      dragNodeRef.current.x = e.clientX - rect.left;
+      dragNodeRef.current.y = e.clientY - rect.top;
+    }
+
+    function handleMouseUp() {
+      dragNodeRef.current = null;
+    }
+
+    canvas.addEventListener("mousedown", handleMouseDown);
+    canvas.addEventListener("mousemove", handleMouseMove);
+    canvas.addEventListener("mouseup", handleMouseUp);
+    canvas.addEventListener("mouseleave", handleMouseUp);
+
+    return () => {
+      canvas.removeEventListener("mousedown", handleMouseDown);
+      canvas.removeEventListener("mousemove", handleMouseMove);
+      canvas.removeEventListener("mouseup", handleMouseUp);
+      canvas.removeEventListener("mouseleave", handleMouseUp);
+    };
+  }, []);
+
+  if (loading) return <p>Loading graph...</p>;
+  if (error) return <div className="error">{error}</div>;
+  if (!data) return <p>No data available</p>;
+
+  return (
+    <div className="graph-container">
+      <div className="header">
+        <h1>Relationship Graph</h1>
+      </div>
+      <p className="graph-hint">
+        Drag nodes to reposition. Closer relationships have shorter, darker edges.
+      </p>
+      <canvas
+        ref={canvasRef}
+        width={900}
+        height={600}
+        style={{
+          border: "1px solid var(--color-border)",
+          borderRadius: "8px",
+          background: "var(--color-bg)",
+          cursor: "grab",
+        }}
+      />
+      {selectedNode && (
+        <div className="selected-info">
+          <h3>{selectedNode.name}</h3>
+          {selectedNode.current_job && <p>Job: {selectedNode.current_job}</p>}
+          <a href={`/contacts/${selectedNode.id}`}>View details</a>
+        </div>
+      )}
+      <div className="legend">
+        <h4>Relationship Closeness (1-10)</h4>
+        <div className="legend-items">
+          <div className="legend-item">
+            <span className="legend-line" style={{ background: getEdgeColorCSS(10), height: "4px" }}></span>
+            <span>10 - Very Close (Spouse, Partner)</span>
+          </div>
+          <div className="legend-item">
+            <span className="legend-line" style={{ background: getEdgeColorCSS(7), height: "3px" }}></span>
+            <span>7 - Close (Family, Best Friend)</span>
+          </div>
+          <div className="legend-item">
+            <span className="legend-line" style={{ background: getEdgeColorCSS(4), height: "2px" }}></span>
+            <span>4 - Moderate (Friend, Colleague)</span>
+          </div>
+          <div className="legend-item">
+            <span className="legend-line" style={{ background: getEdgeColorCSS(2), height: "1px" }}></span>
+            <span>2 - Distant (Acquaintance)</span>
+          </div>
+        </div>
+      </div>
+    </div>
+  );
+}
+
+function getEdgeColorCSS(weight: number): string {
+  // weight is 1-10, normalize to 0-1
+  const normalized = weight / 10;
+  const hue = 220;
+  const saturation = 70;
+  const lightness = 80 - normalized * 40;
+  return `hsl(${hue}, ${saturation}%, ${lightness}%)`;
+}

frontend/src/index.css

@@ -0,0 +1,62 @@
+:root {
+  /* Light theme (default) */
+  --color-bg: #f5f5f5;
+  --color-bg-card: #ffffff;
+  --color-bg-hover: #f0f0f0;
+  --color-bg-muted: #f9f9f9;
+  --color-bg-error: #ffe0e0;
+
+  --color-text: #333333;
+  --color-text-muted: #666666;
+  --color-text-error: #cc0000;
+
+  --color-border: #dddddd;
+  --color-border-light: #eeeeee;
+  --color-border-lighter: #f0f0f0;
+
+  --color-primary: #0066cc;
+  --color-primary-hover: #0055aa;
+
+  --color-danger: #cc3333;
+  --color-danger-hover: #aa2222;
+
+  --color-tag-bg: #e0e0e0;
+
+  --shadow: 0 1px 3px rgba(0, 0, 0, 0.1);
+
+  font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, sans-serif;
+  line-height: 1.5;
+  font-weight: 400;
+  color: var(--color-text);
+  background-color: var(--color-bg);
+  font-synthesis: none;
+  text-rendering: optimizeLegibility;
+  -webkit-font-smoothing: antialiased;
+  -moz-osx-font-smoothing: grayscale;
+}
+
+[data-theme="dark"] {
+  --color-bg: #1a1a1a;
+  --color-bg-card: #2d2d2d;
+  --color-bg-hover: #3d3d3d;
+  --color-bg-muted: #252525;
+  --color-bg-error: #4a2020;
+
+  --color-text: #e0e0e0;
+  --color-text-muted: #a0a0a0;
+  --color-text-error: #ff6b6b;
+
+  --color-border: #404040;
+  --color-border-light: #353535;
+  --color-border-lighter: #303030;
+
+  --color-primary: #4da6ff;
+  --color-primary-hover: #7dbfff;
+
+  --color-danger: #ff6b6b;
+  --color-danger-hover: #ff8a8a;
+
+  --color-tag-bg: #404040;
+
+  --shadow: 0 1px 3px rgba(0, 0, 0, 0.3);
+}

frontend/src/main.tsx

@@ -0,0 +1,13 @@
+import { StrictMode } from "react";
+import { createRoot } from "react-dom/client";
+import { BrowserRouter } from "react-router-dom";
+import App from "./App.tsx";
+import "./index.css";
+
+createRoot(document.getElementById("root")!).render(
+  <StrictMode>
+    <BrowserRouter>
+      <App />
+    </BrowserRouter>
+  </StrictMode>
+);

frontend/src/types/index.ts

@@ -0,0 +1,155 @@
+export interface Need {
+  id: number;
+  name: string;
+  description: string | null;
+}
+
+export interface NeedCreate {
+  name: string;
+  description?: string | null;
+}
+
+export const RELATIONSHIP_TYPES = [
+  { value: 'spouse', displayName: 'Spouse', defaultWeight: 10 },
+  { value: 'partner', displayName: 'Partner', defaultWeight: 10 },
+  { value: 'parent', displayName: 'Parent', defaultWeight: 9 },
+  { value: 'child', displayName: 'Child', defaultWeight: 9 },
+  { value: 'sibling', displayName: 'Sibling', defaultWeight: 9 },
+  { value: 'best_friend', displayName: 'Best Friend', defaultWeight: 8 },
+  { value: 'grandparent', displayName: 'Grandparent', defaultWeight: 7 },
+  { value: 'grandchild', displayName: 'Grandchild', defaultWeight: 7 },
+  { value: 'aunt_uncle', displayName: 'Aunt/Uncle', defaultWeight: 7 },
+  { value: 'niece_nephew', displayName: 'Niece/Nephew', defaultWeight: 7 },
+  { value: 'cousin', displayName: 'Cousin', defaultWeight: 7 },
+  { value: 'in_law', displayName: 'In-Law', defaultWeight: 7 },
+  { value: 'close_friend', displayName: 'Close Friend', defaultWeight: 6 },
+  { value: 'friend', displayName: 'Friend', defaultWeight: 6 },
+  { value: 'mentor', displayName: 'Mentor', defaultWeight: 5 },
+  { value: 'mentee', displayName: 'Mentee', defaultWeight: 5 },
+  { value: 'business_partner', displayName: 'Business Partner', defaultWeight: 5 },
+  { value: 'colleague', displayName: 'Colleague', defaultWeight: 4 },
+  { value: 'manager', displayName: 'Manager', defaultWeight: 4 },
+  { value: 'direct_report', displayName: 'Direct Report', defaultWeight: 4 },
+  { value: 'client', displayName: 'Client', defaultWeight: 4 },
+  { value: 'acquaintance', displayName: 'Acquaintance', defaultWeight: 3 },
+  { value: 'neighbor', displayName: 'Neighbor', defaultWeight: 3 },
+  { value: 'ex', displayName: 'Ex', defaultWeight: 2 },
+  { value: 'other', displayName: 'Other', defaultWeight: 2 },
+] as const;
+
+export type RelationshipTypeValue = typeof RELATIONSHIP_TYPES[number]['value'];
+
+export interface ContactRelationship {
+  contact_id: number;
+  related_contact_id: number;
+  relationship_type: string;
+  closeness_weight: number;
+}
+
+export interface ContactRelationshipCreate {
+  related_contact_id: number;
+  relationship_type: RelationshipTypeValue;
+  closeness_weight?: number;
+}
+
+export interface ContactRelationshipUpdate {
+  relationship_type?: RelationshipTypeValue;
+  closeness_weight?: number;
+}
+
+export interface GraphNode {
+  id: number;
+  name: string;
+  current_job: string | null;
+}
+
+export interface GraphEdge {
+  source: number;
+  target: number;
+  relationship_type: string;
+  closeness_weight: number;
+}
+
+export interface GraphData {
+  nodes: GraphNode[];
+  edges: GraphEdge[];
+}
+
+export interface Contact {
+  id: number;
+  name: string;
+  age: number | null;
+  bio: string | null;
+  current_job: string | null;
+  gender: string | null;
+  goals: string | null;
+  legal_name: string | null;
+  profile_pic: string | null;
+  safe_conversation_starters: string | null;
+  self_sufficiency_score: number | null;
+  social_structure_style: string | null;
+  ssn: string | null;
+  suffix: string | null;
+  timezone: string | null;
+  topics_to_avoid: string | null;
+  needs: Need[];
+  related_to: ContactRelationship[];
+  related_from: ContactRelationship[];
+}
+
+export interface ContactListItem {
+  id: number;
+  name: string;
+  age: number | null;
+  bio: string | null;
+  current_job: string | null;
+  gender: string | null;
+  goals: string | null;
+  legal_name: string | null;
+  profile_pic: string | null;
+  safe_conversation_starters: string | null;
+  self_sufficiency_score: number | null;
+  social_structure_style: string | null;
+  ssn: string | null;
+  suffix: string | null;
+  timezone: string | null;
+  topics_to_avoid: string | null;
+}
+
+export interface ContactCreate {
+  name: string;
+  age?: number | null;
+  bio?: string | null;
+  current_job?: string | null;
+  gender?: string | null;
+  goals?: string | null;
+  legal_name?: string | null;
+  profile_pic?: string | null;
+  safe_conversation_starters?: string | null;
+  self_sufficiency_score?: number | null;
+  social_structure_style?: string | null;
+  ssn?: string | null;
+  suffix?: string | null;
+  timezone?: string | null;
+  topics_to_avoid?: string | null;
+  need_ids?: number[];
+}
+
+export interface ContactUpdate {
+  name?: string | null;
+  age?: number | null;
+  bio?: string | null;
+  current_job?: string | null;
+  gender?: string | null;
+  goals?: string | null;
+  legal_name?: string | null;
+  profile_pic?: string | null;
+  safe_conversation_starters?: string | null;
+  self_sufficiency_score?: number | null;
+  social_structure_style?: string | null;
+  ssn?: string | null;
+  suffix?: string | null;
+  timezone?: string | null;
+  topics_to_avoid?: string | null;
+  need_ids?: number[] | null;
+}

frontend/tsconfig.app.json

@@ -0,0 +1,28 @@
+{
+  "compilerOptions": {
+    "tsBuildInfoFile": "./node_modules/.tmp/tsconfig.app.tsbuildinfo",
+    "target": "ES2022",
+    "useDefineForClassFields": true,
+    "lib": ["ES2022", "DOM", "DOM.Iterable"],
+    "module": "ESNext",
+    "types": ["vite/client"],
+    "skipLibCheck": true,
+
+    /* Bundler mode */
+    "moduleResolution": "bundler",
+    "allowImportingTsExtensions": true,
+    "verbatimModuleSyntax": true,
+    "moduleDetection": "force",
+    "noEmit": true,
+    "jsx": "react-jsx",
+
+    /* Linting */
+    "strict": true,
+    "noUnusedLocals": true,
+    "noUnusedParameters": true,
+    "erasableSyntaxOnly": true,
+    "noFallthroughCasesInSwitch": true,
+    "noUncheckedSideEffectImports": true
+  },
+  "include": ["src"]
+}

frontend/tsconfig.json

@@ -0,0 +1,7 @@
+{
+  "files": [],
+  "references": [
+    { "path": "./tsconfig.app.json" },
+    { "path": "./tsconfig.node.json" }
+  ]
+}

frontend/tsconfig.node.json

@@ -0,0 +1,26 @@
+{
+  "compilerOptions": {
+    "tsBuildInfoFile": "./node_modules/.tmp/tsconfig.node.tsbuildinfo",
+    "target": "ES2023",
+    "lib": ["ES2023"],
+    "module": "ESNext",
+    "types": ["node"],
+    "skipLibCheck": true,
+
+    /* Bundler mode */
+    "moduleResolution": "bundler",
+    "allowImportingTsExtensions": true,
+    "verbatimModuleSyntax": true,
+    "moduleDetection": "force",
+    "noEmit": true,
+
+    /* Linting */
+    "strict": true,
+    "noUnusedLocals": true,
+    "noUnusedParameters": true,
+    "erasableSyntaxOnly": true,
+    "noFallthroughCasesInSwitch": true,
+    "noUncheckedSideEffectImports": true
+  },
+  "include": ["vite.config.ts"]
+}

frontend/vite.config.ts

@@ -0,0 +1,11 @@
+import { defineConfig } from "vite";
+import react from "@vitejs/plugin-react";
+
+export default defineConfig({
+  plugins: [react()],
+  server: {
+    proxy: {
+      "/api": "http://localhost:8000",
+    },
+  },
+});

overlays/default.nix

@@ -24,9 +24,7 @@
         fastapi
         fastapi-cli
         httpx
-        huggingface-hub
         mypy
-        orjson
         polars
         psycopg
         pydantic
@@ -36,15 +34,16 @@
         pytest-mock
         pytest-xdist
         python-multipart
+        requests
         ruff
         scalene
         sqlalchemy
         sqlalchemy
         tenacity
         textual
-        tiktoken
         tinytuya
         typer
+        types-requests
         websockets
       ]
     );

pyproject.toml

@@ -12,13 +12,13 @@ dependencies = [
     "alembic",
     "apprise",
     "apscheduler",
-    "huggingface-hub",
     "httpx",
     "python-multipart",
     "polars",
     "psycopg[binary]",
     "pydantic",
     "pyyaml",
+    "requests",
     "sqlalchemy",
     "typer",
     "websockets",
@@ -27,11 +27,7 @@ dependencies = [
 [project.scripts]
 database = "python.database_cli:app"
 van-inventory = "python.van_inventory.main:serve"
-prompt-bench = "python.prompt_bench.main:cli"
+sheet-music-ocr = "python.sheet_music_ocr.main:app"
-prompt-bench-download = "python.prompt_bench.downloader:cli"
-finetune = "python.prompt_bench.finetune:cli"
-finetune-container = "python.prompt_bench.finetune_container:cli"
-build-finetune-dataset = "python.prompt_bench.build_finetune_dataset:cli"
 
 [dependency-groups]
 dev = [
@@ -42,6 +38,7 @@ dev = [
     "pytest-xdist",
     "pytest",
     "ruff",
+    "types-requests",
 ]
 
 [tool.ruff]
@@ -87,11 +84,6 @@ lint.ignore = [
 "python/eval_warnings/**" = [
     "S607", # (perm) gh and git are expected on PATH in the runner environment
 ]
-"python/prompt_bench/**" = [
-    "FBT002",  # (perm) typer requires boolean defaults for --flag/--no-flag options
-    "PLR0913", # (perm) typer CLIs naturally have many parameters
-    "S607",    # (perm) docker and nvidia-smi are expected on PATH
-]
 "python/alembic/**" = [
     "INP001", # (perm) this creates LSP issues for alembic
 ]
@@ -121,5 +113,4 @@ exclude_lines = [
 
 [tool.pytest.ini_options]
 addopts = "-n auto -ra"
-testpaths = ["tests"]
 # --cov=system_tools --cov-report=term-missing --cov-report=xml --cov-report=html --cov-branch

python/alembic/data_science_dev/versions/2026_03_24-adding_failedingestion_2f43120e3ffc.py

@@ -1,50 +0,0 @@
-"""adding FailedIngestion.
-
-Revision ID: 2f43120e3ffc
-Revises: f99be864fe69
-Create Date: 2026-03-24 23:46:17.277897
-
-"""
-
-from __future__ import annotations
-
-from typing import TYPE_CHECKING
-
-import sqlalchemy as sa
-from alembic import op
-
-from python.orm import DataScienceDevBase
-
-if TYPE_CHECKING:
-    from collections.abc import Sequence
-
-# revision identifiers, used by Alembic.
-revision: str = "2f43120e3ffc"
-down_revision: str | None = "f99be864fe69"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-schema = DataScienceDevBase.schema_name
-
-
-def upgrade() -> None:
-    """Upgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.create_table(
-        "failed_ingestion",
-        sa.Column("raw_line", sa.Text(), nullable=False),
-        sa.Column("error", sa.Text(), nullable=False),
-        sa.Column("id", sa.Integer(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_failed_ingestion")),
-        schema=schema,
-    )
-    # ### end Alembic commands ###
-
-
-def downgrade() -> None:
-    """Downgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.drop_table("failed_ingestion", schema=schema)
-    # ### end Alembic commands ###

python/alembic/data_science_dev/versions/2026_03_25-attach_partitions_to_posts.py

@@ -1,72 +0,0 @@
-"""Attach all partition tables to the posts parent table.
-
-Alembic autogenerate creates partition tables as standalone tables but does not
-emit the ALTER TABLE ... ATTACH PARTITION statements needed for PostgreSQL to
-route inserts to the correct partition.
-
-Revision ID: a1b2c3d4e5f6
-Revises: 605b1794838f
-Create Date: 2026-03-25 10:00:00.000000
-
-"""
-
-from __future__ import annotations
-
-from typing import TYPE_CHECKING
-
-from alembic import op
-from sqlalchemy import text
-
-from python.orm import DataScienceDevBase
-from python.orm.data_science_dev.posts.partitions import (
-    PARTITION_END_YEAR,
-    PARTITION_START_YEAR,
-    iso_weeks_in_year,
-    week_bounds,
-)
-
-if TYPE_CHECKING:
-    from collections.abc import Sequence
-
-# revision identifiers, used by Alembic.
-revision: str = "a1b2c3d4e5f6"
-down_revision: str | None = "605b1794838f"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-schema = DataScienceDevBase.schema_name
-
-ALREADY_ATTACHED_QUERY = text("""
-    SELECT inhrelid::regclass::text
-    FROM pg_inherits
-    WHERE inhparent = :parent::regclass
-""")
-
-
-def upgrade() -> None:
-    """Attach all weekly partition tables to the posts parent table."""
-    connection = op.get_bind()
-    already_attached = {row[0] for row in connection.execute(ALREADY_ATTACHED_QUERY, {"parent": f"{schema}.posts"})}
-
-    for year in range(PARTITION_START_YEAR, PARTITION_END_YEAR + 1):
-        for week in range(1, iso_weeks_in_year(year) + 1):
-            table_name = f"posts_{year}_{week:02d}"
-            qualified_name = f"{schema}.{table_name}"
-            if qualified_name in already_attached:
-                continue
-            start, end = week_bounds(year, week)
-            start_str = start.strftime("%Y-%m-%d %H:%M:%S")
-            end_str = end.strftime("%Y-%m-%d %H:%M:%S")
-            op.execute(
-                f"ALTER TABLE {schema}.posts "
-                f"ATTACH PARTITION {qualified_name} "
-                f"FOR VALUES FROM ('{start_str}') TO ('{end_str}')"
-            )
-
-
-def downgrade() -> None:
-    """Detach all weekly partition tables from the posts parent table."""
-    for year in range(PARTITION_START_YEAR, PARTITION_END_YEAR + 1):
-        for week in range(1, iso_weeks_in_year(year) + 1):
-            table_name = f"posts_{year}_{week:02d}"
-            op.execute(f"ALTER TABLE {schema}.posts DETACH PARTITION {schema}.{table_name}")

python/alembic/data_science_dev/versions/2026_03_27-adding_congress_data_83bfc8af92d8.py

@@ -1,153 +0,0 @@
-"""adding congress data.
-
-Revision ID: 83bfc8af92d8
-Revises: a1b2c3d4e5f6
-Create Date: 2026-03-27 10:43:02.324510
-
-"""
-
-from __future__ import annotations
-
-from typing import TYPE_CHECKING
-
-import sqlalchemy as sa
-from alembic import op
-
-from python.orm import DataScienceDevBase
-
-if TYPE_CHECKING:
-    from collections.abc import Sequence
-
-# revision identifiers, used by Alembic.
-revision: str = "83bfc8af92d8"
-down_revision: str | None = "a1b2c3d4e5f6"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-schema = DataScienceDevBase.schema_name
-
-
-def upgrade() -> None:
-    """Upgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.create_table(
-        "bill",
-        sa.Column("congress", sa.Integer(), nullable=False),
-        sa.Column("bill_type", sa.String(), nullable=False),
-        sa.Column("number", sa.Integer(), nullable=False),
-        sa.Column("title", sa.String(), nullable=True),
-        sa.Column("title_short", sa.String(), nullable=True),
-        sa.Column("official_title", sa.String(), nullable=True),
-        sa.Column("status", sa.String(), nullable=True),
-        sa.Column("status_at", sa.Date(), nullable=True),
-        sa.Column("sponsor_bioguide_id", sa.String(), nullable=True),
-        sa.Column("subjects_top_term", sa.String(), nullable=True),
-        sa.Column("id", sa.Integer(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_bill")),
-        sa.UniqueConstraint("congress", "bill_type", "number", name="uq_bill_congress_type_number"),
-        schema=schema,
-    )
-    op.create_index("ix_bill_congress", "bill", ["congress"], unique=False, schema=schema)
-    op.create_table(
-        "legislator",
-        sa.Column("bioguide_id", sa.Text(), nullable=False),
-        sa.Column("thomas_id", sa.String(), nullable=True),
-        sa.Column("lis_id", sa.String(), nullable=True),
-        sa.Column("govtrack_id", sa.Integer(), nullable=True),
-        sa.Column("opensecrets_id", sa.String(), nullable=True),
-        sa.Column("fec_ids", sa.String(), nullable=True),
-        sa.Column("first_name", sa.String(), nullable=False),
-        sa.Column("last_name", sa.String(), nullable=False),
-        sa.Column("official_full_name", sa.String(), nullable=True),
-        sa.Column("nickname", sa.String(), nullable=True),
-        sa.Column("birthday", sa.Date(), nullable=True),
-        sa.Column("gender", sa.String(), nullable=True),
-        sa.Column("current_party", sa.String(), nullable=True),
-        sa.Column("current_state", sa.String(), nullable=True),
-        sa.Column("current_district", sa.Integer(), nullable=True),
-        sa.Column("current_chamber", sa.String(), nullable=True),
-        sa.Column("id", sa.Integer(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_legislator")),
-        schema=schema,
-    )
-    op.create_index(op.f("ix_legislator_bioguide_id"), "legislator", ["bioguide_id"], unique=True, schema=schema)
-    op.create_table(
-        "bill_text",
-        sa.Column("bill_id", sa.Integer(), nullable=False),
-        sa.Column("version_code", sa.String(), nullable=False),
-        sa.Column("version_name", sa.String(), nullable=True),
-        sa.Column("text_content", sa.String(), nullable=True),
-        sa.Column("date", sa.Date(), nullable=True),
-        sa.Column("id", sa.Integer(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.ForeignKeyConstraint(
-            ["bill_id"], [f"{schema}.bill.id"], name=op.f("fk_bill_text_bill_id_bill"), ondelete="CASCADE"
-        ),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_bill_text")),
-        sa.UniqueConstraint("bill_id", "version_code", name="uq_bill_text_bill_id_version_code"),
-        schema=schema,
-    )
-    op.create_table(
-        "vote",
-        sa.Column("congress", sa.Integer(), nullable=False),
-        sa.Column("chamber", sa.String(), nullable=False),
-        sa.Column("session", sa.Integer(), nullable=False),
-        sa.Column("number", sa.Integer(), nullable=False),
-        sa.Column("vote_type", sa.String(), nullable=True),
-        sa.Column("question", sa.String(), nullable=True),
-        sa.Column("result", sa.String(), nullable=True),
-        sa.Column("result_text", sa.String(), nullable=True),
-        sa.Column("vote_date", sa.Date(), nullable=False),
-        sa.Column("yea_count", sa.Integer(), nullable=True),
-        sa.Column("nay_count", sa.Integer(), nullable=True),
-        sa.Column("not_voting_count", sa.Integer(), nullable=True),
-        sa.Column("present_count", sa.Integer(), nullable=True),
-        sa.Column("bill_id", sa.Integer(), nullable=True),
-        sa.Column("id", sa.Integer(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.ForeignKeyConstraint(["bill_id"], [f"{schema}.bill.id"], name=op.f("fk_vote_bill_id_bill")),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_vote")),
-        sa.UniqueConstraint("congress", "chamber", "session", "number", name="uq_vote_congress_chamber_session_number"),
-        schema=schema,
-    )
-    op.create_index("ix_vote_congress_chamber", "vote", ["congress", "chamber"], unique=False, schema=schema)
-    op.create_index("ix_vote_date", "vote", ["vote_date"], unique=False, schema=schema)
-    op.create_table(
-        "vote_record",
-        sa.Column("vote_id", sa.Integer(), nullable=False),
-        sa.Column("legislator_id", sa.Integer(), nullable=False),
-        sa.Column("position", sa.String(), nullable=False),
-        sa.ForeignKeyConstraint(
-            ["legislator_id"],
-            [f"{schema}.legislator.id"],
-            name=op.f("fk_vote_record_legislator_id_legislator"),
-            ondelete="CASCADE",
-        ),
-        sa.ForeignKeyConstraint(
-            ["vote_id"], [f"{schema}.vote.id"], name=op.f("fk_vote_record_vote_id_vote"), ondelete="CASCADE"
-        ),
-        sa.PrimaryKeyConstraint("vote_id", "legislator_id", name=op.f("pk_vote_record")),
-        schema=schema,
-    )
-    # ### end Alembic commands ###
-
-
-def downgrade() -> None:
-    """Downgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.drop_table("vote_record", schema=schema)
-    op.drop_index("ix_vote_date", table_name="vote", schema=schema)
-    op.drop_index("ix_vote_congress_chamber", table_name="vote", schema=schema)
-    op.drop_table("vote", schema=schema)
-    op.drop_table("bill_text", schema=schema)
-    op.drop_index(op.f("ix_legislator_bioguide_id"), table_name="legislator", schema=schema)
-    op.drop_table("legislator", schema=schema)
-    op.drop_index("ix_bill_congress", table_name="bill", schema=schema)
-    op.drop_table("bill", schema=schema)
-    # ### end Alembic commands ###

python/alembic/data_science_dev/versions/2026_03_29-adding_legislatorsocialmedia_5cd7eee3549d.py

@@ -1,58 +0,0 @@
-"""adding LegislatorSocialMedia.
-
-Revision ID: 5cd7eee3549d
-Revises: 83bfc8af92d8
-Create Date: 2026-03-29 11:53:44.224799
-
-"""
-
-from __future__ import annotations
-
-from typing import TYPE_CHECKING
-
-import sqlalchemy as sa
-from alembic import op
-
-from python.orm import DataScienceDevBase
-
-if TYPE_CHECKING:
-    from collections.abc import Sequence
-
-# revision identifiers, used by Alembic.
-revision: str = "5cd7eee3549d"
-down_revision: str | None = "83bfc8af92d8"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-schema = DataScienceDevBase.schema_name
-
-
-def upgrade() -> None:
-    """Upgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.create_table(
-        "legislator_social_media",
-        sa.Column("legislator_id", sa.Integer(), nullable=False),
-        sa.Column("platform", sa.String(), nullable=False),
-        sa.Column("account_name", sa.String(), nullable=False),
-        sa.Column("url", sa.String(), nullable=True),
-        sa.Column("source", sa.String(), nullable=False),
-        sa.Column("id", sa.Integer(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.ForeignKeyConstraint(
-            ["legislator_id"],
-            [f"{schema}.legislator.id"],
-            name=op.f("fk_legislator_social_media_legislator_id_legislator"),
-        ),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_legislator_social_media")),
-        schema=schema,
-    )
-    # ### end Alembic commands ###
-
-
-def downgrade() -> None:
-    """Downgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.drop_table("legislator_social_media", schema=schema)
-    # ### end Alembic commands ###

python/alembic/env.py

@@ -81,7 +81,6 @@ def include_name(
 
     """
     if type_ == "schema":
-        # allows a database with multiple schemas to have separate alembic revisions
         return name == target_metadata.schema
     return True
 

python/alembic/richie/versions/2026_03_16-adding_roles_to_signal_devices_2ef7ba690159.py

@@ -1,66 +0,0 @@
-"""adding roles to signal devices.
-
-Revision ID: 2ef7ba690159
-Revises: a1b2c3d4e5f6
-Create Date: 2026-03-16 19:22:38.020350
-
-"""
-
-from __future__ import annotations
-
-from typing import TYPE_CHECKING
-
-import sqlalchemy as sa
-from alembic import op
-
-from python.orm import RichieBase
-
-if TYPE_CHECKING:
-    from collections.abc import Sequence
-
-# revision identifiers, used by Alembic.
-revision: str = "2ef7ba690159"
-down_revision: str | None = "a1b2c3d4e5f6"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-schema = RichieBase.schema_name
-
-
-def upgrade() -> None:
-    """Upgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.create_table(
-        "role",
-        sa.Column("name", sa.String(length=50), nullable=False),
-        sa.Column("id", sa.SmallInteger(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_role")),
-        sa.UniqueConstraint("name", name=op.f("uq_role_name")),
-        schema=schema,
-    )
-    op.create_table(
-        "device_role",
-        sa.Column("device_id", sa.Integer(), nullable=False),
-        sa.Column("role_id", sa.SmallInteger(), nullable=False),
-        sa.Column("id", sa.Integer(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.ForeignKeyConstraint(
-            ["device_id"], [f"{schema}.signal_device.id"], name=op.f("fk_device_role_device_id_signal_device")
-        ),
-        sa.ForeignKeyConstraint(["role_id"], [f"{schema}.role.id"], name=op.f("fk_device_role_role_id_role")),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_device_role")),
-        sa.UniqueConstraint("device_id", "role_id", name="uq_device_role_device_role"),
-        schema=schema,
-    )
-    # ### end Alembic commands ###
-
-
-def downgrade() -> None:
-    """Downgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.drop_table("device_role", schema=schema)
-    op.drop_table("role", schema=schema)
-    # ### end Alembic commands ###

python/alembic/richie/versions/2026_03_18-seprating_signal_bot_database_6b275323f435.py

@@ -1,171 +0,0 @@
-"""seprating signal_bot database.
-
-Revision ID: 6b275323f435
-Revises: 2ef7ba690159
-Create Date: 2026-03-18 08:34:28.785885
-
-"""
-
-from __future__ import annotations
-
-from typing import TYPE_CHECKING
-
-import sqlalchemy as sa
-from alembic import op
-from sqlalchemy.dialects import postgresql
-
-from python.orm import RichieBase
-
-if TYPE_CHECKING:
-    from collections.abc import Sequence
-
-# revision identifiers, used by Alembic.
-revision: str = "6b275323f435"
-down_revision: str | None = "2ef7ba690159"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-schema = RichieBase.schema_name
-
-
-def upgrade() -> None:
-    """Upgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.drop_table("device_role", schema=schema)
-    op.drop_table("signal_device", schema=schema)
-    op.drop_table("role", schema=schema)
-    op.drop_table("dead_letter_message", schema=schema)
-    # ### end Alembic commands ###
-
-
-def downgrade() -> None:
-    """Downgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.create_table(
-        "dead_letter_message",
-        sa.Column("source", sa.VARCHAR(), autoincrement=False, nullable=False),
-        sa.Column("message", sa.TEXT(), autoincrement=False, nullable=False),
-        sa.Column("received_at", postgresql.TIMESTAMP(timezone=True), autoincrement=False, nullable=False),
-        sa.Column(
-            "status",
-            postgresql.ENUM("UNPROCESSED", "PROCESSED", name="message_status", schema=schema),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.Column("id", sa.INTEGER(), autoincrement=True, nullable=False),
-        sa.Column(
-            "created",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.Column(
-            "updated",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_dead_letter_message")),
-        schema=schema,
-    )
-    op.create_table(
-        "role",
-        sa.Column("name", sa.VARCHAR(length=50), autoincrement=False, nullable=False),
-        sa.Column(
-            "id",
-            sa.SMALLINT(),
-            server_default=sa.text(f"nextval('{schema}.role_id_seq'::regclass)"),
-            autoincrement=True,
-            nullable=False,
-        ),
-        sa.Column(
-            "created",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.Column(
-            "updated",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_role")),
-        sa.UniqueConstraint(
-            "name", name=op.f("uq_role_name"), postgresql_include=[], postgresql_nulls_not_distinct=False
-        ),
-        schema=schema,
-    )
-    op.create_table(
-        "signal_device",
-        sa.Column("phone_number", sa.VARCHAR(length=50), autoincrement=False, nullable=False),
-        sa.Column("safety_number", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column(
-            "trust_level",
-            postgresql.ENUM("VERIFIED", "UNVERIFIED", "BLOCKED", name="trust_level", schema=schema),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.Column("last_seen", postgresql.TIMESTAMP(timezone=True), autoincrement=False, nullable=False),
-        sa.Column("id", sa.INTEGER(), autoincrement=True, nullable=False),
-        sa.Column(
-            "created",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.Column(
-            "updated",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_signal_device")),
-        sa.UniqueConstraint(
-            "phone_number",
-            name=op.f("uq_signal_device_phone_number"),
-            postgresql_include=[],
-            postgresql_nulls_not_distinct=False,
-        ),
-        schema=schema,
-    )
-    op.create_table(
-        "device_role",
-        sa.Column("device_id", sa.INTEGER(), autoincrement=False, nullable=False),
-        sa.Column("role_id", sa.SMALLINT(), autoincrement=False, nullable=False),
-        sa.Column("id", sa.INTEGER(), autoincrement=True, nullable=False),
-        sa.Column(
-            "created",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.Column(
-            "updated",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.ForeignKeyConstraint(
-            ["device_id"], [f"{schema}.signal_device.id"], name=op.f("fk_device_role_device_id_signal_device")
-        ),
-        sa.ForeignKeyConstraint(["role_id"], [f"{schema}.role.id"], name=op.f("fk_device_role_role_id_role")),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_device_role")),
-        sa.UniqueConstraint(
-            "device_id",
-            "role_id",
-            name=op.f("uq_device_role_device_role"),
-            postgresql_include=[],
-            postgresql_nulls_not_distinct=False,
-        ),
-        schema=schema,
-    )
-    # ### end Alembic commands ###

python/alembic/richie/versions/2026_03_29-removed_ds_table_from_richie_db_c8a794340928.py

@@ -1,187 +0,0 @@
-"""removed ds table from richie DB.
-
-Revision ID: c8a794340928
-Revises: 6b275323f435
-Create Date: 2026-03-29 15:29:23.643146
-
-"""
-
-from __future__ import annotations
-
-from typing import TYPE_CHECKING
-
-import sqlalchemy as sa
-from alembic import op
-from sqlalchemy.dialects import postgresql
-
-from python.orm import RichieBase
-
-if TYPE_CHECKING:
-    from collections.abc import Sequence
-
-# revision identifiers, used by Alembic.
-revision: str = "c8a794340928"
-down_revision: str | None = "6b275323f435"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-schema = RichieBase.schema_name
-
-
-def upgrade() -> None:
-    """Upgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.drop_table("vote_record", schema=schema)
-    op.drop_index(op.f("ix_vote_congress_chamber"), table_name="vote", schema=schema)
-    op.drop_index(op.f("ix_vote_date"), table_name="vote", schema=schema)
-    op.drop_index(op.f("ix_legislator_bioguide_id"), table_name="legislator", schema=schema)
-    op.drop_table("legislator", schema=schema)
-    op.drop_table("vote", schema=schema)
-    op.drop_index(op.f("ix_bill_congress"), table_name="bill", schema=schema)
-    op.drop_table("bill", schema=schema)
-    # ### end Alembic commands ###
-
-
-def downgrade() -> None:
-    """Downgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.create_table(
-        "vote",
-        sa.Column("congress", sa.INTEGER(), autoincrement=False, nullable=False),
-        sa.Column("chamber", sa.VARCHAR(), autoincrement=False, nullable=False),
-        sa.Column("session", sa.INTEGER(), autoincrement=False, nullable=False),
-        sa.Column("number", sa.INTEGER(), autoincrement=False, nullable=False),
-        sa.Column("vote_type", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("question", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("result", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("result_text", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("vote_date", sa.DATE(), autoincrement=False, nullable=False),
-        sa.Column("yea_count", sa.INTEGER(), autoincrement=False, nullable=True),
-        sa.Column("nay_count", sa.INTEGER(), autoincrement=False, nullable=True),
-        sa.Column("not_voting_count", sa.INTEGER(), autoincrement=False, nullable=True),
-        sa.Column("present_count", sa.INTEGER(), autoincrement=False, nullable=True),
-        sa.Column("bill_id", sa.INTEGER(), autoincrement=False, nullable=True),
-        sa.Column("id", sa.INTEGER(), autoincrement=True, nullable=False),
-        sa.Column(
-            "created",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.Column(
-            "updated",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.ForeignKeyConstraint(["bill_id"], [f"{schema}.bill.id"], name=op.f("fk_vote_bill_id_bill")),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_vote")),
-        sa.UniqueConstraint(
-            "congress",
-            "chamber",
-            "session",
-            "number",
-            name=op.f("uq_vote_congress_chamber_session_number"),
-            postgresql_include=[],
-            postgresql_nulls_not_distinct=False,
-        ),
-        schema=schema,
-    )
-    op.create_index(op.f("ix_vote_date"), "vote", ["vote_date"], unique=False, schema=schema)
-    op.create_index(op.f("ix_vote_congress_chamber"), "vote", ["congress", "chamber"], unique=False, schema=schema)
-    op.create_table(
-        "vote_record",
-        sa.Column("vote_id", sa.INTEGER(), autoincrement=False, nullable=False),
-        sa.Column("legislator_id", sa.INTEGER(), autoincrement=False, nullable=False),
-        sa.Column("position", sa.VARCHAR(), autoincrement=False, nullable=False),
-        sa.ForeignKeyConstraint(
-            ["legislator_id"],
-            [f"{schema}.legislator.id"],
-            name=op.f("fk_vote_record_legislator_id_legislator"),
-            ondelete="CASCADE",
-        ),
-        sa.ForeignKeyConstraint(
-            ["vote_id"], [f"{schema}.vote.id"], name=op.f("fk_vote_record_vote_id_vote"), ondelete="CASCADE"
-        ),
-        sa.PrimaryKeyConstraint("vote_id", "legislator_id", name=op.f("pk_vote_record")),
-        schema=schema,
-    )
-    op.create_table(
-        "legislator",
-        sa.Column("bioguide_id", sa.TEXT(), autoincrement=False, nullable=False),
-        sa.Column("thomas_id", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("lis_id", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("govtrack_id", sa.INTEGER(), autoincrement=False, nullable=True),
-        sa.Column("opensecrets_id", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("fec_ids", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("first_name", sa.VARCHAR(), autoincrement=False, nullable=False),
-        sa.Column("last_name", sa.VARCHAR(), autoincrement=False, nullable=False),
-        sa.Column("official_full_name", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("nickname", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("birthday", sa.DATE(), autoincrement=False, nullable=True),
-        sa.Column("gender", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("current_party", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("current_state", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("current_district", sa.INTEGER(), autoincrement=False, nullable=True),
-        sa.Column("current_chamber", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("id", sa.INTEGER(), autoincrement=True, nullable=False),
-        sa.Column(
-            "created",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.Column(
-            "updated",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_legislator")),
-        schema=schema,
-    )
-    op.create_index(op.f("ix_legislator_bioguide_id"), "legislator", ["bioguide_id"], unique=True, schema=schema)
-    op.create_table(
-        "bill",
-        sa.Column("congress", sa.INTEGER(), autoincrement=False, nullable=False),
-        sa.Column("bill_type", sa.VARCHAR(), autoincrement=False, nullable=False),
-        sa.Column("number", sa.INTEGER(), autoincrement=False, nullable=False),
-        sa.Column("title", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("title_short", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("official_title", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("status", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("status_at", sa.DATE(), autoincrement=False, nullable=True),
-        sa.Column("sponsor_bioguide_id", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("subjects_top_term", sa.VARCHAR(), autoincrement=False, nullable=True),
-        sa.Column("id", sa.INTEGER(), autoincrement=True, nullable=False),
-        sa.Column(
-            "created",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.Column(
-            "updated",
-            postgresql.TIMESTAMP(timezone=True),
-            server_default=sa.text("now()"),
-            autoincrement=False,
-            nullable=False,
-        ),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_bill")),
-        sa.UniqueConstraint(
-            "congress",
-            "bill_type",
-            "number",
-            name=op.f("uq_bill_congress_type_number"),
-            postgresql_include=[],
-            postgresql_nulls_not_distinct=False,
-        ),
-        schema=schema,
-    )
-    op.create_index(op.f("ix_bill_congress"), "bill", ["congress"], unique=False, schema=schema)
-    # ### end Alembic commands ###

python/alembic/signal_bot/versions/2026_03_17-seprating_signal_bot_database_6eaf696e07a5.py

@@ -1,100 +0,0 @@
-"""seprating signal_bot database.
-
-Revision ID: 6eaf696e07a5
-Revises:
-Create Date: 2026-03-17 21:35:37.612672
-
-"""
-
-from __future__ import annotations
-
-from typing import TYPE_CHECKING
-
-import sqlalchemy as sa
-from alembic import op
-from sqlalchemy.dialects import postgresql
-
-from python.orm import SignalBotBase
-
-if TYPE_CHECKING:
-    from collections.abc import Sequence
-
-# revision identifiers, used by Alembic.
-revision: str = "6eaf696e07a5"
-down_revision: str | None = None
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-schema = SignalBotBase.schema_name
-
-
-def upgrade() -> None:
-    """Upgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.create_table(
-        "dead_letter_message",
-        sa.Column("source", sa.String(), nullable=False),
-        sa.Column("message", sa.Text(), nullable=False),
-        sa.Column("received_at", sa.DateTime(timezone=True), nullable=False),
-        sa.Column(
-            "status", postgresql.ENUM("UNPROCESSED", "PROCESSED", name="message_status", schema=schema), nullable=False
-        ),
-        sa.Column("id", sa.Integer(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_dead_letter_message")),
-        schema=schema,
-    )
-    op.create_table(
-        "role",
-        sa.Column("name", sa.String(length=50), nullable=False),
-        sa.Column("id", sa.SmallInteger(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_role")),
-        sa.UniqueConstraint("name", name=op.f("uq_role_name")),
-        schema=schema,
-    )
-    op.create_table(
-        "signal_device",
-        sa.Column("phone_number", sa.String(length=50), nullable=False),
-        sa.Column("safety_number", sa.String(), nullable=True),
-        sa.Column(
-            "trust_level",
-            postgresql.ENUM("VERIFIED", "UNVERIFIED", "BLOCKED", name="trust_level", schema=schema),
-            nullable=False,
-        ),
-        sa.Column("last_seen", sa.DateTime(timezone=True), nullable=False),
-        sa.Column("id", sa.Integer(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_signal_device")),
-        sa.UniqueConstraint("phone_number", name=op.f("uq_signal_device_phone_number")),
-        schema=schema,
-    )
-    op.create_table(
-        "device_role",
-        sa.Column("device_id", sa.Integer(), nullable=False),
-        sa.Column("role_id", sa.SmallInteger(), nullable=False),
-        sa.Column("id", sa.Integer(), nullable=False),
-        sa.Column("created", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.Column("updated", sa.DateTime(timezone=True), server_default=sa.text("now()"), nullable=False),
-        sa.ForeignKeyConstraint(
-            ["device_id"], [f"{schema}.signal_device.id"], name=op.f("fk_device_role_device_id_signal_device")
-        ),
-        sa.ForeignKeyConstraint(["role_id"], [f"{schema}.role.id"], name=op.f("fk_device_role_role_id_role")),
-        sa.PrimaryKeyConstraint("id", name=op.f("pk_device_role")),
-        sa.UniqueConstraint("device_id", "role_id", name="uq_device_role_device_role"),
-        schema=schema,
-    )
-    # ### end Alembic commands ###
-
-
-def downgrade() -> None:
-    """Downgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.drop_table("device_role", schema=schema)
-    op.drop_table("signal_device", schema=schema)
-    op.drop_table("role", schema=schema)
-    op.drop_table("dead_letter_message", schema=schema)
-    # ### end Alembic commands ###

python/alembic/signal_bot/versions/2026_03_18-test_66bdd532bcab.py

@@ -1,72 +0,0 @@
-"""test.
-
-Revision ID: 66bdd532bcab
-Revises: 6eaf696e07a5
-Create Date: 2026-03-18 19:21:14.561568
-
-"""
-
-from __future__ import annotations
-
-from typing import TYPE_CHECKING
-
-import sqlalchemy as sa
-from alembic import op
-from sqlalchemy.dialects import postgresql
-
-from python.orm import SignalBotBase
-
-if TYPE_CHECKING:
-    from collections.abc import Sequence
-
-# revision identifiers, used by Alembic.
-revision: str = "66bdd532bcab"
-down_revision: str | None = "6eaf696e07a5"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-schema = SignalBotBase.schema_name
-
-
-def upgrade() -> None:
-    """Upgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.alter_column(
-        "dead_letter_message",
-        "status",
-        existing_type=postgresql.ENUM("UNPROCESSED", "PROCESSED", name="message_status", schema=schema),
-        type_=sa.Enum("UNPROCESSED", "PROCESSED", name="message_status", native_enum=False),
-        existing_nullable=False,
-        schema=schema,
-    )
-    op.alter_column(
-        "signal_device",
-        "trust_level",
-        existing_type=postgresql.ENUM("VERIFIED", "UNVERIFIED", "BLOCKED", name="trust_level", schema=schema),
-        type_=sa.Enum("VERIFIED", "UNVERIFIED", "BLOCKED", name="trust_level", native_enum=False),
-        existing_nullable=False,
-        schema=schema,
-    )
-    # ### end Alembic commands ###
-
-
-def downgrade() -> None:
-    """Downgrade."""
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.alter_column(
-        "signal_device",
-        "trust_level",
-        existing_type=sa.Enum("VERIFIED", "UNVERIFIED", "BLOCKED", name="trust_level", native_enum=False),
-        type_=postgresql.ENUM("VERIFIED", "UNVERIFIED", "BLOCKED", name="trust_level", schema=schema),
-        existing_nullable=False,
-        schema=schema,
-    )
-    op.alter_column(
-        "dead_letter_message",
-        "status",
-        existing_type=sa.Enum("UNPROCESSED", "PROCESSED", name="message_status", native_enum=False),
-        type_=postgresql.ENUM("UNPROCESSED", "PROCESSED", name="message_status", schema=schema),
-        existing_nullable=False,
-        schema=schema,
-    )
-    # ### end Alembic commands ###

python/api/main.py

@@ -1,23 +1,27 @@
 """FastAPI interface for Contact database."""
 
 import logging
+import shutil
+import subprocess
+import tempfile
 from collections.abc import AsyncIterator
 from contextlib import asynccontextmanager
+from os import environ
+from pathlib import Path
 from typing import Annotated
 
 import typer
 import uvicorn
 from fastapi import FastAPI
 
-from python.api.middleware import ZstdMiddleware
+from python.api.routers import contact_router, create_frontend_router
-from python.api.routers import contact_router, views_router
 from python.common import configure_logger
 from python.orm.common import get_postgres_engine
 
 logger = logging.getLogger(__name__)
 
 
-def create_app() -> FastAPI:
+def create_app(frontend_dir: Path | None = None) -> FastAPI:
     """Create and configure the FastAPI application."""
 
     @asynccontextmanager
@@ -28,23 +32,84 @@ def create_app() -> FastAPI:
         app.state.engine.dispose()
 
     app = FastAPI(title="Contact Database API", lifespan=lifespan)
-    app.add_middleware(ZstdMiddleware)
 
     app.include_router(contact_router)
-    app.include_router(views_router)
+
+    if frontend_dir:
+        logger.info(f"Serving frontend from {frontend_dir}")
+        frontend_router = create_frontend_router(frontend_dir)
+        app.include_router(frontend_router)
 
     return app
 
 
+def build_frontend(source_dir: Path | None, cache_dir: Path | None = None) -> Path | None:
+    """Run npm build and copy output to a temp directory.
+
+    Works even if source_dir is read-only by copying to a temp directory first.
+
+    Args:
+        source_dir: Frontend source directory.
+        cache_dir: Optional npm cache directory for faster repeated builds.
+
+    Returns:
+        Path to frontend build directory, or None if no source_dir provided.
+    """
+    if not source_dir:
+        return None
+
+    if not source_dir.exists():
+        error = f"Frontend directory {source_dir} does not exist"
+        raise FileExistsError(error)
+
+    logger.info("Building frontend from %s...", source_dir)
+
+    # Copy source to a writable temp directory
+    build_dir = Path(tempfile.mkdtemp(prefix="contact_frontend_build_"))
+    shutil.copytree(source_dir, build_dir, dirs_exist_ok=True)
+
+    env = dict(environ)
+    if cache_dir:
+        cache_dir.mkdir(parents=True, exist_ok=True)
+        env["npm_config_cache"] = str(cache_dir)
+
+    subprocess.run(["npm", "install"], cwd=build_dir, env=env, check=True)  # noqa: S607
+    subprocess.run(["npm", "run", "build"], cwd=build_dir, env=env, check=True)  # noqa: S607
+
+    dist_dir = build_dir / "dist"
+    if not dist_dir.exists():
+        error = f"Build output not found at {dist_dir}"
+        raise FileNotFoundError(error)
+
+    output_dir = Path(tempfile.mkdtemp(prefix="contact_frontend_"))
+    shutil.copytree(dist_dir, output_dir, dirs_exist_ok=True)
+    logger.info(f"Frontend built and copied to {output_dir}")
+
+    shutil.rmtree(build_dir)
+
+    return output_dir
+
+
 def serve(
     host: Annotated[str, typer.Option("--host", "-h", help="Host to bind to")],
+    frontend_dir: Annotated[
+        Path | None,
+        typer.Option(
+            "--frontend-dir",
+            "-f",
+            help="Frontend source directory. If provided, runs npm build and serves from temp dir.",
+        ),
+    ] = None,
     port: Annotated[int, typer.Option("--port", "-p", help="Port to bind to")] = 8000,
     log_level: Annotated[str, typer.Option("--log-level", "-l", help="Log level")] = "INFO",
 ) -> None:
     """Start the Contact API server."""
     configure_logger(log_level)
 
-    app = create_app()
+    cache_dir = Path(environ["HOME"]) / ".npm"
+    serve_dir = build_frontend(frontend_dir, cache_dir=cache_dir)
+
+    app = create_app(frontend_dir=serve_dir)
     uvicorn.run(app, host=host, port=port)
 
 

python/api/middleware.py

@@ -1,49 +0,0 @@
-"""Middleware for the FastAPI application."""
-
-from compression import zstd
-from starlette.middleware.base import BaseHTTPMiddleware, RequestResponseEndpoint
-from starlette.requests import Request
-from starlette.responses import Response
-
-MINIMUM_RESPONSE_SIZE = 500
-
-
-class ZstdMiddleware(BaseHTTPMiddleware):
-    """Middleware that compresses responses with zstd when the client supports it."""
-
-    async def dispatch(self, request: Request, call_next: RequestResponseEndpoint) -> Response:
-        """Compress the response with zstd if the client accepts it."""
-        accepted_encodings = request.headers.get("accept-encoding", "")
-        if "zstd" not in accepted_encodings:
-            return await call_next(request)
-
-        response = await call_next(request)
-
-        if response.headers.get("content-encoding") or "text/event-stream" in response.headers.get("content-type", ""):
-            return response
-
-        body = b""
-        async for chunk in response.body_iterator:
-            body += chunk if isinstance(chunk, bytes) else chunk.encode()
-
-        if len(body) < MINIMUM_RESPONSE_SIZE:
-            return Response(
-                content=body,
-                status_code=response.status_code,
-                headers=dict(response.headers),
-                media_type=response.media_type,
-            )
-
-        compressed = zstd.compress(body)
-
-        headers = dict(response.headers)
-        headers["content-encoding"] = "zstd"
-        headers["content-length"] = str(len(compressed))
-        headers.pop("transfer-encoding", None)
-
-        return Response(
-            content=compressed,
-            status_code=response.status_code,
-            headers=headers,
-            media_type=response.media_type,
-        )

python/api/routers/__init__.py

@@ -1,6 +1,6 @@
 """API routers."""
 
 from python.api.routers.contact import router as contact_router
-from python.api.routers.views import router as views_router
+from python.api.routers.frontend import create_frontend_router
 
-__all__ = ["contact_router", "views_router"]
+__all__ = ["contact_router", "create_frontend_router"]

python/api/routers/contact.py

@@ -1,10 +1,6 @@
 """Contact API router."""
 
-from pathlib import Path
+from fastapi import APIRouter, HTTPException
-
-from fastapi import APIRouter, HTTPException, Request
-from fastapi.responses import HTMLResponse
-from fastapi.templating import Jinja2Templates
 from pydantic import BaseModel
 from sqlalchemy import select
 from sqlalchemy.orm import selectinload
@@ -12,14 +8,6 @@ from sqlalchemy.orm import selectinload
 from python.api.dependencies import DbSession
 from python.orm.richie.contact import Contact, ContactRelationship, Need, RelationshipType
 
-TEMPLATES_DIR = Path(__file__).parent.parent / "templates"
-templates = Jinja2Templates(directory=TEMPLATES_DIR)
-
-
-def _is_htmx(request: Request) -> bool:
-    """Check if the request is from HTMX."""
-    return request.headers.get("HX-Request") == "true"
-
 
 class NeedBase(BaseModel):
     """Base schema for Need."""
@@ -192,16 +180,14 @@ def get_need(need_id: int, db: DbSession) -> Need:
     return need
 
 
-@router.delete("/needs/{need_id}", response_model=None)
+@router.delete("/needs/{need_id}")
-def delete_need(need_id: int, request: Request, db: DbSession) -> dict[str, bool] | HTMLResponse:
+def delete_need(need_id: int, db: DbSession) -> dict[str, bool]:
     """Delete a need by ID."""
     need = db.get(Need, need_id)
     if not need:
         raise HTTPException(status_code=404, detail="Need not found")
     db.delete(need)
     db.commit()
-    if _is_htmx(request):
-        return HTMLResponse("")
     return {"deleted": True}
 
 
@@ -275,16 +261,14 @@ def update_contact(
     return db_contact
 
 
-@router.delete("/contacts/{contact_id}", response_model=None)
+@router.delete("/contacts/{contact_id}")
-def delete_contact(contact_id: int, request: Request, db: DbSession) -> dict[str, bool] | HTMLResponse:
+def delete_contact(contact_id: int, db: DbSession) -> dict[str, bool]:
     """Delete a contact by ID."""
     contact = db.get(Contact, contact_id)
     if not contact:
         raise HTTPException(status_code=404, detail="Contact not found")
     db.delete(contact)
     db.commit()
-    if _is_htmx(request):
-        return HTMLResponse("")
     return {"deleted": True}
 
 
@@ -310,13 +294,12 @@ def add_need_to_contact(
     return {"added": True}
 
 
-@router.delete("/contacts/{contact_id}/needs/{need_id}", response_model=None)
+@router.delete("/contacts/{contact_id}/needs/{need_id}")
 def remove_need_from_contact(
     contact_id: int,
     need_id: int,
-    request: Request,
     db: DbSession,
-) -> dict[str, bool] | HTMLResponse:
+) -> dict[str, bool]:
     """Remove a need from a contact."""
     contact = db.get(Contact, contact_id)
     if not contact:
@@ -330,8 +313,6 @@ def remove_need_from_contact(
         contact.needs.remove(need)
         db.commit()
 
-    if _is_htmx(request):
-        return HTMLResponse("")
     return {"removed": True}
 
 
@@ -423,13 +404,12 @@ def update_contact_relationship(
     return relationship
 
 
-@router.delete("/contacts/{contact_id}/relationships/{related_contact_id}", response_model=None)
+@router.delete("/contacts/{contact_id}/relationships/{related_contact_id}")
 def remove_contact_relationship(
     contact_id: int,
     related_contact_id: int,
-    request: Request,
     db: DbSession,
-) -> dict[str, bool] | HTMLResponse:
+) -> dict[str, bool]:
     """Remove a relationship between two contacts."""
     relationship = db.scalar(
         select(ContactRelationship).where(
@@ -442,8 +422,6 @@ def remove_contact_relationship(
 
     db.delete(relationship)
     db.commit()
-    if _is_htmx(request):
-        return HTMLResponse("")
     return {"deleted": True}
 
 

python/api/routers/frontend.py

@@ -0,0 +1,24 @@
+"""Frontend SPA router."""
+
+from pathlib import Path
+
+from fastapi import APIRouter
+from fastapi.responses import FileResponse
+from fastapi.staticfiles import StaticFiles
+
+
+def create_frontend_router(frontend_dir: Path) -> APIRouter:
+    """Create a router for serving the frontend SPA."""
+    router = APIRouter(tags=["frontend"])
+
+    router.mount("/assets", StaticFiles(directory=frontend_dir / "assets"), name="assets")
+
+    @router.get("/{full_path:path}")
+    async def serve_spa(full_path: str) -> FileResponse:
+        """Serve React SPA for all non-API routes."""
+        file_path = frontend_dir / full_path
+        if file_path.is_file():
+            return FileResponse(file_path)
+        return FileResponse(frontend_dir / "index.html")
+
+    return router

python/api/routers/views.py

@@ -1,345 +0,0 @@
-"""HTMX server-rendered view router."""
-
-from pathlib import Path
-from typing import Annotated, Any
-
-from fastapi import APIRouter, Form, HTTPException, Request
-from fastapi.responses import HTMLResponse, RedirectResponse
-from fastapi.templating import Jinja2Templates
-from sqlalchemy import select
-from sqlalchemy.orm import Session, selectinload
-
-from python.api.dependencies import DbSession
-from python.orm.richie.contact import Contact, ContactRelationship, Need, RelationshipType
-
-TEMPLATES_DIR = Path(__file__).parent.parent / "templates"
-templates = Jinja2Templates(directory=TEMPLATES_DIR)
-
-router = APIRouter(tags=["views"])
-
-FAMILIAL_TYPES = {
-    "parent",
-    "child",
-    "sibling",
-    "grandparent",
-    "grandchild",
-    "aunt_uncle",
-    "niece_nephew",
-    "cousin",
-    "in_law",
-}
-FRIEND_TYPES = {"best_friend", "close_friend", "friend", "acquaintance", "neighbor"}
-PARTNER_TYPES = {"spouse", "partner"}
-PROFESSIONAL_TYPES = {"mentor", "mentee", "business_partner", "colleague", "manager", "direct_report", "client"}
-
-CONTACT_STRING_FIELDS = (
-    "name",
-    "legal_name",
-    "suffix",
-    "gender",
-    "current_job",
-    "timezone",
-    "profile_pic",
-    "bio",
-    "goals",
-    "social_structure_style",
-    "safe_conversation_starters",
-    "topics_to_avoid",
-    "ssn",
-)
-
-CONTACT_INT_FIELDS = ("age", "self_sufficiency_score")
-
-
-def _group_relationships(relationships: list[ContactRelationship]) -> dict[str, list[ContactRelationship]]:
-    """Group relationships by category."""
-    groups: dict[str, list[ContactRelationship]] = {
-        "familial": [],
-        "partners": [],
-        "friends": [],
-        "professional": [],
-        "other": [],
-    }
-    for rel in relationships:
-        if rel.relationship_type in FAMILIAL_TYPES:
-            groups["familial"].append(rel)
-        elif rel.relationship_type in PARTNER_TYPES:
-            groups["partners"].append(rel)
-        elif rel.relationship_type in FRIEND_TYPES:
-            groups["friends"].append(rel)
-        elif rel.relationship_type in PROFESSIONAL_TYPES:
-            groups["professional"].append(rel)
-        else:
-            groups["other"].append(rel)
-    return groups
-
-
-def _build_contact_name_map(database: Session, contact: Contact) -> dict[int, str]:
-    """Build a mapping of contact IDs to names for relationship display."""
-    related_ids = {rel.related_contact_id for rel in contact.related_to}
-    related_ids |= {rel.contact_id for rel in contact.related_from}
-    related_ids.discard(contact.id)
-
-    if not related_ids:
-        return {}
-
-    related_contacts = list(database.scalars(select(Contact).where(Contact.id.in_(related_ids))).all())
-    return {related.id: related.name for related in related_contacts}
-
-
-def _get_relationship_type_display() -> dict[str, str]:
-    """Build a mapping of relationship type values to display names."""
-    return {rel_type.value: rel_type.display_name for rel_type in RelationshipType}
-
-
-async def _parse_contact_form(request: Request) -> dict[str, Any]:
-    """Parse contact form data from a multipart/form request."""
-    form_data = await request.form()
-    result: dict[str, Any] = {}
-
-    for field in CONTACT_STRING_FIELDS:
-        value = form_data.get(field, "")
-        result[field] = str(value) if value else None
-
-    for field in CONTACT_INT_FIELDS:
-        value = form_data.get(field, "")
-        result[field] = int(value) if value else None
-
-    result["need_ids"] = [int(value) for value in form_data.getlist("need_ids")]
-    return result
-
-
-def _save_contact_from_form(database: Session, contact: Contact, form_result: dict[str, Any]) -> None:
-    """Apply parsed form data to a Contact and save associated needs."""
-    need_ids = form_result.pop("need_ids")
-
-    for key, value in form_result.items():
-        setattr(contact, key, value)
-
-    if need_ids:
-        contact.needs = list(database.scalars(select(Need).where(Need.id.in_(need_ids))).all())
-    else:
-        contact.needs = []
-
-
-@router.get("/", response_class=HTMLResponse)
-@router.get("/contacts", response_class=HTMLResponse)
-def contact_list_page(request: Request, database: DbSession) -> HTMLResponse:
-    """Render the contacts list page."""
-    contacts = list(database.scalars(select(Contact)).all())
-    return templates.TemplateResponse(request, "contact_list.html", {"contacts": contacts})
-
-
-@router.get("/contacts/new", response_class=HTMLResponse)
-def new_contact_page(request: Request, database: DbSession) -> HTMLResponse:
-    """Render the new contact form page."""
-    all_needs = list(database.scalars(select(Need)).all())
-    return templates.TemplateResponse(request, "contact_form.html", {"contact": None, "all_needs": all_needs})
-
-
-@router.post("/htmx/contacts/new")
-async def create_contact_form(request: Request, database: DbSession) -> RedirectResponse:
-    """Handle the create contact form submission."""
-    form_result = await _parse_contact_form(request)
-    contact = Contact()
-    _save_contact_from_form(database, contact, form_result)
-
-    database.add(contact)
-    database.commit()
-    database.refresh(contact)
-    return RedirectResponse(url=f"/contacts/{contact.id}", status_code=303)
-
-
-@router.get("/contacts/{contact_id}", response_class=HTMLResponse)
-def contact_detail_page(contact_id: int, request: Request, database: DbSession) -> HTMLResponse:
-    """Render the contact detail page."""
-    contact = database.scalar(
-        select(Contact)
-        .where(Contact.id == contact_id)
-        .options(
-            selectinload(Contact.needs),
-            selectinload(Contact.related_to),
-            selectinload(Contact.related_from),
-        )
-    )
-    if not contact:
-        raise HTTPException(status_code=404, detail="Contact not found")
-
-    contact_names = _build_contact_name_map(database, contact)
-    grouped_relationships = _group_relationships(contact.related_to)
-    all_contacts = list(database.scalars(select(Contact)).all())
-    all_needs = list(database.scalars(select(Need)).all())
-    available_needs = [need for need in all_needs if need not in contact.needs]
-
-    return templates.TemplateResponse(
-        request,
-        "contact_detail.html",
-        {
-            "contact": contact,
-            "contact_names": contact_names,
-            "grouped_relationships": grouped_relationships,
-            "all_contacts": all_contacts,
-            "available_needs": available_needs,
-            "relationship_types": list(RelationshipType),
-        },
-    )
-
-
-@router.get("/contacts/{contact_id}/edit", response_class=HTMLResponse)
-def edit_contact_page(contact_id: int, request: Request, database: DbSession) -> HTMLResponse:
-    """Render the edit contact form page."""
-    contact = database.scalar(select(Contact).where(Contact.id == contact_id).options(selectinload(Contact.needs)))
-    if not contact:
-        raise HTTPException(status_code=404, detail="Contact not found")
-
-    all_needs = list(database.scalars(select(Need)).all())
-    return templates.TemplateResponse(request, "contact_form.html", {"contact": contact, "all_needs": all_needs})
-
-
-@router.post("/htmx/contacts/{contact_id}/edit")
-async def update_contact_form(contact_id: int, request: Request, database: DbSession) -> RedirectResponse:
-    """Handle the edit contact form submission."""
-    contact = database.get(Contact, contact_id)
-    if not contact:
-        raise HTTPException(status_code=404, detail="Contact not found")
-
-    form_result = await _parse_contact_form(request)
-    _save_contact_from_form(database, contact, form_result)
-
-    database.commit()
-    return RedirectResponse(url=f"/contacts/{contact_id}", status_code=303)
-
-
-@router.post("/htmx/contacts/{contact_id}/add-need", response_class=HTMLResponse)
-def add_need_to_contact_htmx(
-    contact_id: int,
-    request: Request,
-    database: DbSession,
-    need_id: Annotated[int, Form()],
-) -> HTMLResponse:
-    """Add a need to a contact and return updated manage-needs partial."""
-    contact = database.scalar(select(Contact).where(Contact.id == contact_id).options(selectinload(Contact.needs)))
-    if not contact:
-        raise HTTPException(status_code=404, detail="Contact not found")
-
-    need = database.get(Need, need_id)
-    if not need:
-        raise HTTPException(status_code=404, detail="Need not found")
-
-    if need not in contact.needs:
-        contact.needs.append(need)
-        database.commit()
-        database.refresh(contact)
-
-    return templates.TemplateResponse(request, "partials/manage_needs.html", {"contact": contact})
-
-
-@router.post("/htmx/contacts/{contact_id}/add-relationship", response_class=HTMLResponse)
-def add_relationship_htmx(
-    contact_id: int,
-    request: Request,
-    database: DbSession,
-    related_contact_id: Annotated[int, Form()],
-    relationship_type: Annotated[str, Form()],
-) -> HTMLResponse:
-    """Add a relationship and return updated manage-relationships partial."""
-    contact = database.scalar(select(Contact).where(Contact.id == contact_id).options(selectinload(Contact.related_to)))
-    if not contact:
-        raise HTTPException(status_code=404, detail="Contact not found")
-
-    related_contact = database.get(Contact, related_contact_id)
-    if not related_contact:
-        raise HTTPException(status_code=404, detail="Related contact not found")
-
-    rel_type = RelationshipType(relationship_type)
-    weight = rel_type.default_weight
-
-    relationship = ContactRelationship(
-        contact_id=contact_id,
-        related_contact_id=related_contact_id,
-        relationship_type=relationship_type,
-        closeness_weight=weight,
-    )
-    database.add(relationship)
-    database.commit()
-    database.refresh(contact)
-
-    contact_names = _build_contact_name_map(database, contact)
-    return templates.TemplateResponse(
-        request,
-        "partials/manage_relationships.html",
-        {"contact": contact, "contact_names": contact_names},
-    )
-
-
-@router.post("/htmx/contacts/{contact_id}/relationships/{related_contact_id}/weight")
-def update_relationship_weight_htmx(
-    contact_id: int,
-    related_contact_id: int,
-    database: DbSession,
-    closeness_weight: Annotated[int, Form()],
-) -> HTMLResponse:
-    """Update a relationship's closeness weight from HTMX range input."""
-    relationship = database.scalar(
-        select(ContactRelationship).where(
-            ContactRelationship.contact_id == contact_id,
-            ContactRelationship.related_contact_id == related_contact_id,
-        )
-    )
-    if not relationship:
-        raise HTTPException(status_code=404, detail="Relationship not found")
-
-    relationship.closeness_weight = closeness_weight
-    database.commit()
-    return HTMLResponse("")
-
-
-@router.post("/htmx/needs", response_class=HTMLResponse)
-def create_need_htmx(
-    request: Request,
-    database: DbSession,
-    name: Annotated[str, Form()],
-    description: Annotated[str, Form()] = "",
-) -> HTMLResponse:
-    """Create a need via form data and return updated needs list."""
-    need = Need(name=name, description=description or None)
-    database.add(need)
-    database.commit()
-    needs = list(database.scalars(select(Need)).all())
-    return templates.TemplateResponse(request, "partials/need_items.html", {"needs": needs})
-
-
-@router.get("/needs", response_class=HTMLResponse)
-def needs_page(request: Request, database: DbSession) -> HTMLResponse:
-    """Render the needs list page."""
-    needs = list(database.scalars(select(Need)).all())
-    return templates.TemplateResponse(request, "need_list.html", {"needs": needs})
-
-
-@router.get("/graph", response_class=HTMLResponse)
-def graph_page(request: Request, database: DbSession) -> HTMLResponse:
-    """Render the relationship graph page."""
-    contacts = list(database.scalars(select(Contact)).all())
-    relationships = list(database.scalars(select(ContactRelationship)).all())
-
-    graph_data = {
-        "nodes": [{"id": contact.id, "name": contact.name, "current_job": contact.current_job} for contact in contacts],
-        "edges": [
-            {
-                "source": rel.contact_id,
-                "target": rel.related_contact_id,
-                "relationship_type": rel.relationship_type,
-                "closeness_weight": rel.closeness_weight,
-            }
-            for rel in relationships
-        ],
-    }
-
-    return templates.TemplateResponse(
-        request,
-        "graph.html",
-        {
-            "graph_data": graph_data,
-            "relationship_type_display": _get_relationship_type_display(),
-        },
-    )

python/api/templates/base.html

@@ -1,198 +0,0 @@
-<!DOCTYPE html>
-<html lang="en" data-theme="light">
-<head>
-    <meta charset="UTF-8">
-    <meta name="viewport" content="width=device-width, initial-scale=1.0">
-    <title>{% block title %}Contact Database{% endblock %}</title>
-    <script src="https://unpkg.com/htmx.org@2.0.4"></script>
-    <style>
-        :root {
-            --color-bg: #f5f5f5;
-            --color-bg-card: #ffffff;
-            --color-bg-hover: #f0f0f0;
-            --color-bg-muted: #f9f9f9;
-            --color-bg-error: #ffe0e0;
-            --color-text: #333333;
-            --color-text-muted: #666666;
-            --color-text-error: #cc0000;
-            --color-border: #dddddd;
-            --color-border-light: #eeeeee;
-            --color-border-lighter: #f0f0f0;
-            --color-primary: #0066cc;
-            --color-primary-hover: #0055aa;
-            --color-danger: #cc3333;
-            --color-danger-hover: #aa2222;
-            --color-tag-bg: #e0e0e0;
-            --shadow: 0 1px 3px rgba(0, 0, 0, 0.1);
-            font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, sans-serif;
-            line-height: 1.5;
-            color: var(--color-text);
-            background-color: var(--color-bg);
-        }
-        [data-theme="dark"] {
-            --color-bg: #1a1a1a;
-            --color-bg-card: #2d2d2d;
-            --color-bg-hover: #3d3d3d;
-            --color-bg-muted: #252525;
-            --color-bg-error: #4a2020;
-            --color-text: #e0e0e0;
-            --color-text-muted: #a0a0a0;
-            --color-text-error: #ff6b6b;
-            --color-border: #404040;
-            --color-border-light: #353535;
-            --color-border-lighter: #303030;
-            --color-primary: #4da6ff;
-            --color-primary-hover: #7dbfff;
-            --color-danger: #ff6b6b;
-            --color-danger-hover: #ff8a8a;
-            --color-tag-bg: #404040;
-            --shadow: 0 1px 3px rgba(0, 0, 0, 0.3);
-        }
-        * { box-sizing: border-box; }
-        body { margin: 0; background: var(--color-bg); color: var(--color-text); }
-        .app { max-width: 1000px; margin: 0 auto; padding: 20px; }
-        nav { display: flex; align-items: center; gap: 20px; padding: 15px 0; border-bottom: 1px solid var(--color-border); margin-bottom: 20px; }
-        nav a { color: var(--color-primary); text-decoration: none; font-weight: 500; }
-        nav a:hover { text-decoration: underline; }
-        .theme-toggle { margin-left: auto; }
-        main { background: var(--color-bg-card); padding: 20px; border-radius: 8px; box-shadow: var(--shadow); }
-        .header { display: flex; justify-content: space-between; align-items: center; margin-bottom: 20px; }
-        .header h1 { margin: 0; }
-        a { color: var(--color-primary); }
-        a:hover { text-decoration: underline; }
-
-        .btn { display: inline-block; padding: 8px 16px; border: 1px solid var(--color-border); border-radius: 4px; background: var(--color-bg-card); color: var(--color-text); text-decoration: none; cursor: pointer; font-size: 14px; margin-left: 8px; }
-        .btn:hover { background: var(--color-bg-hover); }
-        .btn-primary { background: var(--color-primary); border-color: var(--color-primary); color: white; }
-        .btn-primary:hover { background: var(--color-primary-hover); }
-        .btn-danger { background: var(--color-danger); border-color: var(--color-danger); color: white; }
-        .btn-danger:hover { background: var(--color-danger-hover); }
-        .btn-small { padding: 4px 8px; font-size: 12px; }
-        .btn:disabled { opacity: 0.6; cursor: not-allowed; }
-
-        table { width: 100%; border-collapse: collapse; }
-        th, td { padding: 12px; text-align: left; border-bottom: 1px solid var(--color-border-light); }
-        th { font-weight: 600; background: var(--color-bg-muted); }
-        tr:hover { background: var(--color-bg-muted); }
-
-        .error { background: var(--color-bg-error); color: var(--color-text-error); padding: 10px; border-radius: 4px; margin-bottom: 20px; }
-        .tag { display: inline-block; background: var(--color-tag-bg); padding: 2px 8px; border-radius: 12px; font-size: 12px; color: var(--color-text-muted); }
-
-        .add-form { display: flex; gap: 10px; margin-top: 15px; flex-wrap: wrap; }
-        .add-form select, .add-form input { padding: 8px; border: 1px solid var(--color-border); border-radius: 4px; min-width: 200px; background: var(--color-bg-card); color: var(--color-text); }
-
-        .form-group { margin-bottom: 20px; }
-        .form-group label { display: block; font-weight: 500; margin-bottom: 5px; }
-        .form-group input, .form-group textarea, .form-group select { width: 100%; padding: 10px; border: 1px solid var(--color-border); border-radius: 4px; font-size: 14px; background: var(--color-bg-card); color: var(--color-text); }
-        .form-group textarea { resize: vertical; }
-        .form-row { display: grid; grid-template-columns: 1fr 1fr; gap: 20px; }
-        .checkbox-group { display: flex; flex-wrap: wrap; gap: 15px; }
-        .checkbox-label { display: flex; align-items: center; gap: 5px; cursor: pointer; }
-        .form-actions { display: flex; gap: 10px; margin-top: 30px; padding-top: 20px; border-top: 1px solid var(--color-border-light); }
-
-        .need-form { background: var(--color-bg-muted); padding: 20px; border-radius: 4px; margin-bottom: 20px; }
-        .need-items { list-style: none; padding: 0; }
-        .need-items li { display: flex; justify-content: space-between; align-items: flex-start; padding: 15px; border: 1px solid var(--color-border-light); border-radius: 4px; margin-bottom: 10px; }
-        .need-info p { margin: 5px 0 0; color: var(--color-text-muted); font-size: 14px; }
-
-        .graph-container { width: 100%; }
-        .graph-hint { color: var(--color-text-muted); font-size: 14px; margin-bottom: 15px; }
-        .selected-info { margin-top: 15px; padding: 15px; background: var(--color-bg-muted); border-radius: 8px; }
-        .selected-info h3 { margin: 0 0 10px; }
-        .selected-info p { margin: 5px 0; color: var(--color-text-muted); }
-        .legend { margin-top: 20px; padding: 15px; background: var(--color-bg-muted); border-radius: 8px; }
-        .legend h4 { margin: 0 0 10px; font-size: 14px; }
-        .legend-items { display: flex; flex-wrap: wrap; gap: 15px; }
-        .legend-item { display: flex; align-items: center; gap: 8px; font-size: 12px; color: var(--color-text-muted); }
-        .legend-line { width: 30px; border-radius: 2px; }
-
-        .id-card { width: 100%; }
-        .id-card-inner { background: linear-gradient(135deg, #0a0a0f 0%, #1a1a2e 50%, #0a0a0f 100%); background-image: radial-gradient(white 1px, transparent 1px), linear-gradient(135deg, #0a0a0f 0%, #1a1a2e 50%, #0a0a0f 100%); background-size: 50px 50px, 100% 100%; color: #fff; border-radius: 12px; padding: 25px; min-height: 500px; position: relative; overflow: hidden; }
-        .id-card-header { display: flex; justify-content: space-between; align-items: flex-start; margin-bottom: 15px; }
-        .id-card-header-left { flex: 1; }
-        .id-card-header-right { display: flex; flex-direction: column; align-items: flex-end; gap: 10px; }
-        .id-card-title { font-size: 2.5rem; font-weight: 700; margin: 0; color: #fff; text-shadow: 2px 2px 4px rgba(0,0,0,0.5); }
-        .id-profile-pic { width: 80px; height: 80px; border-radius: 8px; object-fit: cover; border: 2px solid rgba(255,255,255,0.3); }
-        .id-profile-placeholder { width: 80px; height: 80px; border-radius: 8px; background: linear-gradient(135deg, #4ecdc4 0%, #44a8a0 100%); display: flex; align-items: center; justify-content: center; border: 2px solid rgba(255,255,255,0.3); }
-        .id-profile-placeholder span { font-size: 2rem; font-weight: 700; color: #fff; text-shadow: 1px 1px 2px rgba(0,0,0,0.3); }
-        .id-card-actions { display: flex; gap: 8px; }
-        .id-card-actions .btn { background: rgba(255,255,255,0.1); border-color: rgba(255,255,255,0.3); color: #fff; }
-        .id-card-actions .btn:hover { background: rgba(255,255,255,0.2); }
-        .id-card-body { display: grid; grid-template-columns: 1fr 1.5fr; gap: 30px; }
-        .id-card-left { display: flex; flex-direction: column; gap: 8px; }
-        .id-field { font-size: 1rem; line-height: 1.4; }
-        .id-field-block { margin-top: 15px; font-size: 0.95rem; line-height: 1.5; }
-        .id-label { color: #4ecdc4; font-weight: 500; }
-        .id-card-right { display: flex; flex-direction: column; gap: 20px; }
-        .id-bio { font-size: 0.9rem; line-height: 1.6; color: #e0e0e0; }
-        .id-relationships { margin-top: 10px; }
-        .id-section-title { font-size: 1.5rem; margin: 0 0 15px; color: #fff; border-bottom: 1px solid rgba(255,255,255,0.2); padding-bottom: 8px; }
-        .id-rel-group { margin-bottom: 12px; font-size: 0.9rem; line-height: 1.6; }
-        .id-rel-label { color: #a0a0a0; }
-        .id-rel-group a { color: #4ecdc4; text-decoration: none; }
-        .id-rel-group a:hover { text-decoration: underline; }
-        .id-rel-type { color: #888; font-size: 0.85em; }
-        .id-card-warnings { margin-top: 30px; padding-top: 20px; border-top: 1px solid rgba(255,255,255,0.2); display: flex; flex-wrap: wrap; gap: 20px; }
-        .id-warning { display: flex; align-items: center; gap: 8px; font-size: 0.9rem; color: #ff6b6b; }
-        .warning-dot { width: 8px; height: 8px; background: #ff6b6b; border-radius: 50%; flex-shrink: 0; }
-        .warning-desc { color: #ccc; }
-
-        .id-card-manage { margin-top: 20px; background: var(--color-bg-muted); border-radius: 8px; padding: 15px; }
-        .id-card-manage summary { cursor: pointer; font-weight: 600; font-size: 1.1rem; padding: 5px 0; }
-        .id-card-manage[open] summary { margin-bottom: 15px; border-bottom: 1px solid var(--color-border-light); padding-bottom: 10px; }
-        .manage-section { margin-bottom: 25px; }
-        .manage-section h3 { margin: 0 0 15px; font-size: 1rem; }
-        .manage-relationships { display: flex; flex-direction: column; gap: 10px; margin-bottom: 15px; }
-        .manage-rel-item { display: flex; align-items: center; gap: 12px; padding: 10px; background: var(--color-bg-card); border-radius: 6px; flex-wrap: wrap; }
-        .manage-rel-item a { font-weight: 500; min-width: 120px; }
-        .weight-control { display: flex; align-items: center; gap: 8px; font-size: 12px; color: var(--color-text-muted); }
-        .weight-control input[type="range"] { width: 80px; cursor: pointer; }
-        .weight-value { min-width: 20px; text-align: center; font-weight: 600; }
-        .manage-needs-list { list-style: none; padding: 0; margin: 0 0 15px; }
-        .manage-needs-list li { display: flex; align-items: center; gap: 12px; padding: 10px; background: var(--color-bg-card); border-radius: 6px; margin-bottom: 8px; }
-        .manage-needs-list li .btn { margin-left: auto; }
-
-        .htmx-indicator { display: none; }
-        .htmx-request .htmx-indicator { display: inline; }
-        .htmx-request.htmx-indicator { display: inline; }
-
-        @media (max-width: 768px) {
-            .id-card-body { grid-template-columns: 1fr; }
-            .id-card-title { font-size: 1.8rem; }
-            .id-card-header { flex-direction: column; gap: 15px; }
-        }
-    </style>
-</head>
-<body>
-    <div class="app">
-        <nav>
-            <a href="/contacts">Contacts</a>
-            <a href="/graph">Graph</a>
-            <a href="/needs">Needs</a>
-            <button class="btn btn-small theme-toggle" onclick="toggleTheme()">
-                <span id="theme-label">Dark</span>
-            </button>
-        </nav>
-
-        <main id="main-content">
-            {% block content %}{% endblock %}
-        </main>
-    </div>
-
-    <script>
-        function toggleTheme() {
-            const html = document.documentElement;
-            const current = html.getAttribute('data-theme');
-            const next = current === 'light' ? 'dark' : 'light';
-            html.setAttribute('data-theme', next);
-            localStorage.setItem('theme', next);
-            document.getElementById('theme-label').textContent = next === 'light' ? 'Dark' : 'Light';
-        }
-        (function() {
-            const saved = localStorage.getItem('theme') || 'light';
-            document.documentElement.setAttribute('data-theme', saved);
-            document.getElementById('theme-label').textContent = saved === 'light' ? 'Dark' : 'Light';
-        })();
-    </script>
-</body>
-</html>

python/api/templates/contact_detail.html

@@ -1,204 +0,0 @@
-{% extends "base.html" %}
-{% block title %}{{ contact.name }}{% endblock %}
-{% block content %}
-<div class="id-card">
-    <div class="id-card-inner">
-        <div class="id-card-header">
-            <div class="id-card-header-left">
-                <h1 class="id-card-title">I.D.: {{ contact.name }}</h1>
-            </div>
-            <div class="id-card-header-right">
-                {% if contact.profile_pic %}
-                <img src="{{ contact.profile_pic }}" alt="{{ contact.name }}'s profile" class="id-profile-pic">
-                {% else %}
-                <div class="id-profile-placeholder">
-                    <span>{{ contact.name[0]|upper }}</span>
-                </div>
-                {% endif %}
-                <div class="id-card-actions">
-                    <a href="/contacts/{{ contact.id }}/edit" class="btn btn-small">Edit</a>
-                    <a href="/contacts" class="btn btn-small">Back</a>
-                </div>
-            </div>
-        </div>
-
-        <div class="id-card-body">
-            <div class="id-card-left">
-                {% if contact.legal_name %}
-                <div class="id-field">Legal name: {{ contact.legal_name }}</div>
-                {% endif %}
-                {% if contact.suffix %}
-                <div class="id-field">Suffix: {{ contact.suffix }}</div>
-                {% endif %}
-                {% if contact.gender %}
-                <div class="id-field">Gender: {{ contact.gender }}</div>
-                {% endif %}
-                {% if contact.age %}
-                <div class="id-field">Age: {{ contact.age }}</div>
-                {% endif %}
-                {% if contact.current_job %}
-                <div class="id-field">Job: {{ contact.current_job }}</div>
-                {% endif %}
-                {% if contact.social_structure_style %}
-                <div class="id-field">Social style: {{ contact.social_structure_style }}</div>
-                {% endif %}
-                {% if contact.self_sufficiency_score is not none %}
-                <div class="id-field">Self-Sufficiency: {{ contact.self_sufficiency_score }}</div>
-                {% endif %}
-                {% if contact.timezone %}
-                <div class="id-field">Timezone: {{ contact.timezone }}</div>
-                {% endif %}
-                {% if contact.safe_conversation_starters %}
-                <div class="id-field-block">
-                    <span class="id-label">Safe con starters:</span> {{ contact.safe_conversation_starters }}
-                </div>
-                {% endif %}
-                {% if contact.topics_to_avoid %}
-                <div class="id-field-block">
-                    <span class="id-label">Topics to avoid:</span> {{ contact.topics_to_avoid }}
-                </div>
-                {% endif %}
-                {% if contact.goals %}
-                <div class="id-field-block">
-                    <span class="id-label">Goals:</span> {{ contact.goals }}
-                </div>
-                {% endif %}
-            </div>
-
-            <div class="id-card-right">
-                {% if contact.bio %}
-                <div class="id-bio">
-                    <span class="id-label">Bio:</span> {{ contact.bio }}
-                </div>
-                {% endif %}
-
-                <div class="id-relationships">
-                    <h2 class="id-section-title">Relationships</h2>
-
-                    {% if grouped_relationships.familial %}
-                    <div class="id-rel-group">
-                        <span class="id-rel-label">Familial:</span>
-                        {% for rel in grouped_relationships.familial %}
-                        <a href="/contacts/{{ rel.related_contact_id }}">{{ contact_names[rel.related_contact_id] }}</a><span class="id-rel-type">({{ rel.relationship_type|replace("_", " ")|title }})</span>{% if not loop.last %}, {% endif %}
-                        {% endfor %}
-                    </div>
-                    {% endif %}
-
-                    {% if grouped_relationships.partners %}
-                    <div class="id-rel-group">
-                        <span class="id-rel-label">Partners:</span>
-                        {% for rel in grouped_relationships.partners %}
-                        <a href="/contacts/{{ rel.related_contact_id }}">{{ contact_names[rel.related_contact_id] }}</a>{% if not loop.last %}, {% endif %}
-                        {% endfor %}
-                    </div>
-                    {% endif %}
-
-                    {% if grouped_relationships.friends %}
-                    <div class="id-rel-group">
-                        <span class="id-rel-label">Friends:</span>
-                        {% for rel in grouped_relationships.friends %}
-                        <a href="/contacts/{{ rel.related_contact_id }}">{{ contact_names[rel.related_contact_id] }}</a>{% if not loop.last %}, {% endif %}
-                        {% endfor %}
-                    </div>
-                    {% endif %}
-
-                    {% if grouped_relationships.professional %}
-                    <div class="id-rel-group">
-                        <span class="id-rel-label">Professional:</span>
-                        {% for rel in grouped_relationships.professional %}
-                        <a href="/contacts/{{ rel.related_contact_id }}">{{ contact_names[rel.related_contact_id] }}</a><span class="id-rel-type">({{ rel.relationship_type|replace("_", " ")|title }})</span>{% if not loop.last %}, {% endif %}
-                        {% endfor %}
-                    </div>
-                    {% endif %}
-
-                    {% if grouped_relationships.other %}
-                    <div class="id-rel-group">
-                        <span class="id-rel-label">Other:</span>
-                        {% for rel in grouped_relationships.other %}
-                        <a href="/contacts/{{ rel.related_contact_id }}">{{ contact_names[rel.related_contact_id] }}</a><span class="id-rel-type">({{ rel.relationship_type|replace("_", " ")|title }})</span>{% if not loop.last %}, {% endif %}
-                        {% endfor %}
-                    </div>
-                    {% endif %}
-
-                    {% if contact.related_from %}
-                    <div class="id-rel-group">
-                        <span class="id-rel-label">Known by:</span>
-                        {% for rel in contact.related_from %}
-                        <a href="/contacts/{{ rel.contact_id }}">{{ contact_names[rel.contact_id] }}</a>{% if not loop.last %}, {% endif %}
-                        {% endfor %}
-                    </div>
-                    {% endif %}
-                </div>
-            </div>
-        </div>
-
-        {% if contact.needs %}
-        <div class="id-card-warnings">
-            {% for need in contact.needs %}
-            <div class="id-warning">
-                <span class="warning-dot"></span>
-                Warning: {{ need.name }}
-                {% if need.description %}<span class="warning-desc"> - {{ need.description }}</span>{% endif %}
-            </div>
-            {% endfor %}
-        </div>
-        {% endif %}
-    </div>
-
-    <details class="id-card-manage">
-        <summary>Manage Contact</summary>
-
-        <div class="manage-section">
-            <h3>Manage Relationships</h3>
-            <div id="manage-relationships" class="manage-relationships">
-                {% include "partials/manage_relationships.html" %}
-            </div>
-
-            {% if all_contacts %}
-            <form hx-post="/htmx/contacts/{{ contact.id }}/add-relationship"
-                  hx-target="#manage-relationships"
-                  hx-swap="innerHTML"
-                  class="add-form">
-                <select name="related_contact_id" required>
-                    <option value="">Select contact...</option>
-                    {% for other in all_contacts %}
-                    {% if other.id != contact.id %}
-                    <option value="{{ other.id }}">{{ other.name }}</option>
-                    {% endif %}
-                    {% endfor %}
-                </select>
-                <select name="relationship_type" required>
-                    <option value="">Select relationship type...</option>
-                    {% for rel_type in relationship_types %}
-                    <option value="{{ rel_type.value }}">{{ rel_type.display_name }}</option>
-                    {% endfor %}
-                </select>
-                <button type="submit" class="btn btn-primary">Add Relationship</button>
-            </form>
-            {% endif %}
-        </div>
-
-        <div class="manage-section">
-            <h3>Manage Needs/Warnings</h3>
-            <div id="manage-needs">
-                {% include "partials/manage_needs.html" %}
-            </div>
-
-            {% if available_needs %}
-            <form hx-post="/htmx/contacts/{{ contact.id }}/add-need"
-                  hx-target="#manage-needs"
-                  hx-swap="innerHTML"
-                  class="add-form">
-                <select name="need_id" required>
-                    <option value="">Select a need...</option>
-                    {% for need in available_needs %}
-                    <option value="{{ need.id }}">{{ need.name }}</option>
-                    {% endfor %}
-                </select>
-                <button type="submit" class="btn btn-primary">Add Need</button>
-            </form>
-            {% endif %}
-        </div>
-    </details>
-</div>
-{% endblock %}

python/api/templates/contact_form.html

@@ -1,115 +0,0 @@
-{% extends "base.html" %}
-{% block title %}{{ "Edit " + contact.name if contact else "New Contact" }}{% endblock %}
-{% block content %}
-<div class="contact-form">
-    <h1>{{ "Edit Contact" if contact else "New Contact" }}</h1>
-
-    {% if contact %}
-    <form method="post" action="/htmx/contacts/{{ contact.id }}/edit">
-    {% else %}
-    <form method="post" action="/htmx/contacts/new">
-    {% endif %}
-
-        <div class="form-group">
-            <label for="name">Name *</label>
-            <input id="name" name="name" type="text" value="{{ contact.name if contact else '' }}" required>
-        </div>
-
-        <div class="form-row">
-            <div class="form-group">
-                <label for="legal_name">Legal Name</label>
-                <input id="legal_name" name="legal_name" type="text" value="{{ contact.legal_name or '' }}">
-            </div>
-            <div class="form-group">
-                <label for="suffix">Suffix</label>
-                <input id="suffix" name="suffix" type="text" value="{{ contact.suffix or '' }}">
-            </div>
-        </div>
-
-        <div class="form-row">
-            <div class="form-group">
-                <label for="age">Age</label>
-                <input id="age" name="age" type="number" value="{{ contact.age if contact and contact.age is not none else '' }}">
-            </div>
-            <div class="form-group">
-                <label for="gender">Gender</label>
-                <input id="gender" name="gender" type="text" value="{{ contact.gender or '' }}">
-            </div>
-        </div>
-
-        <div class="form-group">
-            <label for="current_job">Current Job</label>
-            <input id="current_job" name="current_job" type="text" value="{{ contact.current_job or '' }}">
-        </div>
-
-        <div class="form-group">
-            <label for="timezone">Timezone</label>
-            <input id="timezone" name="timezone" type="text" value="{{ contact.timezone or '' }}">
-        </div>
-
-        <div class="form-group">
-            <label for="profile_pic">Profile Picture URL</label>
-            <input id="profile_pic" name="profile_pic" type="url" placeholder="https://example.com/photo.jpg" value="{{ contact.profile_pic or '' }}">
-        </div>
-
-        <div class="form-group">
-            <label for="bio">Bio</label>
-            <textarea id="bio" name="bio" rows="3">{{ contact.bio or '' }}</textarea>
-        </div>
-
-        <div class="form-group">
-            <label for="goals">Goals</label>
-            <textarea id="goals" name="goals" rows="3">{{ contact.goals or '' }}</textarea>
-        </div>
-
-        <div class="form-group">
-            <label for="social_structure_style">Social Structure Style</label>
-            <input id="social_structure_style" name="social_structure_style" type="text" value="{{ contact.social_structure_style or '' }}">
-        </div>
-
-        <div class="form-group">
-            <label for="self_sufficiency_score">Self-Sufficiency Score (1-10)</label>
-            <input id="self_sufficiency_score" name="self_sufficiency_score" type="number" min="1" max="10" value="{{ contact.self_sufficiency_score if contact and contact.self_sufficiency_score is not none else '' }}">
-        </div>
-
-        <div class="form-group">
-            <label for="safe_conversation_starters">Safe Conversation Starters</label>
-            <textarea id="safe_conversation_starters" name="safe_conversation_starters" rows="2">{{ contact.safe_conversation_starters or '' }}</textarea>
-        </div>
-
-        <div class="form-group">
-            <label for="topics_to_avoid">Topics to Avoid</label>
-            <textarea id="topics_to_avoid" name="topics_to_avoid" rows="2">{{ contact.topics_to_avoid or '' }}</textarea>
-        </div>
-
-        <div class="form-group">
-            <label for="ssn">SSN</label>
-            <input id="ssn" name="ssn" type="text" value="{{ contact.ssn or '' }}">
-        </div>
-
-        {% if all_needs %}
-        <div class="form-group">
-            <label>Needs/Accommodations</label>
-            <div class="checkbox-group">
-                {% for need in all_needs %}
-                <label class="checkbox-label">
-                    <input type="checkbox" name="need_ids" value="{{ need.id }}"
-                           {% if contact and need in contact.needs %}checked{% endif %}>
-                    {{ need.name }}
-                </label>
-                {% endfor %}
-            </div>
-        </div>
-        {% endif %}
-
-        <div class="form-actions">
-            <button type="submit" class="btn btn-primary">Save</button>
-            {% if contact %}
-            <a href="/contacts/{{ contact.id }}" class="btn">Cancel</a>
-            {% else %}
-            <a href="/contacts" class="btn">Cancel</a>
-            {% endif %}
-        </div>
-    </form>
-</div>
-{% endblock %}

python/api/templates/contact_list.html

@@ -1,14 +0,0 @@
-{% extends "base.html" %}
-{% block title %}Contacts{% endblock %}
-{% block content %}
-<div class="contact-list">
-    <div class="header">
-        <h1>Contacts</h1>
-        <a href="/contacts/new" class="btn btn-primary">Add Contact</a>
-    </div>
-
-    <div id="contact-table">
-        {% include "partials/contact_table.html" %}
-    </div>
-</div>
-{% endblock %}

python/api/templates/graph.html

@@ -1,198 +0,0 @@
-{% extends "base.html" %}
-{% block title %}Relationship Graph{% endblock %}
-{% block content %}
-<div class="graph-container">
-    <div class="header">
-        <h1>Relationship Graph</h1>
-    </div>
-    <p class="graph-hint">Drag nodes to reposition. Closer relationships have shorter, darker edges.</p>
-    <canvas id="graph-canvas" width="900" height="600"
-            style="border: 1px solid var(--color-border); border-radius: 8px; background: var(--color-bg); cursor: grab;">
-    </canvas>
-    <div id="selected-info"></div>
-    <div class="legend">
-        <h4>Relationship Closeness (1-10)</h4>
-        <div class="legend-items">
-            <div class="legend-item">
-                <span class="legend-line" style="background: hsl(220, 70%, 40%); height: 4px; display: inline-block;"></span>
-                <span>10 - Very Close (Spouse, Partner)</span>
-            </div>
-            <div class="legend-item">
-                <span class="legend-line" style="background: hsl(220, 70%, 52%); height: 3px; display: inline-block;"></span>
-                <span>7 - Close (Family, Best Friend)</span>
-            </div>
-            <div class="legend-item">
-                <span class="legend-line" style="background: hsl(220, 70%, 64%); height: 2px; display: inline-block;"></span>
-                <span>4 - Moderate (Friend, Colleague)</span>
-            </div>
-            <div class="legend-item">
-                <span class="legend-line" style="background: hsl(220, 70%, 72%); height: 1px; display: inline-block;"></span>
-                <span>2 - Distant (Acquaintance)</span>
-            </div>
-        </div>
-    </div>
-</div>
-
-<script>
-(function() {
-    const RELATIONSHIP_DISPLAY = {{ relationship_type_display|tojson }};
-    const graphData = {{ graph_data|tojson }};
-
-    const canvas = document.getElementById('graph-canvas');
-    const ctx = canvas.getContext('2d');
-    const width = canvas.width;
-    const height = canvas.height;
-    const centerX = width / 2;
-    const centerY = height / 2;
-
-    const nodes = graphData.nodes.map(function(node) {
-        return Object.assign({}, node, {
-            x: centerX + (Math.random() - 0.5) * 300,
-            y: centerY + (Math.random() - 0.5) * 300,
-            vx: 0,
-            vy: 0
-        });
-    });
-
-    const nodeMap = new Map(nodes.map(function(node) { return [node.id, node]; }));
-
-    const edges = graphData.edges.map(function(edge) {
-        const sourceNode = nodeMap.get(edge.source);
-        const targetNode = nodeMap.get(edge.target);
-        if (!sourceNode || !targetNode) return null;
-        return Object.assign({}, edge, { sourceNode: sourceNode, targetNode: targetNode });
-    }).filter(function(edge) { return edge !== null; });
-
-    let dragNode = null;
-    let selectedNode = null;
-
-    const repulsion = 5000;
-    const springStrength = 0.05;
-    const baseSpringLength = 150;
-    const damping = 0.9;
-    const centerPull = 0.01;
-
-    function simulate() {
-        for (const node of nodes) { node.vx = 0; node.vy = 0; }
-        for (let i = 0; i < nodes.length; i++) {
-            for (let j = i + 1; j < nodes.length; j++) {
-                const dx = nodes[j].x - nodes[i].x;
-                const dy = nodes[j].y - nodes[i].y;
-                const dist = Math.sqrt(dx * dx + dy * dy) || 1;
-                const force = repulsion / (dist * dist);
-                const fx = (dx / dist) * force;
-                const fy = (dy / dist) * force;
-                nodes[i].vx -= fx; nodes[i].vy -= fy;
-                nodes[j].vx += fx; nodes[j].vy += fy;
-            }
-        }
-        for (const edge of edges) {
-            const dx = edge.targetNode.x - edge.sourceNode.x;
-            const dy = edge.targetNode.y - edge.sourceNode.y;
-            const dist = Math.sqrt(dx * dx + dy * dy) || 1;
-            const normalizedWeight = edge.closeness_weight / 10;
-            const idealLength = baseSpringLength * (1.5 - normalizedWeight);
-            const displacement = dist - idealLength;
-            const force = springStrength * displacement;
-            const fx = (dx / dist) * force;
-            const fy = (dy / dist) * force;
-            edge.sourceNode.vx += fx; edge.sourceNode.vy += fy;
-            edge.targetNode.vx -= fx; edge.targetNode.vy -= fy;
-        }
-        for (const node of nodes) {
-            node.vx += (centerX - node.x) * centerPull;
-            node.vy += (centerY - node.y) * centerPull;
-        }
-        for (const node of nodes) {
-            if (node === dragNode) continue;
-            node.x += node.vx * damping;
-            node.y += node.vy * damping;
-            node.x = Math.max(30, Math.min(width - 30, node.x));
-            node.y = Math.max(30, Math.min(height - 30, node.y));
-        }
-    }
-
-    function getEdgeColor(weight) {
-        const normalized = weight / 10;
-        return 'hsl(220, 70%, ' + (80 - normalized * 40) + '%)';
-    }
-
-    function draw() {
-        ctx.clearRect(0, 0, width, height);
-        for (const edge of edges) {
-            const lineWidth = 1 + (edge.closeness_weight / 10) * 3;
-            ctx.strokeStyle = getEdgeColor(edge.closeness_weight);
-            ctx.lineWidth = lineWidth;
-            ctx.beginPath();
-            ctx.moveTo(edge.sourceNode.x, edge.sourceNode.y);
-            ctx.lineTo(edge.targetNode.x, edge.targetNode.y);
-            ctx.stroke();
-            const midX = (edge.sourceNode.x + edge.targetNode.x) / 2;
-            const midY = (edge.sourceNode.y + edge.targetNode.y) / 2;
-            ctx.fillStyle = '#666';
-            ctx.font = '10px sans-serif';
-            ctx.textAlign = 'center';
-            const label = RELATIONSHIP_DISPLAY[edge.relationship_type] || edge.relationship_type;
-            ctx.fillText(label, midX, midY - 5);
-        }
-        for (const node of nodes) {
-            const isSelected = node === selectedNode;
-            const radius = isSelected ? 25 : 20;
-            ctx.beginPath();
-            ctx.arc(node.x, node.y, radius, 0, Math.PI * 2);
-            ctx.fillStyle = isSelected ? '#0066cc' : '#fff';
-            ctx.fill();
-            ctx.strokeStyle = '#0066cc';
-            ctx.lineWidth = 2;
-            ctx.stroke();
-            ctx.fillStyle = isSelected ? '#fff' : '#333';
-            ctx.font = '12px sans-serif';
-            ctx.textAlign = 'center';
-            ctx.textBaseline = 'middle';
-            const name = node.name.length > 10 ? node.name.slice(0, 9) + '\u2026' : node.name;
-            ctx.fillText(name, node.x, node.y);
-        }
-    }
-
-    function animate() {
-        simulate();
-        draw();
-        requestAnimationFrame(animate);
-    }
-    animate();
-
-    function getNodeAt(x, y) {
-        for (const node of nodes) {
-            const dx = x - node.x;
-            const dy = y - node.y;
-            if (dx * dx + dy * dy < 400) return node;
-        }
-        return null;
-    }
-
-    canvas.addEventListener('mousedown', function(event) {
-        const rect = canvas.getBoundingClientRect();
-        const node = getNodeAt(event.clientX - rect.left, event.clientY - rect.top);
-        if (node) {
-            dragNode = node;
-            selectedNode = node;
-            const infoDiv = document.getElementById('selected-info');
-            let html = '<div class="selected-info"><h3>' + node.name + '</h3>';
-            if (node.current_job) html += '<p>Job: ' + node.current_job + '</p>';
-            html += '<a href="/contacts/' + node.id + '">View details</a></div>';
-            infoDiv.innerHTML = html;
-        }
-    });
-
-    canvas.addEventListener('mousemove', function(event) {
-        if (!dragNode) return;
-        const rect = canvas.getBoundingClientRect();
-        dragNode.x = event.clientX - rect.left;
-        dragNode.y = event.clientY - rect.top;
-    });
-
-    canvas.addEventListener('mouseup', function() { dragNode = null; });
-    canvas.addEventListener('mouseleave', function() { dragNode = null; });
-})();
-</script>
-{% endblock %}

python/api/templates/need_list.html

@@ -1,31 +0,0 @@
-{% extends "base.html" %}
-{% block title %}Needs{% endblock %}
-{% block content %}
-<div class="need-list">
-    <div class="header">
-        <h1>Needs / Accommodations</h1>
-        <button class="btn btn-primary" onclick="document.getElementById('need-form').toggleAttribute('hidden')">Add Need</button>
-    </div>
-
-    <form id="need-form" hidden
-          hx-post="/htmx/needs"
-          hx-target="#need-items"
-          hx-swap="innerHTML"
-          hx-on::after-request="if(event.detail.successful) this.reset()"
-          class="need-form">
-        <div class="form-group">
-            <label for="name">Name *</label>
-            <input id="name" name="name" type="text" placeholder="e.g., Light Sensitive, ADHD" required>
-        </div>
-        <div class="form-group">
-            <label for="description">Description</label>
-            <textarea id="description" name="description" placeholder="Optional description..." rows="2"></textarea>
-        </div>
-        <button type="submit" class="btn btn-primary">Create</button>
-    </form>
-
-    <div id="need-items">
-        {% include "partials/need_items.html" %}
-    </div>
-</div>
-{% endblock %}

python/api/templates/partials/contact_table.html

@@ -1,33 +0,0 @@
-{% if contacts %}
-<table>
-    <thead>
-        <tr>
-            <th>Name</th>
-            <th>Job</th>
-            <th>Timezone</th>
-            <th>Actions</th>
-        </tr>
-    </thead>
-    <tbody>
-        {% for contact in contacts %}
-        <tr id="contact-row-{{ contact.id }}">
-            <td><a href="/contacts/{{ contact.id }}">{{ contact.name }}</a></td>
-            <td>{{ contact.current_job or "-" }}</td>
-            <td>{{ contact.timezone or "-" }}</td>
-            <td>
-                <a href="/contacts/{{ contact.id }}/edit" class="btn">Edit</a>
-                <button class="btn btn-danger"
-                        hx-delete="/api/contacts/{{ contact.id }}"
-                        hx-target="#contact-row-{{ contact.id }}"
-                        hx-swap="outerHTML"
-                        hx-confirm="Delete this contact?">
-                    Delete
-                </button>
-            </td>
-        </tr>
-        {% endfor %}
-    </tbody>
-</table>
-{% else %}
-<p>No contacts yet.</p>
-{% endif %}

python/api/templates/partials/manage_needs.html

@@ -1,14 +0,0 @@
-<ul class="manage-needs-list">
-    {% for need in contact.needs %}
-    <li id="contact-need-{{ need.id }}">
-        <strong>{{ need.name }}</strong>
-        {% if need.description %}<span> - {{ need.description }}</span>{% endif %}
-        <button class="btn btn-small btn-danger"
-                hx-delete="/api/contacts/{{ contact.id }}/needs/{{ need.id }}"
-                hx-target="#contact-need-{{ need.id }}"
-                hx-swap="outerHTML">
-            Remove
-        </button>
-    </li>
-    {% endfor %}
-</ul>

python/api/templates/partials/manage_relationships.html

@@ -1,23 +0,0 @@
-{% for rel in contact.related_to %}
-<div class="manage-rel-item" id="rel-{{ contact.id }}-{{ rel.related_contact_id }}">
-    <a href="/contacts/{{ rel.related_contact_id }}">{{ contact_names[rel.related_contact_id] }}</a>
-    <span class="tag">{{ rel.relationship_type|replace("_", " ")|title }}</span>
-    <label class="weight-control">
-        <span>Closeness:</span>
-        <input type="range" min="1" max="10" value="{{ rel.closeness_weight }}"
-               hx-post="/htmx/contacts/{{ contact.id }}/relationships/{{ rel.related_contact_id }}/weight"
-               hx-trigger="change"
-               hx-include="this"
-               name="closeness_weight"
-               hx-swap="none"
-               oninput="this.nextElementSibling.textContent = this.value">
-        <span class="weight-value">{{ rel.closeness_weight }}</span>
-    </label>
-    <button class="btn btn-small btn-danger"
-            hx-delete="/api/contacts/{{ contact.id }}/relationships/{{ rel.related_contact_id }}"
-            hx-target="#rel-{{ contact.id }}-{{ rel.related_contact_id }}"
-            hx-swap="outerHTML">
-        Remove
-    </button>
-</div>
-{% endfor %}

python/api/templates/partials/need_items.html

@@ -1,21 +0,0 @@
-{% if needs %}
-<ul class="need-items">
-    {% for need in needs %}
-    <li id="need-item-{{ need.id }}">
-        <div class="need-info">
-            <strong>{{ need.name }}</strong>
-            {% if need.description %}<p>{{ need.description }}</p>{% endif %}
-        </div>
-        <button class="btn btn-danger"
-                hx-delete="/api/needs/{{ need.id }}"
-                hx-target="#need-item-{{ need.id }}"
-                hx-swap="outerHTML"
-                hx-confirm="Delete this need?">
-            Delete
-        </button>
-    </li>
-    {% endfor %}
-</ul>
-{% else %}
-<p>No needs defined yet.</p>
-{% endif %}

python/data_science/__init__.py

@@ -1,3 +0,0 @@
-"""Data science CLI tools."""
-
-from __future__ import annotations

python/data_science/ingest_congress.py

@@ -1,613 +0,0 @@
-"""Ingestion pipeline for loading congress data from unitedstates/congress JSON files.
-
-Loads legislators, bills, votes, vote records, and bill text into the data_science_dev database.
-Expects the parent directory to contain congress-tracker/ and congress-legislators/ as siblings.
-
-Usage:
-    ingest-congress /path/to/parent/
-    ingest-congress /path/to/parent/ --congress 118
-    ingest-congress /path/to/parent/ --congress 118 --only bills
-"""
-
-from __future__ import annotations
-
-import logging
-from pathlib import Path  # noqa: TC003 needed at runtime for typer CLI argument
-from typing import TYPE_CHECKING, Annotated
-
-import orjson
-import typer
-import yaml
-from sqlalchemy import select
-from sqlalchemy.orm import Session
-
-from python.common import configure_logger
-from python.orm.common import get_postgres_engine
-from python.orm.data_science_dev.congress import Bill, BillText, Legislator, LegislatorSocialMedia, Vote, VoteRecord
-
-if TYPE_CHECKING:
-    from collections.abc import Iterator
-
-    from sqlalchemy.engine import Engine
-
-logger = logging.getLogger(__name__)
-
-BATCH_SIZE = 10_000
-
-app = typer.Typer(help="Ingest unitedstates/congress data into data_science_dev.")
-
-
-@app.command()
-def main(
-    parent_dir: Annotated[
-        Path,
-        typer.Argument(help="Parent directory containing congress-tracker/ and congress-legislators/"),
-    ],
-    congress: Annotated[int | None, typer.Option(help="Only ingest a specific congress number")] = None,
-    only: Annotated[
-        str | None,
-        typer.Option(help="Only run a specific step: legislators, social-media, bills, votes, bill-text"),
-    ] = None,
-) -> None:
-    """Ingest congress data from unitedstates/congress JSON files."""
-    configure_logger(level="INFO")
-
-    data_dir = parent_dir / "congress-tracker/congress/data/"
-    legislators_dir = parent_dir / "congress-legislators"
-
-    if not data_dir.is_dir():
-        typer.echo(f"Expected congress-tracker/ directory: {data_dir}", err=True)
-        raise typer.Exit(code=1)
-
-    if not legislators_dir.is_dir():
-        typer.echo(f"Expected congress-legislators/ directory: {legislators_dir}", err=True)
-        raise typer.Exit(code=1)
-
-    engine = get_postgres_engine(name="DATA_SCIENCE_DEV")
-
-    congress_dirs = _resolve_congress_dirs(data_dir, congress)
-    if not congress_dirs:
-        typer.echo("No congress directories found.", err=True)
-        raise typer.Exit(code=1)
-
-    logger.info("Found %d congress directories to process", len(congress_dirs))
-
-    steps: dict[str, tuple] = {
-        "legislators": (ingest_legislators, (engine, legislators_dir)),
-        "legislators-social-media": (ingest_social_media, (engine, legislators_dir)),
-        "bills": (ingest_bills, (engine, congress_dirs)),
-        "votes": (ingest_votes, (engine, congress_dirs)),
-        "bill-text": (ingest_bill_text, (engine, congress_dirs)),
-    }
-
-    if only:
-        if only not in steps:
-            typer.echo(f"Unknown step: {only}. Choose from: {', '.join(steps)}", err=True)
-            raise typer.Exit(code=1)
-        steps = {only: steps[only]}
-
-    for step_name, (step_func, step_args) in steps.items():
-        logger.info("=== Starting step: %s ===", step_name)
-        step_func(*step_args)
-        logger.info("=== Finished step: %s ===", step_name)
-
-    logger.info("ingest-congress done")
-
-
-def _resolve_congress_dirs(data_dir: Path, congress: int | None) -> list[Path]:
-    """Find congress number directories under data_dir."""
-    if congress is not None:
-        target = data_dir / str(congress)
-        return [target] if target.is_dir() else []
-    return sorted(path for path in data_dir.iterdir() if path.is_dir() and path.name.isdigit())
-
-
-def _flush_batch(session: Session, batch: list[object], label: str) -> int:
-    """Add a batch of ORM objects to the session and commit. Returns count added."""
-    if not batch:
-        return 0
-    session.add_all(batch)
-    session.commit()
-    count = len(batch)
-    logger.info("Committed %d %s", count, label)
-    batch.clear()
-    return count
-
-
-# ---------------------------------------------------------------------------
-# Legislators — loaded from congress-legislators YAML files
-# ---------------------------------------------------------------------------
-
-
-def ingest_legislators(engine: Engine, legislators_dir: Path) -> None:
-    """Load legislators from congress-legislators YAML files."""
-    legislators_data = _load_legislators_yaml(legislators_dir)
-    logger.info("Loaded %d legislators from YAML files", len(legislators_data))
-
-    with Session(engine) as session:
-        existing_legislators = {
-            legislator.bioguide_id: legislator for legislator in session.scalars(select(Legislator)).all()
-        }
-        logger.info("Found %d existing legislators in DB", len(existing_legislators))
-
-        total_inserted = 0
-        total_updated = 0
-        for entry in legislators_data:
-            bioguide_id = entry.get("id", {}).get("bioguide")
-            if not bioguide_id:
-                continue
-
-            fields = _parse_legislator(entry)
-            if existing := existing_legislators.get(bioguide_id):
-                changed = False
-                for field, value in fields.items():
-                    if value is not None and getattr(existing, field) != value:
-                        setattr(existing, field, value)
-                        changed = True
-                if changed:
-                    total_updated += 1
-            else:
-                session.add(Legislator(bioguide_id=bioguide_id, **fields))
-                total_inserted += 1
-
-        session.commit()
-    logger.info("Inserted %d new legislators, updated %d existing", total_inserted, total_updated)
-
-
-def _load_legislators_yaml(legislators_dir: Path) -> list[dict]:
-    """Load and combine legislators-current.yaml and legislators-historical.yaml."""
-    legislators: list[dict] = []
-    for filename in ("legislators-current.yaml", "legislators-historical.yaml"):
-        path = legislators_dir / filename
-        if not path.exists():
-            logger.warning("Legislators file not found: %s", path)
-            continue
-        with path.open() as file:
-            data = yaml.safe_load(file)
-            if isinstance(data, list):
-                legislators.extend(data)
-    return legislators
-
-
-def _parse_legislator(entry: dict) -> dict:
-    """Extract Legislator fields from a congress-legislators YAML entry."""
-    ids = entry.get("id", {})
-    name = entry.get("name", {})
-    bio = entry.get("bio", {})
-    terms = entry.get("terms", [])
-    latest_term = terms[-1] if terms else {}
-
-    fec_ids = ids.get("fec")
-    fec_ids_joined = ",".join(fec_ids) if isinstance(fec_ids, list) else fec_ids
-
-    chamber = latest_term.get("type")
-    chamber_normalized = {"rep": "House", "sen": "Senate"}.get(chamber, chamber)
-
-    return {
-        "thomas_id": ids.get("thomas"),
-        "lis_id": ids.get("lis"),
-        "govtrack_id": ids.get("govtrack"),
-        "opensecrets_id": ids.get("opensecrets"),
-        "fec_ids": fec_ids_joined,
-        "first_name": name.get("first"),
-        "last_name": name.get("last"),
-        "official_full_name": name.get("official_full"),
-        "nickname": name.get("nickname"),
-        "birthday": bio.get("birthday"),
-        "gender": bio.get("gender"),
-        "current_party": latest_term.get("party"),
-        "current_state": latest_term.get("state"),
-        "current_district": latest_term.get("district"),
-        "current_chamber": chamber_normalized,
-    }
-
-
-# ---------------------------------------------------------------------------
-# Social Media — loaded from legislators-social-media.yaml
-# ---------------------------------------------------------------------------
-
-SOCIAL_MEDIA_PLATFORMS = {
-    "twitter": "https://twitter.com/{account}",
-    "facebook": "https://facebook.com/{account}",
-    "youtube": "https://youtube.com/{account}",
-    "instagram": "https://instagram.com/{account}",
-    "mastodon": None,
-}
-
-
-def ingest_social_media(engine: Engine, legislators_dir: Path) -> None:
-    """Load social media accounts from legislators-social-media.yaml."""
-    social_media_path = legislators_dir / "legislators-social-media.yaml"
-    if not social_media_path.exists():
-        logger.warning("Social media file not found: %s", social_media_path)
-        return
-
-    with social_media_path.open() as file:
-        social_media_data = yaml.safe_load(file)
-
-    if not isinstance(social_media_data, list):
-        logger.warning("Unexpected format in %s", social_media_path)
-        return
-
-    logger.info("Loaded %d entries from legislators-social-media.yaml", len(social_media_data))
-
-    with Session(engine) as session:
-        legislator_map = _build_legislator_map(session)
-        existing_accounts = {
-            (account.legislator_id, account.platform)
-            for account in session.scalars(select(LegislatorSocialMedia)).all()
-        }
-        logger.info("Found %d existing social media accounts in DB", len(existing_accounts))
-
-        total_inserted = 0
-        total_updated = 0
-        for entry in social_media_data:
-            bioguide_id = entry.get("id", {}).get("bioguide")
-            if not bioguide_id:
-                continue
-
-            legislator_id = legislator_map.get(bioguide_id)
-            if legislator_id is None:
-                continue
-
-            social = entry.get("social", {})
-            for platform, url_template in SOCIAL_MEDIA_PLATFORMS.items():
-                account_name = social.get(platform)
-                if not account_name:
-                    continue
-
-                url = url_template.format(account=account_name) if url_template else None
-
-                if (legislator_id, platform) in existing_accounts:
-                    total_updated += 1
-                else:
-                    session.add(
-                        LegislatorSocialMedia(
-                            legislator_id=legislator_id,
-                            platform=platform,
-                            account_name=str(account_name),
-                            url=url,
-                            source="https://github.com/unitedstates/congress-legislators",
-                        )
-                    )
-                    existing_accounts.add((legislator_id, platform))
-                    total_inserted += 1
-
-        session.commit()
-    logger.info("Inserted %d new social media accounts, updated %d existing", total_inserted, total_updated)
-
-
-def _iter_voters(position_group: object) -> Iterator[dict]:
-    """Yield voter dicts from a vote position group (handles list, single dict, or string)."""
-    if isinstance(position_group, dict):
-        yield position_group
-    elif isinstance(position_group, list):
-        for voter in position_group:
-            if isinstance(voter, dict):
-                yield voter
-
-
-# ---------------------------------------------------------------------------
-# Bills
-# ---------------------------------------------------------------------------
-
-
-def ingest_bills(engine: Engine, congress_dirs: list[Path]) -> None:
-    """Load bill data.json files."""
-    with Session(engine) as session:
-        existing_bills = {(bill.congress, bill.bill_type, bill.number) for bill in session.scalars(select(Bill)).all()}
-        logger.info("Found %d existing bills in DB", len(existing_bills))
-
-        total_inserted = 0
-        batch: list[Bill] = []
-        for congress_dir in congress_dirs:
-            bills_dir = congress_dir / "bills"
-            if not bills_dir.is_dir():
-                continue
-            logger.info("Scanning bills from %s", congress_dir.name)
-            for bill_file in bills_dir.rglob("data.json"):
-                data = _read_json(bill_file)
-                if data is None:
-                    continue
-                bill = _parse_bill(data, existing_bills)
-                if bill is not None:
-                    batch.append(bill)
-                    if len(batch) >= BATCH_SIZE:
-                        total_inserted += _flush_batch(session, batch, "bills")
-
-        total_inserted += _flush_batch(session, batch, "bills")
-    logger.info("Inserted %d new bills total", total_inserted)
-
-
-def _parse_bill(data: dict, existing_bills: set[tuple[int, str, int]]) -> Bill | None:
-    """Parse a bill data.json dict into a Bill ORM object, skipping existing."""
-    raw_congress = data.get("congress")
-    bill_type = data.get("bill_type")
-    raw_number = data.get("number")
-    if raw_congress is None or bill_type is None or raw_number is None:
-        return None
-    congress = int(raw_congress)
-    number = int(raw_number)
-    if (congress, bill_type, number) in existing_bills:
-        return None
-
-    sponsor_bioguide = None
-    sponsor = data.get("sponsor")
-    if sponsor:
-        sponsor_bioguide = sponsor.get("bioguide_id")
-
-    return Bill(
-        congress=congress,
-        bill_type=bill_type,
-        number=number,
-        title=data.get("short_title") or data.get("official_title"),
-        title_short=data.get("short_title"),
-        official_title=data.get("official_title"),
-        status=data.get("status"),
-        status_at=data.get("status_at"),
-        sponsor_bioguide_id=sponsor_bioguide,
-        subjects_top_term=data.get("subjects_top_term"),
-    )
-
-
-# ---------------------------------------------------------------------------
-# Votes (and vote records)
-# ---------------------------------------------------------------------------
-
-
-def ingest_votes(engine: Engine, congress_dirs: list[Path]) -> None:
-    """Load vote data.json files with their vote records."""
-    with Session(engine) as session:
-        legislator_map = _build_legislator_map(session)
-        logger.info("Loaded %d legislators into lookup map", len(legislator_map))
-        bill_map = _build_bill_map(session)
-        logger.info("Loaded %d bills into lookup map", len(bill_map))
-        existing_votes = {
-            (vote.congress, vote.chamber, vote.session, vote.number) for vote in session.scalars(select(Vote)).all()
-        }
-        logger.info("Found %d existing votes in DB", len(existing_votes))
-
-        total_inserted = 0
-        batch: list[Vote] = []
-        for congress_dir in congress_dirs:
-            votes_dir = congress_dir / "votes"
-            if not votes_dir.is_dir():
-                continue
-            logger.info("Scanning votes from %s", congress_dir.name)
-            for vote_file in votes_dir.rglob("data.json"):
-                data = _read_json(vote_file)
-                if data is None:
-                    continue
-                vote = _parse_vote(data, legislator_map, bill_map, existing_votes)
-                if vote is not None:
-                    batch.append(vote)
-                    if len(batch) >= BATCH_SIZE:
-                        total_inserted += _flush_batch(session, batch, "votes")
-
-        total_inserted += _flush_batch(session, batch, "votes")
-    logger.info("Inserted %d new votes total", total_inserted)
-
-
-def _build_legislator_map(session: Session) -> dict[str, int]:
-    """Build a mapping of bioguide_id -> legislator.id."""
-    return {legislator.bioguide_id: legislator.id for legislator in session.scalars(select(Legislator)).all()}
-
-
-def _build_bill_map(session: Session) -> dict[tuple[int, str, int], int]:
-    """Build a mapping of (congress, bill_type, number) -> bill.id."""
-    return {(bill.congress, bill.bill_type, bill.number): bill.id for bill in session.scalars(select(Bill)).all()}
-
-
-def _parse_vote(
-    data: dict,
-    legislator_map: dict[str, int],
-    bill_map: dict[tuple[int, str, int], int],
-    existing_votes: set[tuple[int, str, int, int]],
-) -> Vote | None:
-    """Parse a vote data.json dict into a Vote ORM object with records."""
-    raw_congress = data.get("congress")
-    chamber = data.get("chamber")
-    raw_number = data.get("number")
-    vote_date = data.get("date")
-    if raw_congress is None or chamber is None or raw_number is None or vote_date is None:
-        return None
-
-    raw_session = data.get("session")
-    if raw_session is None:
-        return None
-
-    congress = int(raw_congress)
-    number = int(raw_number)
-    session_number = int(raw_session)
-
-    # Normalize chamber from "h"/"s" to "House"/"Senate"
-    chamber_normalized = {"h": "House", "s": "Senate"}.get(chamber, chamber)
-
-    if (congress, chamber_normalized, session_number, number) in existing_votes:
-        return None
-
-    # Resolve linked bill
-    bill_id = None
-    bill_ref = data.get("bill")
-    if bill_ref:
-        bill_key = (
-            int(bill_ref.get("congress", congress)),
-            bill_ref.get("type"),
-            int(bill_ref.get("number", 0)),
-        )
-        bill_id = bill_map.get(bill_key)
-
-    raw_votes = data.get("votes", {})
-    vote_counts = _count_votes(raw_votes)
-    vote_records = _build_vote_records(raw_votes, legislator_map)
-
-    return Vote(
-        congress=congress,
-        chamber=chamber_normalized,
-        session=session_number,
-        number=number,
-        vote_type=data.get("type"),
-        question=data.get("question"),
-        result=data.get("result"),
-        result_text=data.get("result_text"),
-        vote_date=vote_date[:10] if isinstance(vote_date, str) else vote_date,
-        bill_id=bill_id,
-        vote_records=vote_records,
-        **vote_counts,
-    )
-
-
-def _count_votes(raw_votes: dict) -> dict[str, int]:
-    """Count voters per position category, correctly handling dict and list formats."""
-    yea_count = 0
-    nay_count = 0
-    not_voting_count = 0
-    present_count = 0
-
-    for position, position_group in raw_votes.items():
-        voter_count = sum(1 for _ in _iter_voters(position_group))
-        if position in ("Yea", "Aye"):
-            yea_count += voter_count
-        elif position in ("Nay", "No"):
-            nay_count += voter_count
-        elif position == "Not Voting":
-            not_voting_count += voter_count
-        elif position == "Present":
-            present_count += voter_count
-
-    return {
-        "yea_count": yea_count,
-        "nay_count": nay_count,
-        "not_voting_count": not_voting_count,
-        "present_count": present_count,
-    }
-
-
-def _build_vote_records(raw_votes: dict, legislator_map: dict[str, int]) -> list[VoteRecord]:
-    """Build VoteRecord objects from raw vote data."""
-    records: list[VoteRecord] = []
-    for position, position_group in raw_votes.items():
-        for voter in _iter_voters(position_group):
-            bioguide_id = voter.get("id")
-            if not bioguide_id:
-                continue
-            legislator_id = legislator_map.get(bioguide_id)
-            if legislator_id is None:
-                continue
-            records.append(
-                VoteRecord(
-                    legislator_id=legislator_id,
-                    position=position,
-                )
-            )
-    return records
-
-
-# ---------------------------------------------------------------------------
-# Bill Text
-# ---------------------------------------------------------------------------
-
-
-def ingest_bill_text(engine: Engine, congress_dirs: list[Path]) -> None:
-    """Load bill text from text-versions directories."""
-    with Session(engine) as session:
-        bill_map = _build_bill_map(session)
-        logger.info("Loaded %d bills into lookup map", len(bill_map))
-        existing_bill_texts = {
-            (bill_text.bill_id, bill_text.version_code) for bill_text in session.scalars(select(BillText)).all()
-        }
-        logger.info("Found %d existing bill text versions in DB", len(existing_bill_texts))
-
-        total_inserted = 0
-        batch: list[BillText] = []
-        for congress_dir in congress_dirs:
-            logger.info("Scanning bill texts from %s", congress_dir.name)
-            for bill_text in _iter_bill_texts(congress_dir, bill_map, existing_bill_texts):
-                batch.append(bill_text)
-                if len(batch) >= BATCH_SIZE:
-                    total_inserted += _flush_batch(session, batch, "bill texts")
-
-        total_inserted += _flush_batch(session, batch, "bill texts")
-    logger.info("Inserted %d new bill text versions total", total_inserted)
-
-
-def _iter_bill_texts(
-    congress_dir: Path,
-    bill_map: dict[tuple[int, str, int], int],
-    existing_bill_texts: set[tuple[int, str]],
-) -> Iterator[BillText]:
-    """Yield BillText objects for a single congress directory, skipping existing."""
-    bills_dir = congress_dir / "bills"
-    if not bills_dir.is_dir():
-        return
-
-    for bill_dir in bills_dir.rglob("text-versions"):
-        if not bill_dir.is_dir():
-            continue
-        bill_key = _bill_key_from_dir(bill_dir.parent, congress_dir)
-        if bill_key is None:
-            continue
-        bill_id = bill_map.get(bill_key)
-        if bill_id is None:
-            continue
-
-        for version_dir in sorted(bill_dir.iterdir()):
-            if not version_dir.is_dir():
-                continue
-            if (bill_id, version_dir.name) in existing_bill_texts:
-                continue
-            text_content = _read_bill_text(version_dir)
-            version_data = _read_json(version_dir / "data.json")
-            yield BillText(
-                bill_id=bill_id,
-                version_code=version_dir.name,
-                version_name=version_data.get("version_name") if version_data else None,
-                date=version_data.get("issued_on") if version_data else None,
-                text_content=text_content,
-            )
-
-
-def _bill_key_from_dir(bill_dir: Path, congress_dir: Path) -> tuple[int, str, int] | None:
-    """Extract (congress, bill_type, number) from directory structure."""
-    congress = int(congress_dir.name)
-    bill_type = bill_dir.parent.name
-    name = bill_dir.name
-    # Directory name is like "hr3590" — strip the type prefix to get the number
-    number_str = name[len(bill_type) :]
-    if not number_str.isdigit():
-        return None
-    return (congress, bill_type, int(number_str))
-
-
-def _read_bill_text(version_dir: Path) -> str | None:
-    """Read bill text from a version directory, preferring .txt over .xml."""
-    for extension in ("txt", "htm", "html", "xml"):
-        candidates = list(version_dir.glob(f"document.{extension}"))
-        if not candidates:
-            candidates = list(version_dir.glob(f"*.{extension}"))
-        if candidates:
-            try:
-                return candidates[0].read_text(encoding="utf-8")
-            except Exception:
-                logger.exception("Failed to read %s", candidates[0])
-    return None
-
-
-# ---------------------------------------------------------------------------
-# Helpers
-# ---------------------------------------------------------------------------
-
-
-def _read_json(path: Path) -> dict | None:
-    """Read and parse a JSON file, returning None on failure."""
-    try:
-        return orjson.loads(path.read_bytes())
-    except FileNotFoundError:
-        return None
-    except Exception:
-        logger.exception("Failed to parse %s", path)
-        return None
-
-
-if __name__ == "__main__":
-    app()

python/data_science/ingest_posts.py

@@ -1,247 +0,0 @@
-"""Ingestion pipeline for loading JSONL post files into the weekly-partitioned posts table.
-
-Usage:
-    ingest-posts /path/to/files/
-    ingest-posts /path/to/single_file.jsonl
-    ingest-posts /data/dir/ --workers 4 --batch-size 5000
-"""
-
-from __future__ import annotations
-
-import logging
-from datetime import UTC, datetime
-from pathlib import Path  # noqa: TC003 this is needed for typer
-from typing import TYPE_CHECKING, Annotated
-
-import orjson
-import psycopg
-import typer
-
-from python.common import configure_logger
-from python.orm.common import get_connection_info
-from python.parallelize import parallelize_process
-
-if TYPE_CHECKING:
-    from collections.abc import Iterator
-
-logger = logging.getLogger(__name__)
-
-
-app = typer.Typer(help="Ingest JSONL post files into the partitioned posts table.")
-
-
-@app.command()
-def main(
-    path: Annotated[Path, typer.Argument(help="Directory containing JSONL files, or a single JSONL file")],
-    batch_size: Annotated[int, typer.Option(help="Rows per INSERT batch")] = 10000,
-    workers: Annotated[int, typer.Option(help="Parallel workers for multi-file ingestion")] = 4,
-    pattern: Annotated[str, typer.Option(help="Glob pattern for JSONL files")] = "*.jsonl",
-) -> None:
-    """Ingest JSONL post files into the weekly-partitioned posts table."""
-    configure_logger(level="INFO")
-
-    logger.info("starting ingest-posts")
-    logger.info("path=%s batch_size=%d workers=%d pattern=%s", path, batch_size, workers, pattern)
-    if path.is_file():
-        ingest_file(path, batch_size=batch_size)
-    elif path.is_dir():
-        ingest_directory(path, batch_size=batch_size, max_workers=workers, pattern=pattern)
-    else:
-        typer.echo(f"Path does not exist: {path}", err=True)
-        raise typer.Exit(code=1)
-
-    logger.info("ingest-posts done")
-
-
-def ingest_directory(
-    directory: Path,
-    *,
-    batch_size: int,
-    max_workers: int,
-    pattern: str = "*.jsonl",
-) -> None:
-    """Ingest all JSONL files in a directory using parallel workers."""
-    files = sorted(directory.glob(pattern))
-    if not files:
-        logger.warning("No JSONL files found in %s", directory)
-        return
-
-    logger.info("Found %d JSONL files to ingest", len(files))
-
-    kwargs_list = [{"path": fp, "batch_size": batch_size} for fp in files]
-    parallelize_process(ingest_file, kwargs_list, max_workers=max_workers)
-
-
-SCHEMA = "main"
-
-COLUMNS = (
-    "post_id",
-    "user_id",
-    "instance",
-    "date",
-    "text",
-    "langs",
-    "like_count",
-    "reply_count",
-    "repost_count",
-    "reply_to",
-    "replied_author",
-    "thread_root",
-    "thread_root_author",
-    "repost_from",
-    "reposted_author",
-    "quotes",
-    "quoted_author",
-    "labels",
-    "sent_label",
-    "sent_score",
-)
-
-INSERT_FROM_STAGING = f"""
-    INSERT INTO {SCHEMA}.posts ({", ".join(COLUMNS)})
-    SELECT {", ".join(COLUMNS)} FROM pg_temp.staging
-    ON CONFLICT (post_id, date) DO NOTHING
-"""  # noqa: S608
-
-FAILED_INSERT = f"""
-    INSERT INTO {SCHEMA}.failed_ingestion (raw_line, error)
-    VALUES (%(raw_line)s, %(error)s)
-"""  # noqa: S608
-
-
-def get_psycopg_connection() -> psycopg.Connection:
-    """Create a raw psycopg3 connection from environment variables."""
-    database, host, port, username, password = get_connection_info("DATA_SCIENCE_DEV")
-    return psycopg.connect(
-        dbname=database,
-        host=host,
-        port=int(port),
-        user=username,
-        password=password,
-        autocommit=False,
-    )
-
-
-def ingest_file(path: Path, *, batch_size: int) -> None:
-    """Ingest a single JSONL file into the posts table."""
-    log_trigger = max(100_000 // batch_size, 1)
-    failed_lines: list[dict] = []
-    try:
-        with get_psycopg_connection() as connection:
-            for index, batch in enumerate(read_jsonl_batches(path, batch_size, failed_lines), 1):
-                ingest_batch(connection, batch)
-                if index % log_trigger == 0:
-                    logger.info("Ingested %d batches (%d rows) from %s", index, index * batch_size, path)
-
-            if failed_lines:
-                logger.warning("Recording %d malformed lines from %s", len(failed_lines), path.name)
-                with connection.cursor() as cursor:
-                    cursor.executemany(FAILED_INSERT, failed_lines)
-                connection.commit()
-    except Exception:
-        logger.exception("Failed to ingest file: %s", path)
-        raise
-
-
-def ingest_batch(connection: psycopg.Connection, batch: list[dict]) -> None:
-    """COPY batch into a temp staging table, then INSERT ... ON CONFLICT into posts."""
-    if not batch:
-        return
-
-    try:
-        with connection.cursor() as cursor:
-            cursor.execute(f"""
-                CREATE TEMP TABLE IF NOT EXISTS staging
-                (LIKE {SCHEMA}.posts INCLUDING DEFAULTS)
-                ON COMMIT DELETE ROWS
-            """)
-            cursor.execute("TRUNCATE pg_temp.staging")
-
-            with cursor.copy(f"COPY pg_temp.staging ({', '.join(COLUMNS)}) FROM STDIN") as copy:
-                for row in batch:
-                    copy.write_row(tuple(row.get(column) for column in COLUMNS))
-
-            cursor.execute(INSERT_FROM_STAGING)
-        connection.commit()
-    except Exception as error:
-        connection.rollback()
-
-        if len(batch) == 1:
-            logger.exception("Skipping bad row post_id=%s", batch[0].get("post_id"))
-            with connection.cursor() as cursor:
-                cursor.execute(
-                    FAILED_INSERT,
-                    {
-                        "raw_line": orjson.dumps(batch[0], default=str).decode(),
-                        "error": str(error),
-                    },
-                )
-            connection.commit()
-            return
-
-        midpoint = len(batch) // 2
-        ingest_batch(connection, batch[:midpoint])
-        ingest_batch(connection, batch[midpoint:])
-
-
-def read_jsonl_batches(file_path: Path, batch_size: int, failed_lines: list[dict]) -> Iterator[list[dict]]:
-    """Stream a JSONL file and yield batches of transformed rows."""
-    batch: list[dict] = []
-    with file_path.open("r", encoding="utf-8") as handle:
-        for raw_line in handle:
-            line = raw_line.strip()
-            if not line:
-                continue
-            batch.extend(parse_line(line, file_path, failed_lines))
-            if len(batch) >= batch_size:
-                yield batch
-                batch = []
-    if batch:
-        yield batch
-
-
-def parse_line(line: str, file_path: Path, failed_lines: list[dict]) -> Iterator[dict]:
-    """Parse a JSONL line, handling concatenated JSON objects."""
-    try:
-        yield transform_row(orjson.loads(line))
-    except orjson.JSONDecodeError:
-        if "}{" not in line:
-            logger.warning("Skipping malformed line in %s: %s", file_path.name, line[:120])
-            failed_lines.append({"raw_line": line, "error": "malformed JSON"})
-            return
-        fragments = line.replace("}{", "}\n{").split("\n")
-        for fragment in fragments:
-            try:
-                yield transform_row(orjson.loads(fragment))
-            except (orjson.JSONDecodeError, KeyError, ValueError) as error:
-                logger.warning("Skipping malformed fragment in %s: %s", file_path.name, fragment[:120])
-                failed_lines.append({"raw_line": fragment, "error": str(error)})
-    except Exception as error:
-        logger.exception("Skipping bad row in %s: %s", file_path.name, line[:120])
-        failed_lines.append({"raw_line": line, "error": str(error)})
-
-
-def transform_row(raw: dict) -> dict:
-    """Transform a raw JSONL row into a dict matching the Posts table columns."""
-    raw["date"] = parse_date(raw["date"])
-    if raw.get("langs") is not None:
-        raw["langs"] = orjson.dumps(raw["langs"])
-    if raw.get("text") is not None:
-        raw["text"] = raw["text"].replace("\x00", "")
-    return raw
-
-
-def parse_date(raw_date: int) -> datetime:
-    """Parse compact YYYYMMDDHHmm integer into a naive datetime (input is UTC by spec)."""
-    return datetime(
-        raw_date // 100000000,
-        (raw_date // 1000000) % 100,
-        (raw_date // 10000) % 100,
-        (raw_date // 100) % 100,
-        raw_date % 100,
-        tzinfo=UTC,
-    )
-
-
-if __name__ == "__main__":
-    app()

python/database_cli.py

@@ -83,20 +83,6 @@ DATABASES: dict[str, DatabaseConfig] = {
         base_class_name="VanInventoryBase",
         models_module="python.orm.van_inventory.models",
     ),
-    "signal_bot": DatabaseConfig(
-        env_prefix="SIGNALBOT",
-        version_location="python/alembic/signal_bot/versions",
-        base_module="python.orm.signal_bot.base",
-        base_class_name="SignalBotBase",
-        models_module="python.orm.signal_bot.models",
-    ),
-    "data_science_dev": DatabaseConfig(
-        env_prefix="DATA_SCIENCE_DEV",
-        version_location="python/alembic/data_science_dev/versions",
-        base_module="python.orm.data_science_dev.base",
-        base_class_name="DataScienceDevBase",
-        models_module="python.orm.data_science_dev.models",
-    ),
 }
 
 

python/orm/__init__.py

@@ -1,13 +1,9 @@
 """ORM package exports."""
 
-from python.orm.data_science_dev.base import DataScienceDevBase
 from python.orm.richie.base import RichieBase
-from python.orm.signal_bot.base import SignalBotBase
 from python.orm.van_inventory.base import VanInventoryBase
 
 __all__ = [
-    "DataScienceDevBase",
     "RichieBase",
-    "SignalBotBase",
     "VanInventoryBase",
 ]

python/orm/data_science_dev/__init__.py

@@ -1,11 +0,0 @@
-"""Data science dev database ORM exports."""
-
-from __future__ import annotations
-
-from python.orm.data_science_dev.base import DataScienceDevBase, DataScienceDevTableBase, DataScienceDevTableBaseBig
-
-__all__ = [
-    "DataScienceDevBase",
-    "DataScienceDevTableBase",
-    "DataScienceDevTableBaseBig",
-]

python/orm/data_science_dev/base.py

@@ -1,52 +0,0 @@
-"""Data science dev database ORM base."""
-
-from __future__ import annotations
-
-from datetime import datetime
-
-from sqlalchemy import BigInteger, DateTime, MetaData, func
-from sqlalchemy.ext.declarative import AbstractConcreteBase
-from sqlalchemy.orm import DeclarativeBase, Mapped, mapped_column
-
-from python.orm.common import NAMING_CONVENTION
-
-
-class DataScienceDevBase(DeclarativeBase):
-    """Base class for data_science_dev database ORM models."""
-
-    schema_name = "main"
-
-    metadata = MetaData(
-        schema=schema_name,
-        naming_convention=NAMING_CONVENTION,
-    )
-
-
-class _TableMixin:
-    """Shared timestamp columns for all table bases."""
-
-    created: Mapped[datetime] = mapped_column(
-        DateTime(timezone=True),
-        server_default=func.now(),
-    )
-    updated: Mapped[datetime] = mapped_column(
-        DateTime(timezone=True),
-        server_default=func.now(),
-        onupdate=func.now(),
-    )
-
-
-class DataScienceDevTableBase(_TableMixin, AbstractConcreteBase, DataScienceDevBase):
-    """Table with Integer primary key."""
-
-    __abstract__ = True
-
-    id: Mapped[int] = mapped_column(primary_key=True)
-
-
-class DataScienceDevTableBaseBig(_TableMixin, AbstractConcreteBase, DataScienceDevBase):
-    """Table with BigInteger primary key."""
-
-    __abstract__ = True
-
-    id: Mapped[int] = mapped_column(BigInteger, primary_key=True)

python/orm/data_science_dev/congress/__init__.py

@@ -1,14 +0,0 @@
-"""init."""
-
-from python.orm.data_science_dev.congress.bill import Bill, BillText
-from python.orm.data_science_dev.congress.legislator import Legislator, LegislatorSocialMedia
-from python.orm.data_science_dev.congress.vote import Vote, VoteRecord
-
-__all__ = [
-    "Bill",
-    "BillText",
-    "Legislator",
-    "LegislatorSocialMedia",
-    "Vote",
-    "VoteRecord",
-]

python/orm/data_science_dev/congress/bill.py

@@ -1,66 +0,0 @@
-"""Bill model - legislation introduced in Congress."""
-
-from __future__ import annotations
-
-from datetime import date
-from typing import TYPE_CHECKING
-
-from sqlalchemy import ForeignKey, Index, UniqueConstraint
-from sqlalchemy.orm import Mapped, mapped_column, relationship
-
-from python.orm.data_science_dev.base import DataScienceDevTableBase
-
-if TYPE_CHECKING:
-    from python.orm.data_science_dev.congress.vote import Vote
-
-
-class Bill(DataScienceDevTableBase):
-    """Legislation with congress number, type, titles, status, and sponsor."""
-
-    __tablename__ = "bill"
-
-    congress: Mapped[int]
-    bill_type: Mapped[str]
-    number: Mapped[int]
-
-    title: Mapped[str | None]
-    title_short: Mapped[str | None]
-    official_title: Mapped[str | None]
-
-    status: Mapped[str | None]
-    status_at: Mapped[date | None]
-
-    sponsor_bioguide_id: Mapped[str | None]
-
-    subjects_top_term: Mapped[str | None]
-
-    votes: Mapped[list[Vote]] = relationship(
-        "Vote",
-        back_populates="bill",
-    )
-    bill_texts: Mapped[list[BillText]] = relationship(
-        "BillText",
-        back_populates="bill",
-        cascade="all, delete-orphan",
-    )
-
-    __table_args__ = (
-        UniqueConstraint("congress", "bill_type", "number", name="uq_bill_congress_type_number"),
-        Index("ix_bill_congress", "congress"),
-    )
-
-
-class BillText(DataScienceDevTableBase):
-    """Stores different text versions of a bill (introduced, enrolled, etc.)."""
-
-    __tablename__ = "bill_text"
-
-    bill_id: Mapped[int] = mapped_column(ForeignKey("main.bill.id", ondelete="CASCADE"))
-    version_code: Mapped[str]
-    version_name: Mapped[str | None]
-    text_content: Mapped[str | None]
-    date: Mapped[date | None]
-
-    bill: Mapped[Bill] = relationship("Bill", back_populates="bill_texts")
-
-    __table_args__ = (UniqueConstraint("bill_id", "version_code", name="uq_bill_text_bill_id_version_code"),)

python/orm/data_science_dev/congress/legislator.py

@@ -1,66 +0,0 @@
-"""Legislator model - members of Congress."""
-
-from __future__ import annotations
-
-from datetime import date
-from typing import TYPE_CHECKING
-
-from sqlalchemy import ForeignKey, Text
-from sqlalchemy.orm import Mapped, mapped_column, relationship
-
-from python.orm.data_science_dev.base import DataScienceDevTableBase
-
-if TYPE_CHECKING:
-    from python.orm.data_science_dev.congress.vote import VoteRecord
-
-
-class Legislator(DataScienceDevTableBase):
-    """Members of Congress with identification and current term info."""
-
-    __tablename__ = "legislator"
-
-    bioguide_id: Mapped[str] = mapped_column(Text, unique=True, index=True)
-
-    thomas_id: Mapped[str | None]
-    lis_id: Mapped[str | None]
-    govtrack_id: Mapped[int | None]
-    opensecrets_id: Mapped[str | None]
-    fec_ids: Mapped[str | None]
-
-    first_name: Mapped[str]
-    last_name: Mapped[str]
-    official_full_name: Mapped[str | None]
-    nickname: Mapped[str | None]
-
-    birthday: Mapped[date | None]
-    gender: Mapped[str | None]
-
-    current_party: Mapped[str | None]
-    current_state: Mapped[str | None]
-    current_district: Mapped[int | None]
-    current_chamber: Mapped[str | None]
-
-    social_media_accounts: Mapped[list[LegislatorSocialMedia]] = relationship(
-        "LegislatorSocialMedia",
-        back_populates="legislator",
-        cascade="all, delete-orphan",
-    )
-    vote_records: Mapped[list[VoteRecord]] = relationship(
-        "VoteRecord",
-        back_populates="legislator",
-        cascade="all, delete-orphan",
-    )
-
-
-class LegislatorSocialMedia(DataScienceDevTableBase):
-    """Social media account linked to a legislator."""
-
-    __tablename__ = "legislator_social_media"
-
-    legislator_id: Mapped[int] = mapped_column(ForeignKey("main.legislator.id"))
-    platform: Mapped[str]
-    account_name: Mapped[str]
-    url: Mapped[str | None]
-    source: Mapped[str]
-
-    legislator: Mapped[Legislator] = relationship(back_populates="social_media_accounts")

python/orm/data_science_dev/congress/vote.py

@@ -1,79 +0,0 @@
-"""Vote model - roll call votes in Congress."""
-
-from __future__ import annotations
-
-from datetime import date
-from typing import TYPE_CHECKING
-
-from sqlalchemy import ForeignKey, Index, UniqueConstraint
-from sqlalchemy.orm import Mapped, mapped_column, relationship
-
-from python.orm.data_science_dev.base import DataScienceDevBase, DataScienceDevTableBase
-
-if TYPE_CHECKING:
-    from python.orm.data_science_dev.congress.bill import Bill
-    from python.orm.data_science_dev.congress.legislator import Legislator
-    from python.orm.data_science_dev.congress.vote import Vote
-
-
-class VoteRecord(DataScienceDevBase):
-    """Links a vote to a legislator with their position (Yea, Nay, etc.)."""
-
-    __tablename__ = "vote_record"
-
-    vote_id: Mapped[int] = mapped_column(
-        ForeignKey("main.vote.id", ondelete="CASCADE"),
-        primary_key=True,
-    )
-    legislator_id: Mapped[int] = mapped_column(
-        ForeignKey("main.legislator.id", ondelete="CASCADE"),
-        primary_key=True,
-    )
-    position: Mapped[str]
-
-    vote: Mapped[Vote] = relationship("Vote", back_populates="vote_records")
-    legislator: Mapped[Legislator] = relationship("Legislator", back_populates="vote_records")
-
-
-class Vote(DataScienceDevTableBase):
-    """Roll call votes with counts and optional bill linkage."""
-
-    __tablename__ = "vote"
-
-    congress: Mapped[int]
-    chamber: Mapped[str]
-    session: Mapped[int]
-    number: Mapped[int]
-
-    vote_type: Mapped[str | None]
-    question: Mapped[str | None]
-    result: Mapped[str | None]
-    result_text: Mapped[str | None]
-
-    vote_date: Mapped[date]
-
-    yea_count: Mapped[int | None]
-    nay_count: Mapped[int | None]
-    not_voting_count: Mapped[int | None]
-    present_count: Mapped[int | None]
-
-    bill_id: Mapped[int | None] = mapped_column(ForeignKey("main.bill.id"))
-
-    bill: Mapped[Bill | None] = relationship("Bill", back_populates="votes")
-    vote_records: Mapped[list[VoteRecord]] = relationship(
-        "VoteRecord",
-        back_populates="vote",
-        cascade="all, delete-orphan",
-    )
-
-    __table_args__ = (
-        UniqueConstraint(
-            "congress",
-            "chamber",
-            "session",
-            "number",
-            name="uq_vote_congress_chamber_session_number",
-        ),
-        Index("ix_vote_date", "vote_date"),
-        Index("ix_vote_congress_chamber", "congress", "chamber"),
-    )

python/orm/data_science_dev/models.py

@@ -1,16 +0,0 @@
-"""Data science dev database ORM models."""
-
-from __future__ import annotations
-
-from python.orm.data_science_dev.congress import Bill, BillText, Legislator, Vote, VoteRecord
-from python.orm.data_science_dev.posts import partitions  # noqa: F401 — registers partition classes in metadata
-from python.orm.data_science_dev.posts.tables import Posts
-
-__all__ = [
-    "Bill",
-    "BillText",
-    "Legislator",
-    "Posts",
-    "Vote",
-    "VoteRecord",
-]

python/orm/data_science_dev/posts/__init__.py

@@ -1,11 +0,0 @@
-"""Posts module — weekly-partitioned posts table and partition ORM models."""
-
-from __future__ import annotations
-
-from python.orm.data_science_dev.posts.failed_ingestion import FailedIngestion
-from python.orm.data_science_dev.posts.tables import Posts
-
-__all__ = [
-    "FailedIngestion",
-    "Posts",
-]

python/orm/data_science_dev/posts/columns.py

@@ -1,33 +0,0 @@
-"""Shared column definitions for the posts partitioned table family."""
-
-from __future__ import annotations
-
-from datetime import datetime
-
-from sqlalchemy import BigInteger, SmallInteger, Text
-from sqlalchemy.orm import Mapped, mapped_column
-
-
-class PostsColumns:
-    """Mixin providing all posts columns. Used by both the parent table and partitions."""
-
-    post_id: Mapped[int] = mapped_column(BigInteger, primary_key=True)
-    user_id: Mapped[int] = mapped_column(BigInteger)
-    instance: Mapped[str]
-    date: Mapped[datetime] = mapped_column(primary_key=True)
-    text: Mapped[str] = mapped_column(Text)
-    langs: Mapped[str | None]
-    like_count: Mapped[int]
-    reply_count: Mapped[int]
-    repost_count: Mapped[int]
-    reply_to: Mapped[int | None] = mapped_column(BigInteger)
-    replied_author: Mapped[int | None] = mapped_column(BigInteger)
-    thread_root: Mapped[int | None] = mapped_column(BigInteger)
-    thread_root_author: Mapped[int | None] = mapped_column(BigInteger)
-    repost_from: Mapped[int | None] = mapped_column(BigInteger)
-    reposted_author: Mapped[int | None] = mapped_column(BigInteger)
-    quotes: Mapped[int | None] = mapped_column(BigInteger)
-    quoted_author: Mapped[int | None] = mapped_column(BigInteger)
-    labels: Mapped[str | None]
-    sent_label: Mapped[int | None] = mapped_column(SmallInteger)
-    sent_score: Mapped[float | None]

python/orm/data_science_dev/posts/failed_ingestion.py

@@ -1,17 +0,0 @@
-"""Table for storing JSONL lines that failed during post ingestion."""
-
-from __future__ import annotations
-
-from sqlalchemy import Text
-from sqlalchemy.orm import Mapped, mapped_column
-
-from python.orm.data_science_dev.base import DataScienceDevTableBase
-
-
-class FailedIngestion(DataScienceDevTableBase):
-    """Stores raw JSONL lines and their error messages when ingestion fails."""
-
-    __tablename__ = "failed_ingestion"
-
-    raw_line: Mapped[str] = mapped_column(Text)
-    error: Mapped[str] = mapped_column(Text)

python/orm/data_science_dev/posts/partitions.py

@@ -1,71 +0,0 @@
-"""Dynamically generated ORM classes for each weekly partition of the posts table.
-
-Each class maps to a PostgreSQL partition table (e.g. posts_2024_01).
-These are real ORM models tracked by Alembic autogenerate.
-
-Uses ISO week numbering (datetime.isocalendar().week). ISO years can have
-52 or 53 weeks, and week boundaries are always Monday to Monday.
-"""
-
-from __future__ import annotations
-
-import sys
-from datetime import UTC, datetime
-
-from python.orm.data_science_dev.base import DataScienceDevBase
-from python.orm.data_science_dev.posts.columns import PostsColumns
-
-PARTITION_START_YEAR = 2023
-PARTITION_END_YEAR = 2026
-
-_current_module = sys.modules[__name__]
-
-
-def iso_weeks_in_year(year: int) -> int:
-    """Return the number of ISO weeks in a given year (52 or 53)."""
-    dec_28 = datetime(year, 12, 28, tzinfo=UTC)
-    return dec_28.isocalendar().week
-
-
-def week_bounds(year: int, week: int) -> tuple[datetime, datetime]:
-    """Return (start, end) datetimes for an ISO week.
-
-    Start = Monday 00:00:00 UTC of the given ISO week.
-    End   = Monday 00:00:00 UTC of the following ISO week.
-    """
-    start = datetime.fromisocalendar(year, week, 1).replace(tzinfo=UTC)
-    if week < iso_weeks_in_year(year):
-        end = datetime.fromisocalendar(year, week + 1, 1).replace(tzinfo=UTC)
-    else:
-        end = datetime.fromisocalendar(year + 1, 1, 1).replace(tzinfo=UTC)
-    return start, end
-
-
-def _build_partition_classes() -> dict[str, type]:
-    """Generate one ORM class per ISO week partition."""
-    classes: dict[str, type] = {}
-
-    for year in range(PARTITION_START_YEAR, PARTITION_END_YEAR + 1):
-        for week in range(1, iso_weeks_in_year(year) + 1):
-            class_name = f"PostsWeek{year}W{week:02d}"
-            table_name = f"posts_{year}_{week:02d}"
-
-            partition_class = type(
-                class_name,
-                (PostsColumns, DataScienceDevBase),
-                {
-                    "__tablename__": table_name,
-                    "__table_args__": ({"implicit_returning": False},),
-                },
-            )
-
-            classes[class_name] = partition_class
-
-    return classes
-
-
-# Generate all partition classes and register them on this module
-_partition_classes = _build_partition_classes()
-for _name, _cls in _partition_classes.items():
-    setattr(_current_module, _name, _cls)
-__all__ = list(_partition_classes.keys())

python/orm/data_science_dev/posts/tables.py

@@ -1,13 +0,0 @@
-"""Posts parent table with PostgreSQL weekly range partitioning on date column."""
-
-from __future__ import annotations
-
-from python.orm.data_science_dev.base import DataScienceDevBase
-from python.orm.data_science_dev.posts.columns import PostsColumns
-
-
-class Posts(PostsColumns, DataScienceDevBase):
-    """Parent partitioned table for posts, partitioned by week on `date`."""
-
-    __tablename__ = "posts"
-    __table_args__ = ({"postgresql_partition_by": "RANGE (date)"},)

python/orm/richie/__init__.py

@@ -2,7 +2,8 @@
 
 from __future__ import annotations
 
-from python.orm.richie.base import RichieBase, TableBase, TableBaseBig, TableBaseSmall
+from python.orm.richie.base import RichieBase, TableBase
+from python.orm.richie.congress import Bill, Legislator, Vote, VoteRecord
 from python.orm.richie.contact import (
     Contact,
     ContactNeed,
@@ -10,15 +11,21 @@ from python.orm.richie.contact import (
     Need,
     RelationshipType,
 )
+from python.orm.richie.dead_letter_message import DeadLetterMessage
+from python.orm.richie.signal_device import SignalDevice
 
 __all__ = [
+    "Bill",
     "Contact",
     "ContactNeed",
     "ContactRelationship",
+    "DeadLetterMessage",
+    "Legislator",
     "Need",
     "RelationshipType",
     "RichieBase",
+    "SignalDevice",
     "TableBase",
-    "TableBaseBig",
+    "Vote",
-    "TableBaseSmall",
+    "VoteRecord",
 ]

python/orm/richie/base.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 
 from datetime import datetime
 
-from sqlalchemy import BigInteger, DateTime, MetaData, SmallInteger, func
+from sqlalchemy import DateTime, MetaData, func
 from sqlalchemy.ext.declarative import AbstractConcreteBase
 from sqlalchemy.orm import DeclarativeBase, Mapped, mapped_column
 
@@ -22,9 +22,12 @@ class RichieBase(DeclarativeBase):
     )
 
 
-class _TableMixin:
+class TableBase(AbstractConcreteBase, RichieBase):
-    """Shared timestamp columns for all table bases."""
+    """Abstract concrete base for richie tables with IDs and timestamps."""
 
+    __abstract__ = True
+
+    id: Mapped[int] = mapped_column(primary_key=True)
     created: Mapped[datetime] = mapped_column(
         DateTime(timezone=True),
         server_default=func.now(),
@@ -34,27 +37,3 @@ class _TableMixin:
         server_default=func.now(),
         onupdate=func.now(),
     )
-
-
-class TableBaseSmall(_TableMixin, AbstractConcreteBase, RichieBase):
-    """Table with SmallInteger primary key."""
-
-    __abstract__ = True
-
-    id: Mapped[int] = mapped_column(SmallInteger, primary_key=True)
-
-
-class TableBase(_TableMixin, AbstractConcreteBase, RichieBase):
-    """Table with Integer primary key."""
-
-    __abstract__ = True
-
-    id: Mapped[int] = mapped_column(primary_key=True)
-
-
-class TableBaseBig(_TableMixin, AbstractConcreteBase, RichieBase):
-    """Table with BigInteger primary key."""
-
-    __abstract__ = True
-
-    id: Mapped[int] = mapped_column(BigInteger, primary_key=True)

python/orm/richie/congress.py

@@ -0,0 +1,150 @@
+"""Congress Tracker database models."""
+
+from __future__ import annotations
+
+from datetime import date
+
+from sqlalchemy import ForeignKey, Index, Text, UniqueConstraint
+from sqlalchemy.orm import Mapped, mapped_column, relationship
+
+from python.orm.richie.base import RichieBase, TableBase
+
+
+class Legislator(TableBase):
+    """Legislator model - members of Congress."""
+
+    __tablename__ = "legislator"
+
+    # Natural key - bioguide ID is the authoritative identifier
+    bioguide_id: Mapped[str] = mapped_column(Text, unique=True, index=True)
+
+    # Other IDs for cross-referencing
+    thomas_id: Mapped[str | None]
+    lis_id: Mapped[str | None]
+    govtrack_id: Mapped[int | None]
+    opensecrets_id: Mapped[str | None]
+    fec_ids: Mapped[str | None]  # JSON array stored as string
+
+    # Name info
+    first_name: Mapped[str]
+    last_name: Mapped[str]
+    official_full_name: Mapped[str | None]
+    nickname: Mapped[str | None]
+
+    # Bio
+    birthday: Mapped[date | None]
+    gender: Mapped[str | None]  # M/F
+
+    # Current term info (denormalized for query efficiency)
+    current_party: Mapped[str | None]
+    current_state: Mapped[str | None]
+    current_district: Mapped[int | None]  # House only
+    current_chamber: Mapped[str | None]  # rep/sen
+
+    # Relationships
+    vote_records: Mapped[list[VoteRecord]] = relationship(
+        "VoteRecord",
+        back_populates="legislator",
+        cascade="all, delete-orphan",
+    )
+
+
+class Bill(TableBase):
+    """Bill model - legislation introduced in Congress."""
+
+    __tablename__ = "bill"
+
+    # Composite natural key: congress + bill_type + number
+    congress: Mapped[int]
+    bill_type: Mapped[str]  # hr, s, hres, sres, hjres, sjres
+    number: Mapped[int]
+
+    # Bill info
+    title: Mapped[str | None]
+    title_short: Mapped[str | None]
+    official_title: Mapped[str | None]
+
+    # Status
+    status: Mapped[str | None]
+    status_at: Mapped[date | None]
+
+    # Sponsor
+    sponsor_bioguide_id: Mapped[str | None]
+
+    # Subjects
+    subjects_top_term: Mapped[str | None]
+
+    # Relationships
+    votes: Mapped[list[Vote]] = relationship(
+        "Vote",
+        back_populates="bill",
+    )
+
+    __table_args__ = (
+        UniqueConstraint("congress", "bill_type", "number", name="uq_bill_congress_type_number"),
+        Index("ix_bill_congress", "congress"),
+    )
+
+
+class Vote(TableBase):
+    """Vote model - roll call votes in Congress."""
+
+    __tablename__ = "vote"
+
+    # Composite natural key: congress + chamber + session + number
+    congress: Mapped[int]
+    chamber: Mapped[str]  # house/senate
+    session: Mapped[int]
+    number: Mapped[int]
+
+    # Vote details
+    vote_type: Mapped[str | None]
+    question: Mapped[str | None]
+    result: Mapped[str | None]
+    result_text: Mapped[str | None]
+
+    # Timing
+    vote_date: Mapped[date]
+
+    # Vote counts (denormalized for efficiency)
+    yea_count: Mapped[int | None]
+    nay_count: Mapped[int | None]
+    not_voting_count: Mapped[int | None]
+    present_count: Mapped[int | None]
+
+    # Related bill (optional - not all votes are on bills)
+    bill_id: Mapped[int | None] = mapped_column(ForeignKey("main.bill.id"))
+
+    # Relationships
+    bill: Mapped[Bill | None] = relationship("Bill", back_populates="votes")
+    vote_records: Mapped[list[VoteRecord]] = relationship(
+        "VoteRecord",
+        back_populates="vote",
+        cascade="all, delete-orphan",
+    )
+
+    __table_args__ = (
+        UniqueConstraint("congress", "chamber", "session", "number", name="uq_vote_congress_chamber_session_number"),
+        Index("ix_vote_date", "vote_date"),
+        Index("ix_vote_congress_chamber", "congress", "chamber"),
+    )
+
+
+class VoteRecord(RichieBase):
+    """Association table: Vote <-> Legislator with position."""
+
+    __tablename__ = "vote_record"
+
+    vote_id: Mapped[int] = mapped_column(
+        ForeignKey("main.vote.id", ondelete="CASCADE"),
+        primary_key=True,
+    )
+    legislator_id: Mapped[int] = mapped_column(
+        ForeignKey("main.legislator.id", ondelete="CASCADE"),
+        primary_key=True,
+    )
+    position: Mapped[str]  # Yea, Nay, Not Voting, Present
+
+    # Relationships
+    vote: Mapped[Vote] = relationship("Vote", back_populates="vote_records")
+    legislator: Mapped[Legislator] = relationship("Legislator", back_populates="vote_records")

python/orm/richie/dead_letter_message.py

@@ -0,0 +1,26 @@
+"""Dead letter queue for Signal bot messages that fail processing."""
+
+from __future__ import annotations
+
+from datetime import datetime
+
+from sqlalchemy import DateTime, Text
+from sqlalchemy.dialects.postgresql import ENUM
+from sqlalchemy.orm import Mapped, mapped_column
+
+from python.orm.richie.base import TableBase
+from python.signal_bot.models import MessageStatus
+
+
+class DeadLetterMessage(TableBase):
+    """A Signal message that failed processing and was sent to the dead letter queue."""
+
+    __tablename__ = "dead_letter_message"
+
+    source: Mapped[str]
+    message: Mapped[str] = mapped_column(Text)
+    received_at: Mapped[datetime] = mapped_column(DateTime(timezone=True))
+    status: Mapped[MessageStatus] = mapped_column(
+        ENUM(MessageStatus, name="message_status", create_type=True, schema="main"),
+        default=MessageStatus.UNPROCESSED,
+    )

python/orm/richie/signal_device.py

@@ -0,0 +1,26 @@
+"""Signal bot device registry models."""
+
+from __future__ import annotations
+
+from datetime import datetime
+
+from sqlalchemy import DateTime, String
+from sqlalchemy.dialects.postgresql import ENUM
+from sqlalchemy.orm import Mapped, mapped_column
+
+from python.orm.richie.base import TableBase
+from python.signal_bot.models import TrustLevel
+
+
+class SignalDevice(TableBase):
+    """A Signal device tracked by phone number and safety number."""
+
+    __tablename__ = "signal_device"
+
+    phone_number: Mapped[str] = mapped_column(String(50), unique=True)
+    safety_number: Mapped[str | None]
+    trust_level: Mapped[TrustLevel] = mapped_column(
+        ENUM(TrustLevel, name="trust_level", create_type=True, schema="main"),
+        default=TrustLevel.UNVERIFIED,
+    )
+    last_seen: Mapped[datetime] = mapped_column(DateTime(timezone=True))

python/orm/signal_bot/__init__.py

@@ -1,16 +0,0 @@
-"""Signal bot database ORM exports."""
-
-from __future__ import annotations
-
-from python.orm.signal_bot.base import SignalBotBase, SignalBotTableBase, SignalBotTableBaseSmall
-from python.orm.signal_bot.models import DeadLetterMessage, DeviceRole, RoleRecord, SignalDevice
-
-__all__ = [
-    "DeadLetterMessage",
-    "DeviceRole",
-    "RoleRecord",
-    "SignalBotBase",
-    "SignalBotTableBase",
-    "SignalBotTableBaseSmall",
-    "SignalDevice",
-]

python/orm/signal_bot/base.py

@@ -1,52 +0,0 @@
-"""Signal bot database ORM base."""
-
-from __future__ import annotations
-
-from datetime import datetime
-
-from sqlalchemy import DateTime, MetaData, SmallInteger, func
-from sqlalchemy.ext.declarative import AbstractConcreteBase
-from sqlalchemy.orm import DeclarativeBase, Mapped, mapped_column
-
-from python.orm.common import NAMING_CONVENTION
-
-
-class SignalBotBase(DeclarativeBase):
-    """Base class for signal_bot database ORM models."""
-
-    schema_name = "main"
-
-    metadata = MetaData(
-        schema=schema_name,
-        naming_convention=NAMING_CONVENTION,
-    )
-
-
-class _TableMixin:
-    """Shared timestamp columns for all table bases."""
-
-    created: Mapped[datetime] = mapped_column(
-        DateTime(timezone=True),
-        server_default=func.now(),
-    )
-    updated: Mapped[datetime] = mapped_column(
-        DateTime(timezone=True),
-        server_default=func.now(),
-        onupdate=func.now(),
-    )
-
-
-class SignalBotTableBaseSmall(_TableMixin, AbstractConcreteBase, SignalBotBase):
-    """Table with SmallInteger primary key."""
-
-    __abstract__ = True
-
-    id: Mapped[int] = mapped_column(SmallInteger, primary_key=True)
-
-
-class SignalBotTableBase(_TableMixin, AbstractConcreteBase, SignalBotBase):
-    """Table with Integer primary key."""
-
-    __abstract__ = True
-
-    id: Mapped[int] = mapped_column(primary_key=True)

python/orm/signal_bot/models.py

@@ -1,62 +0,0 @@
-"""Signal bot device, role, and dead letter ORM models."""
-
-from __future__ import annotations
-
-from datetime import datetime
-
-from sqlalchemy import DateTime, Enum, ForeignKey, SmallInteger, String, Text, UniqueConstraint
-from sqlalchemy.orm import Mapped, mapped_column, relationship
-
-from python.orm.signal_bot.base import SignalBotTableBase, SignalBotTableBaseSmall
-from python.signal_bot.models import MessageStatus, TrustLevel
-
-
-class RoleRecord(SignalBotTableBaseSmall):
-    """Lookup table for RBAC roles, keyed by smallint."""
-
-    __tablename__ = "role"
-
-    name: Mapped[str] = mapped_column(String(50), unique=True)
-
-
-class DeviceRole(SignalBotTableBase):
-    """Association between a device and a role."""
-
-    __tablename__ = "device_role"
-    __table_args__ = (
-        UniqueConstraint("device_id", "role_id", name="uq_device_role_device_role"),
-        {"schema": "main"},
-    )
-
-    device_id: Mapped[int] = mapped_column(ForeignKey("main.signal_device.id"))
-    role_id: Mapped[int] = mapped_column(SmallInteger, ForeignKey("main.role.id"))
-
-
-class SignalDevice(SignalBotTableBase):
-    """A Signal device tracked by phone number and safety number."""
-
-    __tablename__ = "signal_device"
-
-    phone_number: Mapped[str] = mapped_column(String(50), unique=True)
-    safety_number: Mapped[str | None]
-    trust_level: Mapped[TrustLevel] = mapped_column(
-        Enum(TrustLevel, name="trust_level", create_constraint=False, native_enum=False),
-        default=TrustLevel.UNVERIFIED,
-    )
-    last_seen: Mapped[datetime] = mapped_column(DateTime(timezone=True))
-
-    roles: Mapped[list[RoleRecord]] = relationship(secondary=DeviceRole.__table__)
-
-
-class DeadLetterMessage(SignalBotTableBase):
-    """A Signal message that failed processing and was sent to the dead letter queue."""
-
-    __tablename__ = "dead_letter_message"
-
-    source: Mapped[str]
-    message: Mapped[str] = mapped_column(Text)
-    received_at: Mapped[datetime] = mapped_column(DateTime(timezone=True))
-    status: Mapped[MessageStatus] = mapped_column(
-        Enum(MessageStatus, name="message_status", create_constraint=False, native_enum=False),
-        default=MessageStatus.UNPROCESSED,
-    )

python/prompt_bench/Dockerfile.finetune

@@ -1,25 +0,0 @@
-# Unsloth fine-tuning container for Qwen 3.5 4B on RTX 3090.
-#
-# Build:
-#   docker build -f python/prompt_bench/Dockerfile.finetune -t bill-finetune .
-#
-# Run:
-#   docker run --rm --device=nvidia.com/gpu=all --ipc=host \
-#     -v $(pwd)/output:/workspace/output \
-#     -v $(pwd)/output/finetune_dataset.jsonl:/workspace/dataset.jsonl:ro \
-#     -v /zfs/models/hf:/models \
-#     bill-finetune \
-#     --dataset /workspace/dataset.jsonl \
-#     --output-dir /workspace/output/qwen-bill-summarizer
-
-FROM ghcr.io/unslothai/unsloth:latest
-
-RUN pip install --no-cache-dir typer
-
-WORKDIR /workspace
-COPY python/prompt_bench/finetune.py python/prompt_bench/finetune.py
-COPY python/prompt_bench/summarization_prompts.py python/prompt_bench/summarization_prompts.py
-COPY python/prompt_bench/__init__.py python/prompt_bench/__init__.py
-COPY python/__init__.py python/__init__.py
-
-ENTRYPOINT ["python", "-m", "python.prompt_bench.finetune"]

python/prompt_bench/__init__.py

@@ -1 +0,0 @@
-"""Prompt benchmarking system for evaluating LLMs via vLLM."""

python/prompt_bench/batch_bill_summarizer.py

@@ -1,233 +0,0 @@
-"""Submit an OpenAI Batch API bill-summarization job over compressed text.
-
-Reads the first N bills from a CSV with a `text_content` column, compresses
-each via `bill_token_compression.compress_bill_text`, builds a JSONL file of
-summarization requests, and submits it as an asynchronous Batch API job
-against `/v1/chat/completions`. Also writes a CSV of per-bill pre/post-
-compression token counts.
-"""
-
-from __future__ import annotations
-
-import csv
-import json
-import logging
-import re
-import sys
-from os import getenv
-from pathlib import Path
-from typing import Annotated
-
-import httpx
-import typer
-from tiktoken import Encoding, get_encoding
-
-from python.prompt_bench.bill_token_compression import compress_bill_text
-from python.prompt_bench.summarization_prompts import SUMMARIZATION_SYSTEM_PROMPT, SUMMARIZATION_USER_TEMPLATE
-
-logger = logging.getLogger(__name__)
-
-OPENAI_API_BASE = "https://api.openai.com/v1"
-
-
-def load_bills(csv_path: Path, count: int = 0) -> list[tuple[str, str]]:
-    """Return (bill_id, text_content) tuples with non-empty text.
-
-    If `count` is 0 or negative, all rows are returned.
-    """
-    csv.field_size_limit(sys.maxsize)
-    bills: list[tuple[str, str]] = []
-    with csv_path.open(newline="", encoding="utf-8") as handle:
-        reader = csv.DictReader(handle)
-        for row in reader:
-            text_content = (row.get("text_content") or "").strip()
-            if not text_content:
-                continue
-            bill_id = row.get("bill_id") or row.get("id") or f"row-{len(bills)}"
-            version_code = row.get("version_code") or ""
-            unique_id = f"{bill_id}-{version_code}" if version_code else bill_id
-            bills.append((unique_id, text_content))
-            if count > 0 and len(bills) >= count:
-                break
-    return bills
-
-
-def safe_filename(value: str) -> str:
-    """Make a string safe for use as a filename or batch custom_id."""
-    return re.sub(r"[^A-Za-z0-9._-]+", "_", value).strip("_") or "unnamed"
-
-
-def build_request(custom_id: str, model: str, bill_text: str) -> dict:
-    """Build one OpenAI batch request line."""
-    return {
-        "custom_id": custom_id,
-        "method": "POST",
-        "url": "/v1/chat/completions",
-        "body": {
-            "model": model,
-            "messages": [
-                {"role": "system", "content": SUMMARIZATION_SYSTEM_PROMPT},
-                {"role": "user", "content": SUMMARIZATION_USER_TEMPLATE.format(text_content=bill_text)},
-            ],
-        },
-    }
-
-
-def write_jsonl(path: Path, lines: list[dict]) -> None:
-    """Write a list of dicts as JSONL."""
-    with path.open("w", encoding="utf-8") as handle:
-        for line in lines:
-            handle.write(json.dumps(line, ensure_ascii=False))
-            handle.write("\n")
-
-
-def upload_file(client: httpx.Client, path: Path) -> str:
-    """Upload a JSONL file to the OpenAI Files API and return its file id."""
-    with path.open("rb") as handle:
-        response = client.post(
-            f"{OPENAI_API_BASE}/files",
-            files={"file": (path.name, handle, "application/jsonl")},
-            data={"purpose": "batch"},
-        )
-    response.raise_for_status()
-    return response.json()["id"]
-
-
-def prepare_requests(
-    bills: list[tuple[str, str]],
-    *,
-    model: str,
-    encoder: Encoding,
-) -> tuple[list[dict], list[dict]]:
-    """Build (request_lines, token_rows) from bills.
-
-    Each bill is compressed before being turned into a request line.
-    Each `token_rows` entry has chars + token counts for one bill so the caller
-    can write a per-bill CSV.
-    """
-    request_lines: list[dict] = []
-    token_rows: list[dict] = []
-    for bill_id, text_content in bills:
-        raw_token_count = len(encoder.encode(text_content))
-        compressed_text = compress_bill_text(text_content)
-        compressed_token_count = len(encoder.encode(compressed_text))
-        token_rows.append(
-            {
-                "bill_id": bill_id,
-                "raw_chars": len(text_content),
-                "compressed_chars": len(compressed_text),
-                "raw_tokens": raw_token_count,
-                "compressed_tokens": compressed_token_count,
-                "token_ratio": (compressed_token_count / raw_token_count) if raw_token_count else None,
-            },
-        )
-        safe_id = safe_filename(bill_id)
-        request_lines.append(build_request(safe_id, model, compressed_text))
-    return request_lines, token_rows
-
-
-def write_token_csv(path: Path, token_rows: list[dict]) -> tuple[int, int]:
-    """Write per-bill token counts to CSV. Returns (raw_total, compressed_total)."""
-    with path.open("w", newline="", encoding="utf-8") as handle:
-        writer = csv.DictWriter(
-            handle,
-            fieldnames=["bill_id", "raw_chars", "compressed_chars", "raw_tokens", "compressed_tokens", "token_ratio"],
-        )
-        writer.writeheader()
-        writer.writerows(token_rows)
-    raw_total = sum(row["raw_tokens"] for row in token_rows)
-    compressed_total = sum(row["compressed_tokens"] for row in token_rows)
-    return raw_total, compressed_total
-
-
-def create_batch(client: httpx.Client, input_file_id: str, description: str) -> dict:
-    """Create a batch job and return its full response payload."""
-    response = client.post(
-        f"{OPENAI_API_BASE}/batches",
-        json={
-            "input_file_id": input_file_id,
-            "endpoint": "/v1/chat/completions",
-            "completion_window": "24h",
-            "metadata": {"description": description},
-        },
-    )
-    response.raise_for_status()
-    return response.json()
-
-
-def main(
-    csv_path: Annotated[Path, typer.Option("--csv", help="Bills CSV path")] = Path("bills.csv"),
-    output_dir: Annotated[Path, typer.Option("--output-dir", help="Where to write JSONL + metadata")] = Path(
-        "output/openai_batch",
-    ),
-    model: Annotated[str, typer.Option(help="OpenAI model id")] = "gpt-5-mini",
-    count: Annotated[int, typer.Option(help="Max bills to process, 0 = all")] = 0,
-    log_level: Annotated[str, typer.Option(help="Log level")] = "INFO",
-) -> None:
-    """Submit an OpenAI Batch job of compressed bill summaries."""
-    logging.basicConfig(level=log_level, format="%(asctime)s %(levelname)s %(name)s: %(message)s")
-
-    api_key = getenv("CLOSEDAI_TOKEN") or getenv("OPENAI_API_KEY")
-    if not api_key:
-        message = "Neither CLOSEDAI_TOKEN nor OPENAI_API_KEY is set"
-        raise typer.BadParameter(message)
-    if not csv_path.is_file():
-        message = f"CSV not found: {csv_path}"
-        raise typer.BadParameter(message)
-
-    output_dir.mkdir(parents=True, exist_ok=True)
-
-    logger.info("Loading %d bills from %s", count, csv_path)
-    bills = load_bills(csv_path, count)
-    if len(bills) < count:
-        logger.warning("Only %d bills available (requested %d)", len(bills), count)
-
-    encoder = get_encoding("o200k_base")
-    request_lines, token_rows = prepare_requests(bills, model=model, encoder=encoder)
-
-    token_csv_path = output_dir / "token_counts.csv"
-    raw_tokens_total, compressed_tokens_total = write_token_csv(token_csv_path, token_rows)
-    logger.info(
-        "Token counts: raw=%d compressed=%d ratio=%.3f -> %s",
-        raw_tokens_total,
-        compressed_tokens_total,
-        (compressed_tokens_total / raw_tokens_total) if raw_tokens_total else 0.0,
-        token_csv_path,
-    )
-
-    jsonl_path = output_dir / "requests.jsonl"
-    write_jsonl(jsonl_path, request_lines)
-    logger.info("Wrote %s (%d bills)", jsonl_path, len(request_lines))
-
-    headers = {"Authorization": f"Bearer {api_key}"}
-    with httpx.Client(headers=headers, timeout=httpx.Timeout(300.0)) as client:
-        logger.info("Uploading JSONL")
-        file_id = upload_file(client, jsonl_path)
-        logger.info("Uploaded: %s", file_id)
-
-        logger.info("Creating batch")
-        batch = create_batch(client, file_id, f"compressed bill summaries x{len(request_lines)} ({model})")
-        logger.info("Batch created: %s", batch["id"])
-
-    metadata = {
-        "model": model,
-        "count": len(bills),
-        "jsonl": str(jsonl_path),
-        "input_file_id": file_id,
-        "batch_id": batch["id"],
-        "raw_tokens_total": raw_tokens_total,
-        "compressed_tokens_total": compressed_tokens_total,
-        "batch": batch,
-    }
-    metadata_path = output_dir / "batch.json"
-    metadata_path.write_text(json.dumps(metadata, indent=2))
-    logger.info("Wrote metadata to %s", metadata_path)
-
-
-def cli() -> None:
-    """Typer entry point."""
-    typer.run(main)
-
-
-if __name__ == "__main__":
-    cli()

python/prompt_bench/bill_token_compression.py

@@ -1,162 +0,0 @@
-"""Lossless-ish text compression for Congressional bill text."""
-
-from __future__ import annotations
-
-import re
-
-STATES = (
-    "Alabama",
-    "Alaska",
-    "Arizona",
-    "Arkansas",
-    "California",
-    "Colorado",
-    "Connecticut",
-    "Delaware",
-    "Florida",
-    "Georgia",
-    "Hawaii",
-    "Idaho",
-    "Illinois",
-    "Indiana",
-    "Iowa",
-    "Kansas",
-    "Kentucky",
-    "Louisiana",
-    "Maine",
-    "Maryland",
-    "Massachusetts",
-    "Michigan",
-    "Minnesota",
-    "Mississippi",
-    "Missouri",
-    "Montana",
-    "Nebraska",
-    "Nevada",
-    "New Hampshire",
-    "New Jersey",
-    "New Mexico",
-    "New York",
-    "North Carolina",
-    "North Dakota",
-    "Ohio",
-    "Oklahoma",
-    "Oregon",
-    "Pennsylvania",
-    "Rhode Island",
-    "South Carolina",
-    "South Dakota",
-    "Tennessee",
-    "Texas",
-    "Utah",
-    "Vermont",
-    "Virginia",
-    "Washington",
-    "West Virginia",
-    "Wisconsin",
-    "Wyoming",
-    "Puerto Rico",
-    "Guam",
-    "American Samoa",
-    "District of Columbia",
-    "US Virgin Islands",
-)
-STATE_PATTERNS = [(re.compile(re.escape(state), re.IGNORECASE), state) for state in STATES]
-
-
-def normalize_state_names(text: str) -> str:
-    """Replace any casing of state names with title case."""
-    for pattern, replacement in STATE_PATTERNS:
-        text = pattern.sub(replacement, text)
-    return text
-
-
-def strip_number_commas(text: str) -> str:
-    """Remove commas from numeric thousands separators."""
-    return re.sub(r"(\d{1,3}(?:,\d{3})+)", lambda match: match.group().replace(",", ""), text)
-
-
-def strip_horizontal_rules(text: str) -> str:
-    """Remove ASCII horizontal-rule lines built from underscores, dashes, equals, or asterisks."""
-    return re.sub(r"^\s*[_\-=\*]{3,}\s*$", "", text, flags=re.MULTILINE)
-
-
-def collapse_double_dashes(text: str) -> str:
-    """Replace ``--`` em-dash stand-ins with a single space so they don't tokenize oddly."""
-    return text.replace("--", " ")
-
-
-def collapse_inline_whitespace(text: str) -> str:
-    """Collapse runs of horizontal whitespace (spaces, tabs) into a single space, leaving newlines intact."""
-    return re.sub(r"[^\S\n]+", " ", text)
-
-
-def collapse_blank_lines(text: str) -> str:
-    """Collapse three-or-more consecutive newlines down to a blank-line separator."""
-    return re.sub(r"\n{3,}", "\n\n", text)
-
-
-def trim_line_edges(text: str) -> str:
-    """Strip spaces immediately before and after newline characters on every line."""
-    text = re.sub(r" +\n", "\n", text)
-    return re.sub(r"\n +", "\n", text)
-
-
-def shorten_section_markers(text: str) -> str:
-    """Rewrite ``Sec. 12.`` style section headings as the more compact ``SEC 12``."""
-    return re.sub(r"(?i)sec\.\s*(\d+[a-zA-Z]?)\.", r"SEC \1", text)
-
-
-def unwrap_parens(text: str) -> str:
-    """Strip parentheses around short alphanumeric labels like ``(a)`` or ``(12)``."""
-    return re.sub(r"\(([a-zA-Z0-9]+)\)", r"\1", text)
-
-
-def strip_typeset_quotes(text: str) -> str:
-    """Remove the `` and '' typeset quote markers used in the GPO bill format."""
-    return text.replace("``", "").replace("''", "")
-
-
-def normalize_usc_acronym(text: str) -> str:
-    """Collapse ``U.S.C.`` to ``USC`` to save tokens on the common citation."""
-    return text.replace("U.S.C.", "USC")
-
-
-def normalize_us_acronym(text: str) -> str:
-    """Normalize the various ``U.S.``/``U. S.`` spellings to the bare ``US`` form."""
-    for acronym in ("U. S.", "u. s.", "U.S. ", "u.s. "):
-        text = text.replace(acronym, "US ")
-    return text
-
-
-def collapse_ellipses(text: str) -> str:
-    """Collapse runs of two-or-more periods (``...``, ``....``) down to a single period."""
-    return re.sub(r"\.{2,}", ".", text)
-
-
-COMPRESSION_STEPS = (
-    strip_horizontal_rules,
-    collapse_double_dashes,
-    collapse_inline_whitespace,
-    collapse_blank_lines,
-    trim_line_edges,
-    shorten_section_markers,
-    unwrap_parens,
-    strip_typeset_quotes,
-    normalize_usc_acronym,
-    normalize_us_acronym,
-    strip_number_commas,
-    collapse_ellipses,
-    normalize_state_names,
-)
-
-
-def compress_bill_text(text: str) -> str:
-    """Apply lossless-ish whitespace and boilerplate compression to bill text.
-
-    Runs every transform in :data:`COMPRESSION_STEPS` in order, then strips
-    leading/trailing whitespace from the final result.
-    """
-    for step in COMPRESSION_STEPS:
-        text = step(text)
-    return text.strip()

python/prompt_bench/compresion_test.py

@@ -1,236 +0,0 @@
-"""Run two interactive OpenAI chat-completion sweeps over bill text.
-
-Reads the first N bills from a CSV with a `text_content` column and sends two
-sweeps through `/v1/chat/completions` concurrently — one with the raw bill
-text, one with the compressed bill text. Each request's prompt is saved to
-disk alongside the OpenAI response id so the prompts and responses can be
-correlated later.
-"""
-
-from __future__ import annotations
-
-import csv
-import json
-import logging
-import re
-import sys
-import time
-from concurrent.futures import ThreadPoolExecutor, as_completed
-from os import getenv
-from pathlib import Path
-from typing import Annotated
-
-import httpx
-import typer
-
-from python.prompt_bench.bill_token_compression import compress_bill_text
-from python.prompt_bench.summarization_prompts import SUMMARIZATION_SYSTEM_PROMPT, SUMMARIZATION_USER_TEMPLATE
-
-logger = logging.getLogger(__name__)
-
-OPENAI_API_BASE = "https://api.openai.com/v1"
-DEFAULT_MODEL = "gpt-5.4-mini"
-DEFAULT_COUNT = 100
-SEED = 42
-
-
-def load_bills(csv_path: Path, count: int) -> list[tuple[str, str]]:
-    """Return up to `count` (bill_id, text_content) tuples with non-empty text."""
-    csv.field_size_limit(sys.maxsize)
-    bills: list[tuple[str, str]] = []
-    with csv_path.open(newline="", encoding="utf-8") as handle:
-        reader = csv.DictReader(handle)
-        for row in reader:
-            text_content = (row.get("text_content") or "").strip()
-            if not text_content:
-                continue
-            bill_id = row.get("bill_id") or row.get("id") or f"row-{len(bills)}"
-            version_code = row.get("version_code") or ""
-            unique_id = f"{bill_id}-{version_code}" if version_code else bill_id
-            bills.append((unique_id, text_content))
-            if len(bills) >= count:
-                break
-    return bills
-
-
-def build_messages(bill_text: str) -> list[dict]:
-    """Return the system + user message pair for a bill."""
-    return [
-        {"role": "system", "content": SUMMARIZATION_SYSTEM_PROMPT},
-        {"role": "user", "content": SUMMARIZATION_USER_TEMPLATE.format(text_content=bill_text)},
-    ]
-
-
-def safe_filename(value: str) -> str:
-    """Make a string safe for use as a filename."""
-    return re.sub(r"[^A-Za-z0-9._-]+", "_", value).strip("_") or "unnamed"
-
-
-def run_one_request(
-    client: httpx.Client,
-    *,
-    bill_id: str,
-    label: str,
-    bill_text: str,
-    model: str,
-    output_path: Path,
-) -> tuple[bool, float, str | None]:
-    """Send one chat-completion request and persist prompt + response.
-
-    Returns (success, elapsed_seconds, response_id).
-    """
-    messages = build_messages(bill_text)
-    payload = {
-        "model": model,
-        "messages": messages,
-        "seed": SEED,
-    }
-    start = time.monotonic()
-    record: dict = {
-        "bill_id": bill_id,
-        "label": label,
-        "model": model,
-        "seed": SEED,
-        "input_chars": len(bill_text),
-        "messages": messages,
-    }
-    try:
-        response = client.post(f"{OPENAI_API_BASE}/chat/completions", json=payload)
-        response.raise_for_status()
-        body = response.json()
-    except httpx.HTTPStatusError as error:
-        elapsed = time.monotonic() - start
-        record["error"] = {
-            "status_code": error.response.status_code,
-            "body": error.response.text,
-            "elapsed_seconds": elapsed,
-        }
-        output_path.write_text(json.dumps(record, ensure_ascii=False, indent=2))
-        logger.exception("HTTP error for %s/%s after %.2fs", label, bill_id, elapsed)
-        return False, elapsed, None
-    except Exception as error:
-        elapsed = time.monotonic() - start
-        record["error"] = {"message": str(error), "elapsed_seconds": elapsed}
-        output_path.write_text(json.dumps(record, ensure_ascii=False, indent=2))
-        logger.exception("Failed: %s/%s after %.2fs", label, bill_id, elapsed)
-        return False, elapsed, None
-
-    elapsed = time.monotonic() - start
-    response_id = body.get("id")
-    record["response_id"] = response_id
-    record["elapsed_seconds"] = elapsed
-    record["usage"] = body.get("usage")
-    record["response"] = body
-    output_path.write_text(json.dumps(record, ensure_ascii=False, indent=2))
-    logger.info("Done: %s/%s id=%s in %.2fs", label, bill_id, response_id, elapsed)
-    return True, elapsed, response_id
-
-
-def main(
-    csv_path: Annotated[Path, typer.Option("--csv", help="Bills CSV path")] = Path("bills.csv"),
-    output_dir: Annotated[Path, typer.Option("--output-dir", help="Where to write per-request JSON")] = Path(
-        "output/openai_runs",
-    ),
-    model: Annotated[str, typer.Option(help="OpenAI model id")] = DEFAULT_MODEL,
-    count: Annotated[int, typer.Option(help="Number of bills per set")] = DEFAULT_COUNT,
-    concurrency: Annotated[int, typer.Option(help="Concurrent in-flight requests")] = 16,
-    log_level: Annotated[str, typer.Option(help="Log level")] = "INFO",
-) -> None:
-    """Run two interactive OpenAI sweeps (compressed + uncompressed) over bill text."""
-    logging.basicConfig(level=log_level, format="%(asctime)s %(levelname)s %(name)s: %(message)s")
-
-    api_key = getenv("CLOSEDAI_TOKEN") or getenv("OPENAI_API_KEY")
-    if not api_key:
-        message = "Neither CLOSEDAI_TOKEN nor OPENAI_API_KEY is set"
-        raise typer.BadParameter(message)
-    if not csv_path.is_file():
-        message = f"CSV not found: {csv_path}"
-        raise typer.BadParameter(message)
-
-    compressed_dir = output_dir / "compressed"
-    uncompressed_dir = output_dir / "uncompressed"
-    compressed_dir.mkdir(parents=True, exist_ok=True)
-    uncompressed_dir.mkdir(parents=True, exist_ok=True)
-
-    logger.info("Loading %d bills from %s", count, csv_path)
-    bills = load_bills(csv_path, count)
-    if len(bills) < count:
-        logger.warning("Only %d bills available (requested %d)", len(bills), count)
-
-    tasks: list[tuple[str, str, str, Path]] = []
-    for bill_id, text_content in bills:
-        filename = f"{safe_filename(bill_id)}.json"
-        tasks.append((bill_id, "compressed", compress_bill_text(text_content), compressed_dir / filename))
-        tasks.append((bill_id, "uncompressed", text_content, uncompressed_dir / filename))
-
-    logger.info("Submitting %d requests at concurrency=%d", len(tasks), concurrency)
-
-    headers = {"Authorization": f"Bearer {api_key}"}
-    completed = 0
-    failed = 0
-    index: list[dict] = []
-    wall_start = time.monotonic()
-    with (
-        httpx.Client(headers=headers, timeout=httpx.Timeout(300.0)) as client,
-        ThreadPoolExecutor(
-            max_workers=concurrency,
-        ) as executor,
-    ):
-        future_to_task = {
-            executor.submit(
-                run_one_request,
-                client,
-                bill_id=bill_id,
-                label=label,
-                bill_text=bill_text,
-                model=model,
-                output_path=output_path,
-            ): (bill_id, label, output_path)
-            for bill_id, label, bill_text, output_path in tasks
-        }
-        for future in as_completed(future_to_task):
-            bill_id, label, output_path = future_to_task[future]
-            success, elapsed, response_id = future.result()
-            if success:
-                completed += 1
-            else:
-                failed += 1
-            index.append(
-                {
-                    "bill_id": bill_id,
-                    "label": label,
-                    "response_id": response_id,
-                    "elapsed_seconds": elapsed,
-                    "success": success,
-                    "path": str(output_path),
-                },
-            )
-    wall_elapsed = time.monotonic() - wall_start
-
-    summary = {
-        "model": model,
-        "count": len(bills),
-        "completed": completed,
-        "failed": failed,
-        "wall_seconds": wall_elapsed,
-        "concurrency": concurrency,
-        "results": index,
-    }
-    summary_path = output_dir / "summary.json"
-    summary_path.write_text(json.dumps(summary, indent=2))
-    logger.info(
-        "Done: completed=%d failed=%d wall=%.1fs summary=%s",
-        completed,
-        failed,
-        wall_elapsed,
-        summary_path,
-    )
-
-
-def cli() -> None:
-    """Typer entry point."""
-    typer.run(main)
-
-
-if __name__ == "__main__":
-    cli()

python/prompt_bench/containers/__init__.py

@@ -1 +0,0 @@
-"""Prompt benchmarking system for evaluating LLMs via vLLM."""

python/prompt_bench/containers/finetune.py

@@ -1,165 +0,0 @@
-"""Docker container lifecycle management for Unsloth fine-tuning."""
-
-from __future__ import annotations
-
-import logging
-import subprocess
-from pathlib import Path
-from typing import Annotated
-
-import typer
-
-from python.prompt_bench.containers.lib import check_gpu_free
-
-logger = logging.getLogger(__name__)
-
-CONTAINER_NAME = "bill-finetune"
-FINETUNE_IMAGE = "bill-finetune:latest"
-DOCKERFILE_PATH = "/home/richie/dotfiles/python/prompt_bench/Dockerfile.finetune"
-DEFAULT_HF_CACHE = Path("/zfs/models/hf")
-
-
-def build_image() -> None:
-    """Build the fine-tuning Docker image."""
-    logger.info("Building fine-tuning image: %s", FINETUNE_IMAGE)
-    result = subprocess.run(
-        ["docker", "build", "-f", DOCKERFILE_PATH, "-t", FINETUNE_IMAGE, "."],
-        text=True,
-        check=False,
-    )
-    if result.returncode != 0:
-        message = "Failed to build fine-tuning image"
-        raise RuntimeError(message)
-    logger.info("Image built: %s", FINETUNE_IMAGE)
-
-
-def start_finetune(
-    *,
-    dataset_path: Path,
-    output_dir: Path,
-    hf_cache: Path = DEFAULT_HF_CACHE,
-) -> None:
-    """Run the fine-tuning container.
-
-    Args:
-        dataset_path: Host path to the fine-tuning JSONL dataset.
-        output_dir: Host path where the trained model will be saved.
-        hf_cache: Host path to HuggingFace model cache (bind-mounted to avoid re-downloading).
-        validation_split: Fraction of data held out for validation.
-    """
-    dataset_path = dataset_path.resolve()
-    output_dir = output_dir.resolve()
-
-    if not dataset_path.is_file():
-        message = f"Dataset not found: {dataset_path}"
-        raise FileNotFoundError(message)
-
-    output_dir.mkdir(parents=True, exist_ok=True)
-    stop_finetune()
-
-    hf_cache = hf_cache.resolve()
-    hf_cache.mkdir(parents=True, exist_ok=True)
-
-    command = [
-        "docker",
-        "run",
-        "--name",
-        CONTAINER_NAME,
-        "--device=nvidia.com/gpu=all",
-        "--ipc=host",
-        "-v",
-        f"{hf_cache}:/root/.cache/huggingface",
-        "-v",
-        f"{output_dir}:/workspace/output/qwen-bill-summarizer",
-        "-v",
-        f"{dataset_path}:/workspace/dataset.jsonl:ro",
-        FINETUNE_IMAGE,
-        "--dataset",
-        "/workspace/dataset.jsonl",
-        "--output-dir",
-        "/workspace/output/qwen-bill-summarizer",
-    ]
-
-    logger.info("Starting fine-tuning container")
-    logger.info("  Dataset:    %s", dataset_path)
-    logger.info("  Output:     %s", output_dir)
-
-    result = subprocess.run(command, text=True, check=False)
-    if result.returncode != 0:
-        message = f"Fine-tuning container exited with code {result.returncode}"
-        raise RuntimeError(message)
-    logger.info("Fine-tuning complete. Model saved to %s", output_dir)
-
-
-def stop_finetune() -> None:
-    """Stop and remove the fine-tuning container."""
-    logger.info("Stopping fine-tuning container")
-    subprocess.run(["docker", "stop", CONTAINER_NAME], capture_output=True, check=False)
-    subprocess.run(["docker", "rm", "-f", CONTAINER_NAME], capture_output=True, check=False)
-
-
-def logs_finetune() -> str | None:
-    """Return recent logs from the fine-tuning container, or None if not running."""
-    result = subprocess.run(
-        ["docker", "logs", "--tail", "50", CONTAINER_NAME],
-        capture_output=True,
-        text=True,
-        check=False,
-    )
-    if result.returncode != 0:
-        return None
-    return result.stdout + result.stderr
-
-
-app = typer.Typer(help="Fine-tuning container management.")
-
-
-@app.command()
-def build() -> None:
-    """Build the fine-tuning Docker image."""
-    build_image()
-
-
-@app.command()
-def run(
-    dataset: Annotated[Path, typer.Option(help="Fine-tuning JSONL")] = Path(
-        "/home/richie/dotfiles/data/finetune_dataset.jsonl"
-    ),
-    output_dir: Annotated[Path, typer.Option(help="Where to save the trained model")] = Path(
-        "/home/richie/dotfiles/data/output/qwen-bill-summarizer",
-    ),
-    hf_cache: Annotated[Path, typer.Option(help="Host path to HuggingFace model cache")] = DEFAULT_HF_CACHE,
-    log_level: Annotated[str, typer.Option(help="Log level")] = "INFO",
-) -> None:
-    """Run fine-tuning inside a Docker container."""
-    logging.basicConfig(level=log_level, format="%(asctime)s %(levelname)s %(name)s: %(message)s")
-    check_gpu_free()
-    start_finetune(
-        dataset_path=dataset,
-        output_dir=output_dir,
-        hf_cache=hf_cache,
-    )
-
-@app.command()
-def stop() -> None:
-    """Stop and remove the fine-tuning container."""
-    stop_finetune()
-
-
-@app.command()
-def logs() -> None:
-    """Show recent logs from the fine-tuning container."""
-    output = logs_finetune()
-    if output is None:
-        typer.echo("No running fine-tuning container found.")
-        raise typer.Exit(code=1)
-    typer.echo(output)
-
-
-def cli() -> None:
-    """Typer entry point."""
-    app()
-
-
-if __name__ == "__main__":
-    cli()

python/prompt_bench/containers/lib.py

@@ -1,23 +0,0 @@
-from __future__ import annotations
-
-import logging
-import subprocess
-
-logger = logging.getLogger(__name__)
-
-
-def check_gpu_free() -> None:
-    """Warn if GPU-heavy processes (e.g. Ollama) are running."""
-    result = subprocess.run(
-        ["nvidia-smi", "--query-compute-apps=pid,process_name", "--format=csv,noheader"],
-        capture_output=True,
-        text=True,
-        check=False,
-    )
-    if result.returncode != 0:
-        logger.warning("Could not query GPU processes: %s", result.stderr.strip())
-        return
-    processes = result.stdout.strip()
-    if processes:
-        logger.warning("GPU processes detected:\n%s", processes)
-        logger.warning("Consider stopping Ollama (sudo systemctl stop ollama) before benchmarking")

python/prompt_bench/containers/vllm.py

@@ -1,70 +0,0 @@
-"""Docker container lifecycle management for vLLM."""
-
-from __future__ import annotations
-
-import logging
-import subprocess
-
-logger = logging.getLogger(__name__)
-
-CONTAINER_NAME = "vllm-bench"
-VLLM_IMAGE = "vllm/vllm-openai:v0.19.0"
-
-
-def start_vllm(
-    *,
-    model: str,
-    port: int,
-    model_dir: str,
-    gpu_memory_utilization: float,
-) -> None:
-    """Start a vLLM container serving the given model.
-
-    Args:
-        model: HuggingFace model directory name (relative to model_dir).
-        port: Host port to bind.
-        model_dir: Host path containing HuggingFace model directories.
-        gpu_memory_utilization: Fraction of GPU memory to use (0-1).
-    """
-    command = [
-        "docker",
-        "run",
-        "-d",
-        "--name",
-        CONTAINER_NAME,
-        "--device=nvidia.com/gpu=all",
-        "--ipc=host",
-        "-v",
-        f"{model_dir}:/models",
-        "-p",
-        f"{port}:8000",
-        VLLM_IMAGE,
-        "--model",
-        f"/models/{model}",
-        "--served-model-name",
-        model,
-        "--gpu-memory-utilization",
-        str(gpu_memory_utilization),
-        "--max-model-len",
-        "4096",
-    ]
-    logger.info("Starting vLLM container with model: %s", model)
-    stop_vllm()
-    result = subprocess.run(command, capture_output=True, text=True, check=False)
-    if result.returncode != 0:
-        msg = f"Failed to start vLLM container: {result.stderr.strip()}"
-        raise RuntimeError(msg)
-    logger.info("vLLM container started: %s", result.stdout.strip()[:12])
-
-
-def stop_vllm() -> None:
-    """Stop and remove the vLLM benchmark container."""
-    logger.info("Stopping vLLM container")
-    subprocess.run(["docker", "stop", CONTAINER_NAME], capture_output=True, check=False)
-    subprocess.run(["docker", "rm", "-f", CONTAINER_NAME], capture_output=True, check=False)
-    subprocess.run(
-        ["docker", "network", "disconnect", "-f", "bridge", CONTAINER_NAME],
-        capture_output=True,
-        check=False,
-    )
-    logger.info("vLLM container stopped and removed")

python/prompt_bench/downloader.py

@@ -1,75 +0,0 @@
-"""HuggingFace model downloader."""
-
-from __future__ import annotations
-
-import logging
-from pathlib import Path
-from typing import Annotated
-
-import typer
-from huggingface_hub import snapshot_download
-
-from python.prompt_bench.models import BenchmarkConfig
-
-logger = logging.getLogger(__name__)
-
-
-def local_model_path(repo: str, model_dir: str) -> Path:
-    """Return the local directory path for a HuggingFace repo."""
-    return Path(model_dir) / repo
-
-
-def is_model_present(repo: str, model_dir: str) -> bool:
-    """Check if a model has already been downloaded."""
-    path = local_model_path(repo, model_dir)
-    return path.exists() and any(path.iterdir())
-
-
-def download_model(repo: str, model_dir: str) -> Path:
-    """Download a HuggingFace model to the local model directory.
-
-    Skips the download if the model directory already exists and contains files.
-    """
-    local_path = local_model_path(repo, model_dir)
-
-    if is_model_present(repo, model_dir):
-        logger.info("Model already exists: %s", local_path)
-        return local_path
-
-    logger.info("Downloading model: %s -> %s", repo, local_path)
-    snapshot_download(
-        repo_id=repo,
-        local_dir=str(local_path),
-    )
-    logger.info("Download complete: %s", repo)
-    return local_path
-
-
-def download_all(config: BenchmarkConfig) -> None:
-    """Download every model listed in the config, top to bottom."""
-    for repo in config.models:
-        download_model(repo, config.model_dir)
-
-
-def main(
-    config: Annotated[Path, typer.Option(help="Path to TOML config file")] = Path("bench.toml"),
-    log_level: Annotated[str, typer.Option(help="Log level")] = "INFO",
-) -> None:
-    """Download all models listed in the benchmark config."""
-    logging.basicConfig(level=log_level, format="%(asctime)s %(levelname)s %(name)s: %(message)s")
-
-    if not config.is_file():
-        message = f"Config file does not exist: {config}"
-        raise typer.BadParameter(message)
-
-    benchmark_config = BenchmarkConfig.from_toml(config)
-    download_all(benchmark_config)
-
-
-def cli() -> None:
-    """Typer entry point."""
-    typer.run(main)
-
-
-if __name__ == "__main__":
-    cli()

python/prompt_bench/finetune.py

@@ -1,214 +0,0 @@
-"""Fine-tune Qwen 3.5 4B on bill summarization data using Unsloth.
-
-Loads a ChatML-style JSONL dataset (system/user/assistant messages),
-applies QLoRA with 4-bit quantization, and saves the merged model
-in HuggingFace format. Designed for a single RTX 3090 (24GB).
-
-Usage:
-    python -m python.prompt_bench.finetune \
-        --dataset output/finetune_dataset.jsonl \
-        --output-dir output/qwen-bill-summarizer
-"""
-
-from __future__ import annotations
-
-import json
-import logging
-from dataclasses import dataclass
-from pathlib import Path
-from typing import Annotated
-
-import tomllib
-import typer
-from unsloth import FastLanguageModel
-from datasets import Dataset
-from transformers import TrainingArguments
-from trl import SFTTrainer
-
-logger = logging.getLogger(__name__)
-
-
-@dataclass
-class LoraConfig:
-    """LoRA adapter hyperparameters."""
-
-    rank: int
-    alpha: int
-    dropout: float
-    targets: list[str]
-
-
-@dataclass
-class TrainingConfig:
-    """Training loop hyperparameters."""
-
-    learning_rate: float
-    epochs: int
-    batch_size: int
-    gradient_accumulation: int
-    max_seq_length: int
-    warmup_ratio: float
-    weight_decay: float
-    logging_steps: int
-    save_steps: int
-
-
-@dataclass
-class FinetuneConfig:
-    """Top-level finetune configuration."""
-
-    base_model: str
-    lora: LoraConfig
-    training: TrainingConfig
-
-    @classmethod
-    def from_toml(cls, config_path: Path) -> FinetuneConfig:
-        """Load finetune config from a TOML file."""
-        raw = tomllib.loads(config_path.read_text())["finetune"]
-        return cls(
-            base_model=raw["base_model"],
-            lora=LoraConfig(**raw["lora"]),
-            training=TrainingConfig(**raw["training"]),
-        )
-
-
-def _messages_to_chatml(messages: list[dict]) -> str:
-    r"""Convert a message list to Qwen ChatML format.
-
-    Produces:
-        <|im_start|>system\n...\n<|im_end|>
-        <|im_start|>user\n...\n<|im_end|>
-        <|im_start|>assistant\n...\n<|im_end|>
-    """
-    parts = []
-    for message in messages:
-        role = message["role"]
-        content = message["content"]
-        parts.append(f"<|im_start|>{role}\n{content}<|im_end|>")
-    return "\n".join(parts)
-
-
-def load_dataset_from_jsonl(path: Path) -> Dataset:
-    """Load a ChatML JSONL file into a HuggingFace Dataset.
-
-    Each line must have {"messages": [{"role": ..., "content": ...}, ...]}.
-    Pre-formats into a `text` column with the Qwen ChatML template applied,
-    which SFTTrainer consumes directly.
-    """
-    records = []
-    with path.open(encoding="utf-8") as handle:
-        for raw_line in handle:
-            stripped = raw_line.strip()
-            if stripped:
-                entry = json.loads(stripped)
-                records.append({"text": _messages_to_chatml(entry["messages"])})
-    logger.info("Loaded %d examples from %s", len(records), path)
-    return Dataset.from_list(records)
-
-
-def main(
-    dataset_path: Annotated[Path, typer.Option("--dataset", help="Fine-tuning JSONL")] = Path(
-        "output/finetune_dataset.jsonl",
-    ),
-    validation_split: Annotated[float, typer.Option("--val-split", help="Fraction held out for validation")] = 0.1,
-    output_dir: Annotated[Path, typer.Option("--output-dir", help="Where to save the merged model")] = Path(
-        "output/qwen-bill-summarizer",
-    ),
-    config_path: Annotated[
-        Path,
-        typer.Option("--config", help="TOML config file"),
-    ] = Path(__file__).parent / "config.toml",
-    save_gguf: Annotated[bool, typer.Option("--save-gguf/--no-save-gguf", help="Also save GGUF")] = False,
-) -> None:
-    """Fine-tune Qwen 3.5 4B on bill summarization with Unsloth + QLoRA."""
-    logging.basicConfig(level="INFO", format="%(asctime)s %(levelname)s %(name)s: %(message)s")
-
-    if not dataset_path.is_file():
-        message = f"Dataset not found: {dataset_path}"
-        raise typer.BadParameter(message)
-
-    config = FinetuneConfig.from_toml(config_path)
-
-    logger.info("Loading base model: %s", config.base_model)
-    model, tokenizer = FastLanguageModel.from_pretrained(
-        model_name=config.base_model,
-        max_seq_length=config.training.max_seq_length,
-        load_in_4bit=True,
-        dtype=None,
-    )
-
-    logger.info("Applying LoRA (rank=%d, alpha=%d)", config.lora.rank, config.lora.alpha)
-    model = FastLanguageModel.get_peft_model(
-        model,
-        r=config.lora.rank,
-        lora_alpha=config.lora.alpha,
-        lora_dropout=config.lora.dropout,
-        target_modules=config.lora.targets,
-        bias="none",
-        use_gradient_checkpointing="unsloth",
-        random_state=42,
-    )
-
-    full_dataset = load_dataset_from_jsonl(dataset_path)
-    split = full_dataset.train_test_split(test_size=validation_split, seed=42)
-    train_dataset = split["train"]
-    validation_dataset = split["test"]
-    logger.info("Split: %d train, %d validation", len(train_dataset), len(validation_dataset))
-    training_args = TrainingArguments(
-        output_dir=str(output_dir / "checkpoints"),
-        num_train_epochs=config.training.epochs,
-        per_device_train_batch_size=config.training.batch_size,
-        gradient_accumulation_steps=config.training.gradient_accumulation,
-        learning_rate=config.training.learning_rate,
-        warmup_ratio=config.training.warmup_ratio,
-        weight_decay=config.training.weight_decay,
-        lr_scheduler_type="cosine",
-        logging_steps=config.training.logging_steps,
-        save_steps=config.training.save_steps,
-        save_total_limit=3,
-        eval_strategy="steps",
-        eval_steps=config.training.save_steps,
-        load_best_model_at_end=True,
-        bf16=True,
-        optim="adamw_8bit",
-        seed=42,
-        report_to="none",
-    )
-
-    trainer = SFTTrainer(
-        model=model,
-        tokenizer=tokenizer,
-        train_dataset=train_dataset,
-        eval_dataset=validation_dataset,
-        args=training_args,
-        max_seq_length=config.training.max_seq_length,
-        packing=True,
-    )
-
-    logger.info(
-        "Starting training: %d train, %d val, %d epochs",
-        len(train_dataset),
-        len(validation_dataset),
-        config.training.epochs,
-    )
-    trainer.train()
-
-    merged_path = str(output_dir / "merged")
-    logger.info("Saving merged model to %s", merged_path)
-    model.save_pretrained_merged(merged_path, tokenizer, save_method="merged_16bit")
-
-    if save_gguf:
-        gguf_path = str(output_dir / "gguf")
-        logger.info("Saving GGUF to %s", gguf_path)
-        model.save_pretrained_gguf(gguf_path, tokenizer, quantization_method="q4_k_m")
-
-    logger.info("Done! Model saved to %s", output_dir)
-
-
-def cli() -> None:
-    """Typer entry point."""
-    typer.run(main)
-
-
-if __name__ == "__main__":
-    cli()

python/prompt_bench/main.py

@@ -1,215 +0,0 @@
-"""CLI entry point for the prompt benchmarking system."""
-
-from __future__ import annotations
-
-import json
-import logging
-import time
-from concurrent.futures import ThreadPoolExecutor, as_completed
-from pathlib import Path
-from typing import Annotated
-
-import typer
-
-from python.prompt_bench.containers.lib import check_gpu_free
-from python.prompt_bench.containers.vllm import start_vllm, stop_vllm
-from python.prompt_bench.downloader import is_model_present
-from python.prompt_bench.models import BenchmarkConfig
-from python.prompt_bench.vllm_client import VLLMClient
-
-logger = logging.getLogger(__name__)
-
-
-def discover_prompts(input_dir: Path) -> list[Path]:
-    """Find all .txt files in the input directory."""
-    prompts = list(input_dir.glob("*.txt"))
-    if not prompts:
-        message = f"No .txt files found in {input_dir}"
-        raise FileNotFoundError(message)
-    return prompts
-
-
-def _run_prompt(
-    client: VLLMClient,
-    prompt_path: Path,
-    *,
-    repo: str,
-    model_dir_name: str,
-    model_output: Path,
-    temperature: float,
-) -> tuple[bool, float]:
-    """Run a single prompt. Returns (success, elapsed_seconds)."""
-    filename = prompt_path.name
-    output_path = model_output / filename
-    start = time.monotonic()
-    try:
-        prompt_text = prompt_path.read_text()
-        response = client.complete(prompt_text, model_dir_name, temperature=temperature)
-        output_path.write_text(response)
-        elapsed = time.monotonic() - start
-        logger.info("Completed: %s / %s in %.2fs", repo, filename, elapsed)
-    except Exception:
-        elapsed = time.monotonic() - start
-        error_path = model_output / f"{filename}.error"
-        logger.exception("Failed: %s / %s after %.2fs", repo, filename, elapsed)
-        error_path.write_text(f"Error processing {filename}")
-        return False, elapsed
-    return True, elapsed
-
-
-def benchmark_model(
-    client: VLLMClient,
-    prompts: list[Path],
-    *,
-    repo: str,
-    model_dir_name: str,
-    model_output: Path,
-    temperature: float,
-    concurrency: int,
-) -> tuple[int, int]:
-    """Run all prompts against a single model in parallel.
-
-    vLLM batches concurrent requests internally, so submitting many at once is
-    significantly faster than running them serially.
-    """
-    pending = [prompt for prompt in prompts if not (model_output / prompt.name).exists()]
-    skipped = len(prompts) - len(pending)
-    if skipped:
-        logger.info("Skipping %d prompts with existing output for %s", skipped, repo)
-
-    if not pending:
-        logger.info("Nothing to do for %s", repo)
-        return 0, 0
-
-    completed = 0
-    failed = 0
-    latencies: list[float] = []
-
-    wall_start = time.monotonic()
-    with ThreadPoolExecutor(max_workers=concurrency) as executor:
-        futures = [
-            executor.submit(
-                _run_prompt,
-                client,
-                prompt_path,
-                repo=repo,
-                model_dir_name=model_dir_name,
-                model_output=model_output,
-                temperature=temperature,
-            )
-            for prompt_path in pending
-        ]
-        for future in as_completed(futures):
-            success, elapsed = future.result()
-            latencies.append(elapsed)
-            if success:
-                completed += 1
-            else:
-                failed += 1
-    wall_elapsed = time.monotonic() - wall_start
-
-    attempted = completed + failed
-    avg_latency = sum(latencies) / attempted
-    throughput = attempted / wall_elapsed if wall_elapsed > 0 else 0.0
-    timing = {
-        "repo": repo,
-        "wall_seconds": wall_elapsed,
-        "attempted": attempted,
-        "completed": completed,
-        "failed": failed,
-        "avg_latency_seconds": avg_latency,
-        "throughput_prompts_per_second": throughput,
-        "concurrency": concurrency,
-    }
-    timing_path = model_output / "_timing.json"
-    timing_path.write_text(json.dumps(timing, indent=2))
-
-    return completed, failed
-
-
-def run_benchmark(
-    config: BenchmarkConfig,
-    input_dir: Path,
-    output_dir: Path,
-) -> None:
-    """Execute the benchmark across all models and prompts."""
-    prompts = discover_prompts(input_dir)
-    logger.info("Found %d prompts in %s", len(prompts), input_dir)
-
-    check_gpu_free()
-
-    total_completed = 0
-    total_failed = 0
-
-    for repo in config.models:
-        if not is_model_present(repo, config.model_dir):
-            logger.warning("Skipping (not downloaded): %s", repo)
-            continue
-
-        model_output = output_dir / repo
-        model_output.mkdir(parents=True, exist_ok=True)
-
-        logger.info("=== Benchmarking model: %s ===", repo)
-
-        stop_vllm()
-        try:
-            start_vllm(
-                model=repo,
-                port=config.port,
-                model_dir=config.model_dir,
-                gpu_memory_utilization=config.gpu_memory_utilization,
-            )
-        except RuntimeError:
-            logger.exception("Failed to start vLLM for %s, skipping", repo)
-            continue
-        logger.info("vLLM started for %s", repo)
-        try:
-            with VLLMClient(port=config.port, timeout=config.timeout) as client:
-                client.wait_ready(max_wait=config.vllm_startup_timeout)
-                completed, failed = benchmark_model(
-                    client,
-                    prompts,
-                    repo=repo,
-                    model_dir_name=repo,
-                    model_output=model_output,
-                    temperature=config.temperature,
-                    concurrency=config.concurrency,
-                )
-                total_completed += completed
-                total_failed += failed
-        finally:
-            stop_vllm()
-
-    logger.info("=== Benchmark complete ===")
-    logger.info("Completed: %d | Failed: %d", total_completed, total_failed)
-
-
-def main(
-    input_dir: Annotated[Path, typer.Argument(help="Directory containing input .txt prompt files")],
-    config: Annotated[Path, typer.Option(help="Path to TOML config file")] = Path("bench.toml"),
-    output_dir: Annotated[Path, typer.Option(help="Output directory for results")] = Path("output"),
-    log_level: Annotated[str, typer.Option(help="Log level")] = "INFO",
-) -> None:
-    """Run prompts through multiple LLMs via vLLM and save results."""
-    logging.basicConfig(level=log_level, format="%(asctime)s %(levelname)s %(name)s: %(message)s")
-
-    if not input_dir.is_dir():
-        message = f"Input directory does not exist: {input_dir}"
-        raise typer.BadParameter(message)
-    if not config.is_file():
-        message = f"Config file does not exist: {config}"
-        raise typer.BadParameter(message)
-
-    benchmark_config = BenchmarkConfig.from_toml(config)
-    output_dir.mkdir(parents=True, exist_ok=True)
-
-    run_benchmark(benchmark_config, input_dir, output_dir)
-
-
-def cli() -> None:
-    """Typer entry point."""
-    typer.run(main)
-
-
-if __name__ == "__main__":
-    cli()

python/prompt_bench/models.py

@@ -1,30 +0,0 @@
-"""Pydantic models for benchmark configuration."""
-
-from __future__ import annotations
-
-import tomllib
-from typing import TYPE_CHECKING
-
-from pydantic import BaseModel
-
-if TYPE_CHECKING:
-    from pathlib import Path
-
-
-class BenchmarkConfig(BaseModel):
-    """Top-level benchmark configuration loaded from TOML."""
-
-    models: list[str]
-    model_dir: str = "/zfs/models/hf"
-    port: int = 8000
-    gpu_memory_utilization: float = 0.90
-    temperature: float = 0.0
-    timeout: int = 300
-    concurrency: int = 4
-    vllm_startup_timeout: int = 900
-
-    @classmethod
-    def from_toml(cls, config_path: Path) -> BenchmarkConfig:
-        """Load benchmark config from a TOML file."""
-        raw = tomllib.loads(config_path.read_text())["bench"]
-        return cls(**raw)

python/prompt_bench/summarization_prompts.py

@@ -1,34 +0,0 @@
-SUMMARIZATION_SYSTEM_PROMPT = """You are a legislative analyst extracting policy substance from Congressional bill text.
-
-Your job is to compress a bill into a dense, neutral structured summary that captures every distinct policy action — including secondary effects that might be buried in subsections.
-
-EXTRACTION RULES:
-- IGNORE: whereas clauses, congressional findings that are purely political statements, recitals, preambles, citations of existing law by number alone, and procedural boilerplate.
-- FOCUS ON: operative verbs — what the bill SHALL do, PROHIBIT, REQUIRE, AUTHORIZE, AMEND, APPROPRIATE, or ESTABLISH.
-- SURFACE ALL THREADS: If the bill touches multiple policy areas, list each thread separately. Do not collapse them.
-- BE CONCRETE: Name the affected population, the mechanism, and the direction (expands/restricts/maintains).
-- STAY NEUTRAL: No political framing. Describe what the text does, not what its sponsors claim it does.
-
-OUTPUT FORMAT — plain structured text, not JSON:
-
-OPERATIVE ACTIONS:
-[Numbered list of what the bill actually does, one action per line, max 20 words each]
-
-AFFECTED POPULATIONS:
-[Who gains something, who loses something, or whose behavior is regulated]
-
-MECHANISMS:
-[How it works: new funding, mandate, prohibition, amendment to existing statute, grant program, study commission, etc.]
-
-POLICY THREADS:
-[List each distinct policy domain this bill touches, even minor ones. Use plain language, not domain codes.]
-
-SYMBOLIC/PROCEDURAL ONLY:
-[Yes or No — is this bill primarily a resolution, designation, or awareness declaration with no operative effect?]
-
-LENGTH TARGET: 150-250 words total. Be ruthless about cutting. Density over completeness."""
-
-SUMMARIZATION_USER_TEMPLATE = """Summarize the following Congressional bill according to your instructions.
-
-BILL TEXT:
-{text_content}"""

python/prompt_bench/tools/build_finetune_dataset.py

@@ -1,114 +0,0 @@
-"""Build a fine-tuning JSONL dataset from batch request + output files.
-
-Joins the original request JSONL (system + user messages) with the batch
-output JSONL (assistant completions) by custom_id to produce a ChatML-style
-messages JSONL suitable for fine-tuning.
-"""
-
-from __future__ import annotations
-
-import json
-import logging
-from pathlib import Path
-from typing import Annotated
-
-import typer
-
-logger = logging.getLogger(__name__)
-
-HTTP_OK = 200
-
-
-def load_requests(path: Path) -> dict[str, list[dict]]:
-    """Parse request JSONL into {custom_id: messages}."""
-    results: dict[str, list[dict]] = {}
-    with path.open(encoding="utf-8") as handle:
-        for raw_line in handle:
-            stripped = raw_line.strip()
-            if not stripped:
-                continue
-            record = json.loads(stripped)
-            custom_id = record["custom_id"]
-            messages = record["body"]["messages"]
-            results[custom_id] = messages
-    return results
-
-
-def load_completions(path: Path) -> dict[str, str]:
-    """Parse batch output JSONL into {custom_id: assistant_content}."""
-    results: dict[str, str] = {}
-    with path.open(encoding="utf-8") as handle:
-        for line_number, raw_line in enumerate(handle, 1):
-            stripped = raw_line.strip()
-            if not stripped:
-                continue
-            record = json.loads(stripped)
-            custom_id = record["custom_id"]
-            response = record.get("response", {})
-            if response.get("status_code") != HTTP_OK:
-                logger.warning("Skipping %s (line %d): status %s", custom_id, line_number, response.get("status_code"))
-                continue
-            body = response.get("body", {})
-            choices = body.get("choices", [])
-            if not choices:
-                logger.warning("Skipping %s (line %d): no choices", custom_id, line_number)
-                continue
-            content = choices[0].get("message", {}).get("content", "")
-            if not content:
-                logger.warning("Skipping %s (line %d): empty content", custom_id, line_number)
-                continue
-            results[custom_id] = content
-    return results
-
-
-def main(
-    requests_path: Annotated[Path, typer.Option("--requests", help="Batch request JSONL")] = Path(
-        "output/openai_batch/requests.jsonl",
-    ),
-    batch_output: Annotated[Path, typer.Option("--batch-output", help="Batch output JSONL")] = Path(
-        "batch_69d84558d91c819091d53f08d78f9fd6_output.jsonl",
-    ),
-    output_path: Annotated[Path, typer.Option("--output", help="Fine-tuning JSONL output")] = Path(
-        "output/finetune_dataset.jsonl",
-    ),
-    log_level: Annotated[str, typer.Option(help="Log level")] = "INFO",
-) -> None:
-    """Build fine-tuning dataset by joining request and output JSONL files."""
-    logging.basicConfig(level=log_level, format="%(asctime)s %(levelname)s %(name)s: %(message)s")
-
-    logger.info("Loading requests from %s", requests_path)
-    requests = load_requests(requests_path)
-    logger.info("Loaded %d requests", len(requests))
-
-    logger.info("Loading completions from %s", batch_output)
-    completions = load_completions(batch_output)
-    logger.info("Loaded %d completions", len(completions))
-
-    output_path.parent.mkdir(parents=True, exist_ok=True)
-    matched = 0
-    skipped = 0
-
-    with output_path.open("w", encoding="utf-8") as handle:
-        for custom_id, messages in requests.items():
-            assistant_content = completions.get(custom_id)
-            if assistant_content is None:
-                skipped += 1
-                continue
-
-            example = {
-                "messages": [*messages, {"role": "assistant", "content": assistant_content}],
-            }
-            handle.write(json.dumps(example, ensure_ascii=False))
-            handle.write("\n")
-            matched += 1
-
-    logger.info("Wrote %d examples to %s (skipped %d unmatched)", matched, output_path, skipped)
-
-
-def cli() -> None:
-    """Typer entry point."""
-    typer.run(main)
-
-
-if __name__ == "__main__":
-    cli()

python/prompt_bench/tools/count_tokens.py

@@ -1,97 +0,0 @@
-"""Sum token usage across compressed and uncompressed run directories."""
-
-from __future__ import annotations
-
-import json
-import logging
-from dataclasses import dataclass, field
-from pathlib import Path
-from typing import Annotated
-
-import typer
-
-logger = logging.getLogger(__name__)
-
-
-@dataclass
-class UsageTotals:
-    """Aggregate usage counters for a directory of run records."""
-
-    files: int = 0
-    errors: int = 0
-    prompt_tokens: int = 0
-    cached_tokens: int = 0
-    completion_tokens: int = 0
-    reasoning_tokens: int = 0
-    total_tokens: int = 0
-    per_file: list[tuple[str, int, int, int]] = field(default_factory=list)
-
-
-def tally_directory(directory: Path) -> UsageTotals:
-    """Return aggregated usage stats for every JSON record in a directory."""
-    totals = UsageTotals()
-    decoder = json.JSONDecoder()
-    for path in sorted(directory.glob("*.json")):
-        text = path.read_text().lstrip()
-        record, _ = decoder.raw_decode(text)
-        totals.files += 1
-        usage = record.get("usage")
-        if not usage:
-            totals.errors += 1
-            continue
-        prompt_tokens = usage.get("prompt_tokens", 0)
-        completion_tokens = usage.get("completion_tokens", 0)
-        total_tokens = usage.get("total_tokens", 0)
-        cached_tokens = (usage.get("prompt_tokens_details") or {}).get("cached_tokens", 0)
-        reasoning_tokens = (usage.get("completion_tokens_details") or {}).get("reasoning_tokens", 0)
-        totals.prompt_tokens += prompt_tokens
-        totals.completion_tokens += completion_tokens
-        totals.total_tokens += total_tokens
-        totals.cached_tokens += cached_tokens
-        totals.reasoning_tokens += reasoning_tokens
-        totals.per_file.append((path.name, prompt_tokens, completion_tokens, total_tokens))
-    return totals
-
-
-def log_totals(label: str, totals: UsageTotals) -> None:
-    """Log a one-block summary for a directory."""
-    counted = totals.files - totals.errors
-    average_total = totals.total_tokens / counted if counted else 0
-    logger.info("[%s]", label)
-    logger.info("  files          : %d (with usage: %d, errors: %d)", totals.files, counted, totals.errors)
-    logger.info("  prompt tokens  : %d", totals.prompt_tokens)
-    logger.info("  cached tokens  : %d", totals.cached_tokens)
-    logger.info("  completion tok : %d", totals.completion_tokens)
-    logger.info("  reasoning tok  : %d", totals.reasoning_tokens)
-    logger.info("  total tokens   : %d", totals.total_tokens)
-    logger.info("  avg total/file : %.1f", average_total)
-
-
-def main(
-    runs_dir: Annotated[Path, typer.Option("--runs-dir")] = Path("output/openai_runs_temp_1"),
-    log_level: Annotated[str, typer.Option("--log-level")] = "INFO",
-) -> None:
-    """Print token usage totals for the compressed and uncompressed run directories."""
-    logging.basicConfig(level=log_level, format="%(message)s")
-
-    grand = UsageTotals()
-    for label in ("compressed", "uncompressed"):
-        directory = runs_dir / label
-        if not directory.is_dir():
-            logger.warning("%s: directory not found at %s", label, directory)
-            continue
-        totals = tally_directory(directory)
-        log_totals(label, totals)
-        grand.files += totals.files
-        grand.errors += totals.errors
-        grand.prompt_tokens += totals.prompt_tokens
-        grand.cached_tokens += totals.cached_tokens
-        grand.completion_tokens += totals.completion_tokens
-        grand.reasoning_tokens += totals.reasoning_tokens
-        grand.total_tokens += totals.total_tokens
-
-    log_totals("grand total", grand)
-
-
-if __name__ == "__main__":
-    typer.run(main)

python/prompt_bench/vllm_client.py

@@ -1,68 +0,0 @@
-"""OpenAI-compatible client for vLLM's API."""
-
-from __future__ import annotations
-
-import logging
-import time
-from typing import Self
-
-import httpx
-
-logger = logging.getLogger(__name__)
-
-READY_POLL_INTERVAL = 2.0
-
-
-class VLLMClient:
-    """Talk to a vLLM server via its OpenAI-compatible API.
-
-    Args:
-        host: vLLM host.
-        port: vLLM port.
-        timeout: Per-request timeout in seconds.
-    """
-
-    def __init__(self, *, host: str = "localhost", port: int = 8000, timeout: int = 300) -> None:
-        """Create a client connected to a vLLM server."""
-        self._client = httpx.Client(base_url=f"http://{host}:{port}", timeout=timeout)
-
-    def wait_ready(self, max_wait: int) -> None:
-        """Poll /v1/models until the server is ready or timeout."""
-        deadline = time.monotonic() + max_wait
-        while time.monotonic() < deadline:
-            try:
-                response = self._client.get("/v1/models")
-                if response.is_success:
-                    logger.info("vLLM server is ready")
-                    return
-            except httpx.TransportError:
-                pass
-            time.sleep(READY_POLL_INTERVAL)
-        msg = f"vLLM server not ready after {max_wait}s"
-        raise TimeoutError(msg)
-
-    def complete(self, prompt: str, model: str, *, temperature: float = 0.0, max_tokens: int = 4096) -> str:
-        """Send a prompt to /v1/completions and return the response text."""
-        payload = {
-            "model": model,
-            "prompt": prompt,
-            "temperature": temperature,
-            "max_tokens": max_tokens,
-        }
-        logger.info("Sending prompt to %s (%d chars)", model, len(prompt))
-        response = self._client.post("/v1/completions", json=payload)
-        response.raise_for_status()
-        data = response.json()
-        return data["choices"][0]["text"]
-
-    def close(self) -> None:
-        """Close the HTTP client."""
-        self._client.close()
-
-    def __enter__(self) -> Self:
-        """Enter the context manager."""
-        return self
-
-    def __exit__(self, *args: object) -> None:
-        """Close the HTTP client on exit."""
-        self.close()

python/sheet_music_ocr/__init__.py

@@ -0,0 +1 @@
+"""Sheet music OCR tool using Audiveris."""