added ebook embedding to orm

python/orm/richie/__init__.py

@@ -11,6 +11,15 @@ from python.orm.richie.contact import (
     Need,
     RelationshipType,
 )
+from python.orm.richie.ebook import (
+    EbookChapter,
+    EbookChunk,
+    EbookChunkEmbedding1024,
+    EbookChunkEmbedding2560,
+    EbookChunkEmbedding4096,
+    EbookEmbeddingModel,
+    EbookSource,
+)
 
 __all__ = [
     "Audiobook",
@@ -19,6 +28,13 @@ __all__ = [
     "Contact",
     "ContactNeed",
     "ContactRelationship",
+    "EbookChapter",
+    "EbookChunk",
+    "EbookChunkEmbedding1024",
+    "EbookChunkEmbedding2560",
+    "EbookChunkEmbedding4096",
+    "EbookEmbeddingModel",
+    "EbookSource",
     "Need",
     "RelationshipType",
     "RichieBase",

python/orm/richie/ebook.py

@@ -0,0 +1,130 @@
+"""EPUB search models."""
+
+from __future__ import annotations
+
+from datetime import datetime
+
+from pgvector.sqlalchemy import Vector
+from sqlalchemy import BigInteger, Boolean, DateTime, ForeignKey, String, UniqueConstraint
+from sqlalchemy.orm import Mapped, mapped_column, relationship
+
+from python.orm.richie.base import TableBase, TableBaseBig
+
+
+class EbookSource(TableBase):
+    """One indexed EPUB file."""
+
+    __tablename__ = "ebook_source"
+    __table_args__ = (
+        UniqueConstraint("file_path"),
+        UniqueConstraint("file_sha256"),
+    )
+
+    title: Mapped[str]
+    author: Mapped[str | None]
+    language: Mapped[str | None]
+    publisher: Mapped[str | None]
+    identifier: Mapped[str | None]
+    file_path: Mapped[str]
+    file_sha256: Mapped[str] = mapped_column(String(64))
+    file_mtime: Mapped[datetime] = mapped_column(DateTime(timezone=True))
+    file_size: Mapped[int] = mapped_column(BigInteger)
+
+    chapters: Mapped[list[EbookChapter]] = relationship(
+        "EbookChapter",
+        back_populates="source",
+        cascade="all, delete-orphan",
+        passive_deletes=True,
+    )
+    chunks: Mapped[list[EbookChunk]] = relationship(
+        "EbookChunk",
+        back_populates="source",
+        cascade="all, delete-orphan",
+        passive_deletes=True,
+    )
+
+
+class EbookChapter(TableBase):
+    """A chapter or spine document inside an EPUB."""
+
+    __tablename__ = "ebook_chapter"
+    __table_args__ = (UniqueConstraint("source_id", "spine_index"),)
+
+    source_id: Mapped[int] = mapped_column(ForeignKey("main.ebook_source.id", ondelete="CASCADE"))
+    spine_index: Mapped[int]
+    title: Mapped[str | None]
+    href: Mapped[str | None]
+
+    source: Mapped[EbookSource] = relationship("EbookSource", back_populates="chapters")
+    chunks: Mapped[list[EbookChunk]] = relationship(
+        "EbookChunk",
+        back_populates="chapter",
+        cascade="all, delete-orphan",
+        passive_deletes=True,
+    )
+
+
+class EbookChunk(TableBaseBig):
+    """A searchable text chunk."""
+
+    __tablename__ = "ebook_chunk"
+    __table_args__ = (
+        UniqueConstraint("source_id", "chunk_index", name="uq_ebook_chunk_source_id_chunk_index"),
+        UniqueConstraint("source_id", "content_sha256", name="uq_ebook_chunk_source_id_content_sha256"),
+    )
+
+    source_id: Mapped[int] = mapped_column(ForeignKey("main.ebook_source.id", ondelete="CASCADE"))
+    chapter_id: Mapped[int | None] = mapped_column(ForeignKey("main.ebook_chapter.id", ondelete="SET NULL"))
+    chunk_index: Mapped[int]
+    text: Mapped[str]
+    token_start: Mapped[int]
+    token_count: Mapped[int]
+    page_label: Mapped[str | None]
+    content_sha256: Mapped[str] = mapped_column(String(64))
+    search_text: Mapped[str]
+
+    source: Mapped[EbookSource] = relationship("EbookSource", back_populates="chunks")
+    chapter: Mapped[EbookChapter | None] = relationship("EbookChapter", back_populates="chunks")
+
+
+class EbookEmbeddingModel(TableBase):
+    """A supported embedding model."""
+
+    __tablename__ = "ebook_embedding_model"
+
+    name: Mapped[str] = mapped_column(String, unique=True)
+    dimension: Mapped[int]
+    is_default: Mapped[bool] = mapped_column(Boolean, default=False)
+
+
+class EbookChunkEmbedding1024(TableBaseBig):
+    """1024-dimensional chunk embedding."""
+
+    __tablename__ = "ebook_chunk_embedding_1024"
+    __table_args__ = (UniqueConstraint("chunk_id", "model_id"),)
+
+    chunk_id: Mapped[int] = mapped_column(ForeignKey("main.ebook_chunk.id", ondelete="CASCADE"))
+    model_id: Mapped[int] = mapped_column(ForeignKey("main.ebook_embedding_model.id", ondelete="CASCADE"))
+    embedding: Mapped[list[float]] = mapped_column(Vector(1024))
+
+
+class EbookChunkEmbedding2560(TableBaseBig):
+    """2560-dimensional chunk embedding."""
+
+    __tablename__ = "ebook_chunk_embedding_2560"
+    __table_args__ = (UniqueConstraint("chunk_id", "model_id"),)
+
+    chunk_id: Mapped[int] = mapped_column(ForeignKey("main.ebook_chunk.id", ondelete="CASCADE"))
+    model_id: Mapped[int] = mapped_column(ForeignKey("main.ebook_embedding_model.id", ondelete="CASCADE"))
+    embedding: Mapped[list[float]] = mapped_column(Vector(2560))
+
+
+class EbookChunkEmbedding4096(TableBaseBig):
+    """4096-dimensional chunk embedding."""
+
+    __tablename__ = "ebook_chunk_embedding_4096"
+    __table_args__ = (UniqueConstraint("chunk_id", "model_id"),)
+
+    chunk_id: Mapped[int] = mapped_column(ForeignKey("main.ebook_chunk.id", ondelete="CASCADE"))
+    model_id: Mapped[int] = mapped_column(ForeignKey("main.ebook_embedding_model.id", ondelete="CASCADE"))
+    embedding: Mapped[list[float]] = mapped_column(Vector(4096))