Add models and database persistence for protected phrase extraction

- Introduced dataclasses for phrase candidates, judgments, and matches in `models.py`.
- Implemented database operations for candidate and protected phrases in `store.py`, including loading, saving, and deleting phrases.
- Enhanced text normalization functions in `text_normalization.py` with detailed docstrings.
- Refactored search functionality to utilize new models and methods for detecting protected phrases.

python/ebook_search/api/routes/admin.py

@@ -14,11 +14,8 @@ from python.ebook_search.api.dependencies import (
 from python.ebook_search.api.web import templates
 from python.ebook_search.embeddings import embed_missing_chunks, embedding_model_stats
 from python.ebook_search.ingest import ingest_configured_paths
-from python.ebook_search.protected_phrases.lib import (
-    build_missing_protected_phrases,
-    generate_candidate_phrases_for_books,
-    judge_candidate_phrases_for_books,
-)
+from python.ebook_search.protected_phrases.generate_ngrams import generate_candidate_phrases_for_books
+from python.ebook_search.protected_phrases.judge_ngrams import judge_candidate_phrases_for_books
 from python.fastapi_tools import DbSession  # noqa: TC001 FastAPI resolves this annotated dependency at runtime
 
 logger = logging.getLogger(__name__)
@@ -50,36 +47,6 @@ def scan_library(request: Request, config: AppConfig, session: DbSession) -> HTM
     return templates.TemplateResponse(request, "partials/admin_status.html", {"message": f"Indexed {count} EPUBs"})
 
 
-@router.post("/build-phrases", response_class=HTMLResponse)
-def build_phrases(request: Request, config: AppConfig, session: DbSession) -> HTMLResponse:
-    """Build protected phrases for indexed books that are missing them."""
-    try:
-        result = build_missing_protected_phrases(session, config)
-        session.commit()
-    except Exception as error:
-        session.rollback()
-        logger.exception("ebook_admin_build_phrases_failed")
-        return templates.TemplateResponse(request, "partials/error.html", {"message": str(error)}, status_code=500)
-
-    logger.info(
-        "ebook_admin_build_phrases_complete books_seen=%s books_built=%s protected=%s mentions=%s",
-        result.books_seen,
-        result.books_built,
-        result.protected_phrases,
-        result.phrase_mentions,
-    )
-    return templates.TemplateResponse(
-        request,
-        "partials/admin_status.html",
-        {
-            "message": (
-                f"Built phrases for {result.books_built} of {result.books_seen} books; "
-                f"{result.protected_phrases} protected phrases, {result.phrase_mentions} mentions"
-            )
-        },
-    )
-
-
 @router.post("/generate-ngrams", response_class=HTMLResponse)
 def generate_ngrams(request: Request, config: AppConfig, session: DbSession) -> HTMLResponse:
     """Generate candidate n-grams for indexed books without LLM judging."""