Commit Graph

7 Commits

Author SHA1 Message Date
Tarik Moussa
5dfa6e4bae feat(graph): paper_identifiers table — multi-ID alias support
Fixes preprint/journal duplicate problem: arXiv papers were ingested
under their preprint OpenAlex ID, but corpus citations reference the
published-version OpenAlex ID → those citations were invisible in the
graph (appeared as dangling).

Changes:
- infra/schema.sql: CREATE TABLE paper_identifiers (paper_id, openalex_id)
  with UNIQUE index; seeded from papers.openalex_id on migration
- codex/graph.py: build_citation_graph LEFT JOINs paper_identifiers as
  a second resolution path: COALESCE(p.id, pi.paper_id, c.cited_id)
- codex/ingest.py: every ingest inserts openalex_id into paper_identifiers
  (ON CONFLICT DO NOTHING) — aliases can be added manually alongside

Live DB: 35 rows seeded + 4 published-version aliases added:
  math/0503219 → W2163787581 (DCG 2007)
  math/0306167 → W2136126748 (Commun Analysis Geom 2004)
  math/0203250 → W1567166970 (Trans AMS 2003)
  1005.2698    → W1511400044 (Geom & Topol 2015)

Effect: dangling 574→570; Bobenko+Springborn 2007 now IN-KB rank 9.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 07:42:21 +02:00
9647897173 feat(F-16): chunk quality gate + section classification
3-signal quality filter (length/alpha-ratio/bib-score) + rule-based section classifier + retroactive CLI pass. 35 new tests, 287 total green.
2026-06-15 01:22:44 +00:00
Tarik Moussa
b87087d3c6 fix(ingest): D-04 bibkey heuristic + D-05a/b/c regression tests + __main__
D-04: add _make_bibkey(paper) — ConcatSurnames+Year, ≤3 authors full list,
      >3 authors FirstEtAlYear. Applied in ingest_paper when paper.bibkey is
      None. ON CONFLICT now fills NULL bibkeys via COALESCE(papers.bibkey,
      EXCLUDED.bibkey) while preserving manually-set values.

D-05b: add codex/__main__.py so `python -m codex` dispatches to the CLI.

D-05c: fix test_ingest.py:75 — rename var to raw_api_citations (the raw
       OpenAlex payload with citing_id=openalex_id) and add regression
       assertion: citing_id written to DB must equal paper.id (DOI), not
       paper.openalex_id.

D-05a (proxy): add TestEntryPoint.test_python_m_codex_help — runs
       `sys.executable -m codex --help` via subprocess, covering the
       non-uv-run entry path. 253 tests green, ruff+mypy clean.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 03:07:06 +02:00
Tarik Moussa
5025ff3a88 fix(ingest): support .txt sources + fix citing_id FK violation
Two bugs exposed during first batch ingest run after D-03 fix:

1. `.txt` files were silently skipped ("Unbekannter Dateityp") — added
   plain-text read path alongside .tex and .pdf.

2. After D-03, OpenAlex resolves arXiv IDs → paper.id uses the DOI
   (https://doi.org/…) but fetch_citations emits citing_id=openalex_id
   (https://openalex.org/W…). FK constraint citations_citing_id_fkey
   → papers.id violated. Fix: rewrite citing_id to paper.id after
   fetching from OpenAlex.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-14 15:54:05 +02:00
Tarik Moussa
cd42e9abc8 fix(F-09): address review-gate findings — resource leak + idempotency
- mathpix.py + figures.py: wrap fitz page-loop in try/finally so
  doc.close() is guaranteed even on exception (no file-handle leak)
- ingest.py: DELETE FROM formulas/figures WHERE paper_id before re-insert
  so --rich re-ingest is idempotent (BIGSERIAL has no natural UNIQUE key;
  ON CONFLICT DO NOTHING was a no-op)

Gate: 158 passed, ruff clean, mypy clean

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-14 02:23:21 +02:00
Tarik Moussa
1df9be6563 feat(F-09): rich parsing — formula + figure extraction
- 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>
2026-06-14 01:16:24 +02:00
Tarik Moussa
5344975bb1 feat(ingest): end-to-end idempotent ingest pipeline
Implements codex/ingest.py with:
- OpenAlex primary / Semantic Scholar fallback metadata fetch
- Abstract dense embedding (zero-vector for missing abstracts)
- Idempotent paper upsert (ON CONFLICT DO UPDATE)
- Source file parsing (.tex via latex_to_text, .pdf via nougat + grobid)
- Chunk deletion + bulk re-insert with dense embeddings
- Citation merge from OpenAlex/S2 API and GROBID PDF refs (dedup via set)
- 7 tests covering basic, tex, pdf, not-found, idempotent, arxiv-fallback,
  and no-abstract scenarios (all mocked, offline)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-05 07:04:12 +02:00