1df9be6563
- codex/parsing/mathpix.py: pix2tex (local, CPU) primary + MathPix API optional; bbox heuristic h>15px, math-char-count>5; singleton model cache - codex/parsing/figures.py: pymupdf embedded-image extraction → PNG; caption detection via proximity + "Figure/Fig./Abbildung" prefix - codex/models.py: FormulaChunk + FigureChunk dataclasses (R-10/R-11) - codex/ingest.py: --rich flag wires formula+figure extraction post-ingest - codex/cli.py: search_app sub-typer (paper + formula subcommands), --rich flag on ingest; wiki_app from F-12 preserved intact - codex/config.py: mathpix_app_id/key, pix2tex_fallback, figures_dir - infra/schema.sql: formulas + figures tables with HNSW pgvector indexes - pyproject.toml: pymupdf>=1.24, pix2tex>=0.1.4 - tests/parsing/test_mathpix.py + test_figures.py: 31 tests (mock pix2tex + MathPix HTTP, real pymupdf on synthetic PDF) Gate: 158 passed, ruff clean, mypy clean (20 files) Requirements: R-10 R-11 R-12 R-13 R-14 → done Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
117 lines
3.1 KiB
Python
117 lines
3.1 KiB
Python
"""Domain dataclasses.
|
|
|
|
Each class maps directly to a table in ``infra/schema.sql``.
|
|
Field names and types are kept in sync with the DDL.
|
|
|
|
All fields that are nullable in the schema are typed as ``… | None``
|
|
with a default of ``None`` so instances can be constructed incrementally.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
from dataclasses import dataclass, field
|
|
from datetime import datetime
|
|
|
|
|
|
@dataclass
|
|
class Paper:
|
|
"""Maps to the ``papers`` table (Layer 1 — semantic search).
|
|
|
|
``id`` is the canonical identifier: an arXiv ID (e.g. ``2301.00001``)
|
|
or a DOI (e.g. ``10.1145/3592430``).
|
|
"""
|
|
|
|
id: str
|
|
title: str
|
|
openalex_id: str | None = None
|
|
bibkey: str | None = None
|
|
authors: list[str] = field(default_factory=list)
|
|
year: int | None = None
|
|
abstract: str | None = None
|
|
source_path: str | None = None
|
|
# abstract_emb is a list of floats representing the pgvector column;
|
|
# pgvector's Python driver accepts list[float] for vector columns.
|
|
abstract_emb: list[float] | None = None
|
|
added_at: datetime | None = None
|
|
|
|
|
|
@dataclass
|
|
class Chunk:
|
|
"""Maps to the ``chunks`` table (Layer 1 — full-text chunks).
|
|
|
|
``id`` is set by the database (BIGSERIAL); leave it as ``None``
|
|
when constructing a new chunk before insertion.
|
|
"""
|
|
|
|
paper_id: str
|
|
ord: int
|
|
content: str
|
|
id: int | None = None
|
|
embedding: list[float] | None = None
|
|
|
|
|
|
@dataclass
|
|
class Citation:
|
|
"""Maps to the ``citations`` table (Layer 2 — citation graph).
|
|
|
|
``cited_id`` intentionally has no foreign-key constraint in the schema;
|
|
citations to not-yet-ingested papers are preserved as discovery leads.
|
|
"""
|
|
|
|
citing_id: str
|
|
cited_id: str
|
|
context: str | None = None
|
|
|
|
|
|
@dataclass
|
|
class CodeLink:
|
|
"""Maps to the ``code_links`` table (Layer 3 — provenance).
|
|
|
|
``symbol`` is a C++ qualified name or ``file.cpp:line`` reference.
|
|
``role`` is a free-text description such as 'implements Thm 3.2'.
|
|
``id`` is set by the database (BIGSERIAL).
|
|
"""
|
|
|
|
symbol: str
|
|
paper_id: str | None = None
|
|
role: str | None = None
|
|
note: str | None = None
|
|
id: int | None = None
|
|
added_at: datetime | None = None
|
|
|
|
|
|
@dataclass
|
|
class FormulaChunk:
|
|
"""Maps to the ``formulas`` table (F-09 Rich Parsing).
|
|
|
|
Stores a single extracted mathematical formula (LaTeX) from a PDF page.
|
|
``id`` is set by the database (BIGSERIAL).
|
|
``embedding`` is reserved for future pgvector similarity search.
|
|
"""
|
|
|
|
paper_id: str
|
|
page: int
|
|
raw_latex: str
|
|
context: str
|
|
id: int | None = None
|
|
eq_label: str | None = None
|
|
embedding: list[float] | None = None
|
|
|
|
|
|
@dataclass
|
|
class FigureChunk:
|
|
"""Maps to the ``figures`` table (F-09 Rich Parsing).
|
|
|
|
Stores metadata for a single extracted figure from a PDF page.
|
|
``image_path`` points to the saved PNG on disk.
|
|
``id`` is set by the database (BIGSERIAL).
|
|
``embedding`` is reserved for future pgvector similarity search.
|
|
"""
|
|
|
|
paper_id: str
|
|
page: int
|
|
image_path: str
|
|
caption: str
|
|
id: int | None = None
|
|
embedding: list[float] | None = None
|