codex-py/docs/audit/DATA-QUALITY-2026-06-15.md

DATA-QUALITY AUDIT — codex knowledge base (handoff for a fresh session)

Status: baseline scan done 2026-06-15. DQ-1 RESOLVED 2026-06-16 (S2 supplement applied to live DB + wired into ingest). DQ-4 MEASURED 2026-06-16 — retrieval is strong after fixing a P0 codex search paper crash. DQ-2 RESOLVED 2026-06-16 (1911.00966 fully recovered from arXiv; s00454 abstract from S2; book chapter is a documented limit; recovery wired into ingest). DQ-3 RESOLVED 2026-06-16 (content fidelity clean — full coverage, only OCR junk dropped). All four assessment items (DQ-1..DQ-4) DONE. DQ-5 RESOLVED 2026-06-17 (DOI id normalized to bare canonical form in openalex._map_paper; live re-ingest of 10.1007/s00454-019-00132-8 confirmed idempotent → unblocks R-A/R-C). Roadmap R-A..R-E added 2026-06-16 (free-source acquisition levers — see Roadmap section). Author: Audit-Loop (Opus). Intended reader: a cold session with no memory of the audit conversation — everything needed to continue is in this file.

---

Why this exists — the second audit axis

The code/pipeline audit (AUDIT-2026-06-15-loop-1.md, -full-repo.md) verified the loader is correct: IDs are canonical, the citation graph is consistent, the ingest pipeline does what it claims. It did not ask whether the information in the database is actually good. A correct loader can faithfully load thin, incomplete, or unrepresentative data. This document is that second axis — a data-quality assessment of the live corpus, independent of code correctness.

---

How to reach the data (self-contained)

• Live DB: Jetson PostgreSQL via an SSH tunnel (open it first):

  ssh -f -N -L 5433:localhost:5432 alfred@192.168.178.103
  

• Credentials: .env.jetson-ingest (gitignored) holds DATABASE_URL (postgresql://researcher:…@localhost:5433/papers). The researcher role is DML-only (read freely; no DDL).
• Run probes: set -a; source .env.jetson-ingest; set +a then PYTHONPATH=. .venv/bin/python with psycopg + psycopg.rows.dict_row. Reusable domain helpers: codex.quality (is_quality_chunk, _bib_score, classify_section), codex.graph, codex.discover.
• Corpus snapshot (2026-06-15, post canonical-id migration): 29 papers, 1507 chunks, 590 citations.

---

Baseline — what is already GOOD (measured, no action needed)

• Chunk content is clean. Re-running the F-16 quality gate over all stored chunks: 0 / 1507 fail is_quality_chunk (none too-short, none low-alpha / OCR-ish, none bibliography-like). The gate did its job on the .txt ingest.
• Chunk sizing healthy. chars: min 478 / median 2913 / max 4154; 0 exact duplicate chunk bodies.
• section column is low-signal (98 % body; 13 bibliography, 9 theorem, 2 intro, 1 proof). This confirms code-audit R-12 — word-window chunks rarely start at a section header. Not a data defect; the column is just weak.

---

Findings (open investigation items)

DQ-1 — Citation coverage: 12 / 29 papers (41 %) had ZERO citations · RESOLVED 2026-06-16

Resolution (what was done):

• Verified upstream, not a bug. Queried the OpenAlex API directly for all 11 zero-citation papers that have an openalex_id: every one returns referenced_works_count = 0 server-side, W-ids all match what we stored. They are type: preprint (arXiv) / article / dissertation — works OpenAlex indexes without a parsed bibliography. Confirmed source-coverage limit.
• Semantic Scholar supplement applied to the live DB (idempotent backfill, ON CONFLICT (citing_id,cited_id) DO NOTHING): +330 citations (590 → 920). Zero-out-edge papers 12 → 2; in-corpus citing coverage 17/29 → 27/29; graph 489→704 nodes, 590→920 edges; discovery-lead targets 472→677. 9 of the new edges are in-corpus (e.g. 2305.10988 → 10.1007/s00454-019-00132-8).
• Wired into ingest so future re-ingests self-heal: when OpenAlex returns an empty referenced_works, ingest.py now falls back to _s2_reference_supplement() (citing_id rewritten to canonical papers.id, DOI cited-ids lowercased). Same helper drove the one-off backfill.
• Bug fixed (would have crashed the batch): semanticscholar.fetch_references used data.get("data", []), but S2 returns HTTP 200 {"data": null} for a known paper with no parsed refs (e.g. depositonce-5415) → TypeError. Changed to data.get("data") or []. Removed a now-redundant discarded S2 probe in the OpenAlex-404 arXiv branch. New regression test added.
• Irreducible remainder (documented limit): the 2 depositonce theses (10.14279/depositonce-20357, -5415) stay at zero — neither OpenAlex nor S2 indexes references for TU-Berlin institutional DOIs. Accept as inherent.
• Files touched (uncommitted in working tree): codex/sources/semanticscholar.py, codex/ingest.py, tests/ingest/test_ingest.py. Full suite: 331 passed.

Original finding (for context):

• Measured: 12 papers contribute no out-edges; the 590-edge citation graph comes from only ~17 papers. The whole F-15 layer (PageRank / coupling / co-citation / discovery leads) therefore runs on ~59 % of the corpus.
• Sharpening — it is mostly a source-coverage limit, not an ingest bug: 11 of the 12 zero-citation papers do have an openalex_id, i.e. ingest called openalex.fetch_citations(openalex_id) and OpenAlex returned an empty referenced_works list. Only 1911.00966 has no openalex_id (OpenAlex 404 → arXiv/S2 fallback, which also yielded nothing).
• Zero-citation ids: 1005.2698, 1505.01341, 1911.00966 (no oa), 2206.13461, 2305.10988, 2310.17529, 2601.22903, math/0001176, math/0503219, math/0603097, 10.14279/depositonce-20357, 10.14279/depositonce-5415. (Pattern: arXiv preprints + the two depositonce theses — works OpenAlex indexes without parsed references.)
• To investigate next:
  1. Confirm it is OpenAlex coverage, not a fetch_citations bug: for 2–3 of the ids, hit GET https://api.openalex.org/works/<openalex_id> and check whether referenced_works is genuinely empty server-side.
  2. If genuinely empty: enrich via the Semantic Scholar references path (semanticscholar.fetch_references("arXiv:<id>")) as a supplement for OpenAlex-empty papers — S2 often has references where OpenAlex doesn't. Note S2 cited_ids are bare DOI/arXiv, so they flow through the existing RESOLVED_CITATIONS_SQL resolver (audit C-1) fine.
  3. Decide whether the F-15 graph_min_corpus_size warning should also flag low citing-paper coverage, not just paper count.
• Acceptance: either (a) citation coverage materially improves after an S2 supplement, or (b) documented as an inherent OpenAlex-coverage limit with the graph caveated accordingly.

DQ-2 — Metadata gaps: 3 no-abstract, 1 fully metadata-less · RESOLVED 2026-06-16

Resolution (what was done):

• 1911.00966 fully recovered from the arXiv Atom API (OpenAlex 404'd on it): title "A discrete version of Liouville's theorem on conformal maps", Pinkall & Springborn, 2019, 384-char abstract → bibkey PinkallSpringborn2019 auto-generated, abstract embedded (non-zero). It was the worst node in the corpus — now @cite-able, in wiki grounding, and its 10 chunks are visible to chunk search (previously filtered out by bibkey IS NOT NULL). Citations were already added in DQ-1 (27 S2 refs).
• 10.1007/s00454-019-00132-8 abstract recovered from S2 (1186 chars) and re-embedded (targeted UPDATE, not full re-ingest — see DQ-5).
• 10.1007/978-3-642-17413-1_7 (book chapter): documented inherent limit. No abstract in OpenAlex, S2, or Crossref (verified). Left as the single remaining zero-vector paper; user may add an abstract by hand.
• Corpus now: 1 paper without abstract (the book chapter), down from 3+1. Zero-vector search pollution (DQ-4 secondary finding) reduced from 2 tail fillers to 1.
• Wired into ingest so this self-heals: OpenAlex-404 on an arXiv id now recovers metadata via arxiv.fetch_metadata (new); an empty abstract is supplemented via _recover_abstract → S2 then Crossref (new codex/sources/crossref.py). New semanticscholar.fetch_abstract.
• Files touched: codex/sources/arxiv.py, codex/sources/crossref.py (new), codex/sources/semanticscholar.py, codex/ingest.py, + tests. Full suite: 342.

Original finding (for context):

• Measured:
  • No abstract (3): 10.1007/978-3-642-17413-1_7, 10.1007/s00454-019-00132-8, 1911.00966. These get a zero-vector abstract_emb, so paper-level semantic search can't place them.
  • No bibkey / year (1): 1911.00966 only — and its authors array is empty too. This paper is fully degraded (OpenAlex 404 → Paper(id, title="") fallback): no abstract, no bibkey, no year, empty authors, no citations. With no bibkey it cannot be @cited and is invisible to wiki grounding (which keys on bibkey).
• To investigate next:
  1. 1911.00966 — confirm the arXiv id is correct and whether OpenAlex/S2 has it under a different id (DOI?); if recoverable, re-ingest to populate metadata. If not, decide: keep as a degraded node or drop.
  2. The 2 no-abstract DOIs — check if OpenAlex has an abstract_inverted_index that the mapper missed, or if the abstract is genuinely absent upstream.
• Acceptance: every paper has at least a bibkey + non-zero abstract embedding, or the exceptions are documented with rationale.

DQ-3 — Content fidelity (chunks vs source .txt): VERIFIED CLEAN · RESOLVED 2026-06-16

Resolution (read-only; re-ran chunk_text + is_quality_chunk on all 29 source files and compared to stored chunks):

• No paper is silently gutted. Coverage (stored chunk chars / source chars) is 109–114% for every paper — the >100% is exactly the 64/512-word overlap inflation, i.e. full retention. stored == kept for all 29, so the live DB matches what the current pipeline produces.
• Only 18 chunks dropped corpus-wide, all low_alpha (0 short, 0 bib), concentrated in 10.1007/0-387-29555-0_13 (6) and depositonce-5415 (10). Inspected: they are garbled OCR tables from scanned-book sources (alpha 0.30–0.36, e.g. "(u: 1) 7 < u D:= ud3424 — else H®: out.v. Family 1…") — exactly the artefacts the alpha-ratio gate targets. No real math content lost (the handoff's feared failure mode did not occur; math prose stays > 0.40 alpha).
• Head/tail alignment exact on sampled papers: chunk 0 begins at the source head (title/authors/abstract captured), last chunk ends at the source tail → no front/back truncation.
• Note: 0 bibliography drops across the corpus → the upstream .txt extraction had already stripped reference lists, which is why DQ-1's citation graph relies on the API/GROBID sources rather than text-mined refs.
• Acceptance met for all 29 papers (not just a sample). No action needed.
• ⚠ SCOPE LIMIT — DQ-3 covers .txt → chunks ONLY. It treats each source .txt as ground truth and verifies the ingest pipeline preserves it. It does not measure the layer above — PDF/source → .txt — which was done by an upstream tool in ConformalLabpp (outside codex-py) and is unexamined. There is already evidence that layer is lossy: the 18 dropped chunks were garbled OCR tables already present in the .txt (i.e. the PDF→txt step garbled the table/formula regions), and the bibliographies were stripped upstream. So "VERIFIED CLEAN" means the loader faithfully preserves whatever the .txt contains, not that no information was lost from the source paper. See DQ-6.

DQ-6 — PDF/source → .txt extraction fidelity: NOT MEASURED · MED, open

• Open question: how much did the upstream PDF→.txt extraction (done in ConformalLabpp, outside codex-py) drop or garble vs the original papers? Math/equations, tables, multi-column layouts, and figure captions are the usual casualties of PDF text extraction — and we have direct evidence of garble (DQ-3's low_alpha drops were already-garbled tables in the .txt).
• Checkable: the 36 original PDFs are present in …/ConformalLabpp/papers/ (map each paper's source_path .txt to its sibling .pdf by filename stem).
• To investigate: for a sample (incl. the known-bad scanned book 0-387-29555-0_13 and a born-digital arXiv PDF), extract PDF text independently and compare length / page-coverage to the .txt; spot-check whether specific theorem statements & equations survive. Flag papers where the .txt is missing large spans.
• Acceptance: a per-sample source→txt coverage estimate, or a list of papers whose .txt is materially degraded (candidates for re-extraction — note roadmap R-A/R-C would re-ingest from PDF/.tex directly and largely sidestep this).

Original open question (for context):

• Open question: does the stored chunk set faithfully reconstruct each source file, or did the chunker / filter_chunks silently drop material (e.g. an abstract, a section, math-heavy passages)? The whole corpus was ingested from .txt (PAPERS_DIR=/Users/tarikmoussa/Desktop/ConformalLabpp/papers/txt), so the .tex/.pdf paths never ran.
• To investigate next: for a sample of ~5 papers, compare len("".join(stored chunks)) against len(source.txt) (coverage %), and eyeball the first/last chunk vs the file head/tail. Flag papers where coverage is low (content lost) — F-16 dropping >X % of a paper is a signal.
• Acceptance: sampled papers retain ≳ the expected fraction of source text; no paper is silently gutted by the quality gate.

DQ-4 — Retrieval quality · MEASURED 2026-06-16 — good, after fixing a P0 crash

P0 bug found + fixed: codex search paper was crash-broken on the live DB. The CLI passed the query embedding uncast, so pgvector raised operator does not exist: vector <-> double precision[] on every paper search. The working call sites (mcp_server.py, wiki.py) cast %(emb)s::vector; the CLI did not. Fixed both <-> occurrences in codex/cli.py search_paper; added a regression guard in tests/cli/test_cli.py asserting the cast (mocked-DB unit tests can't catch the type error — this is why it survived the code audit).

Relevance verdict (post-fix): strong. Six domain queries, both surfaces (paper-level abstract search + chunk-level). rank-1 was the exactly-correct paper in 6/6 queries for chunk search and 5/6 for paper-level (Laplace–Beltrami put the exact paper at rank-2 behind the closely-related "Polygon Laplacian Made Simple" — acceptable). Examples: "combinatorial Yamabe flow" → math/0306167 (exact, all 5 top chunks that paper); "variational principle for Delaunay triangulations" → math/0603097 (exact); "circle packing rigidity" → 2601.22903 (exact). Three of the DQ-1-rescued papers (2305.10988, 1005.2698, depositonce-5415) now surface as top hits. The KB is trustworthy for lookups.

Secondary finding → reinforces DQ-2: zero-vector abstract pollution. Papers with no abstract get a zero abstract_emb, which sits at constant L2 distance ≈1.000 from every query and appears as filler in the paper-level tail whenever < 5 strongly-relevant papers exist. 1911.00966 (fully degraded, DQ-2) showed up at distance 1.000 in 4/6 queries. Harmless at rank>1 with a visible 1.000 score, but noise — and a concrete reason to fix DQ-2. Chunk-level search is immune (it filters bibkey IS NOT NULL, and 1911.00966 has no bibkey/chunks).

Note: the MCP search docstring claims "hybrid dense + FTS" but the SQL is dense-only; score = 1.0 - dist can go negative (bge-m3 L2 ranges 0–2). Ranking is fine (monotonic); only the absolute score label is misleading.

Original open question (for context):

• Open question: end-to-end, do real queries return relevant results? Clean chunks + a working index don't guarantee useful retrieval.
• To investigate next: run a handful of domain queries through the actual search path (codex search paper "<q>" and the chunk-level MCP search), e.g. "discrete conformal map", "circle packing rigidity", "combinatorial Yamabe flow", "discrete Laplace-Beltrami operator", and judge whether the top-5 hits are on-topic and point at the right papers. Cross-check a couple against --cite-boost to see if the boost helps or hurts on this corpus.
• Acceptance: a short relevance table (query → top-5 → on/off-topic) good enough to trust the KB for lookups, or a list of failure modes to fix.

DQ-5 — ingest_paper is NOT idempotent for DOI papers · RESOLVED 2026-06-17 (found 2026-06-16)

Resolution (what was done):

• Root-cause fix, minimal + contained: added openalex._normalize_doi() and applied it in _map_paper so the DOI branch yields the canonical bare, lower-cased form (strips https://doi.org/ / http://doi.org/ / doi:; lower-cases — DOIs are case-insensitive, matching ingest._norm_cited_id). The no-DOI fallback (OpenAlex W-id URL) is left untouched. _map_paper→fetch_paper →ingest_paper is the only consumer chain, so blast radius is one function.
• Checked nothing depended on the URL form. It was the opposite: the full-URL id silently defeated _s2_id_for(paper.id) and _recover_abstract's paper.id.startswith("10.") (a https://… string is neither 10.- nor doi:-prefixed), so S2/Crossref recovery never fired for DOI papers. The fix repairs those paths too.
• Tests (offline, mocked DB): closed audit T-3 — the OpenAlex fixture now emits the real URL-form DOI and asserts paper.id is the bare form; added _normalize_doi unit tests (https/http/doi:/upper-case/already-bare) and a W-id-fallback test; added ingest regression test_ingest_doi_paper_upserts_bare _canonical_id (runs the real _map_paper through ingest, asserts the papers upsert carries the bare id into %(id)s and uses ON CONFLICT (id)). Verified it has teeth (fails on the pre-fix code). Full suite 348 passed; ruff + mypy --strict clean.
• Live idempotency confirmed (tunnel up): re-ingested 10.1007/s00454-019-00132-8 against the Jetson DB → no UniqueViolation; IngestResult(paper_id='10.1007/s00454-019-00132-8', citations_upserted=45). After-state: UPDATE in place — id and openalex_id unchanged, added_at unchanged (proves no fresh INSERT), exactly 1 row for the DOI (no URL-form duplicate), citations 45→45, total papers 29→29.
• Unblocks R-A (GROBID re-ingest) and R-C (.tex re-ingest) — both re-ingest existing papers. Files touched: codex/sources/openalex.py, tests/sources/test_openalex.py, tests/ingest/test_ingest.py.
• Note (out of scope, pre-existing): arXiv papers stored as bare ids (2305.10988) are still not round-trip idempotent if OpenAlex returns an arXiv-DOI/W-id instead of the bare arXiv id — a separate canonicalization concern, not this UniqueViolation. The DOI path (this finding) is fixed.

Original finding (for context):

• Symptom: re-ingesting an existing DOI paper crashes with UniqueViolation: papers_openalex_id_key. Surfaced trying to re-ingest 10.1007/s00454-019-00132-8 during the DQ-2 backfill.
• Root cause: openalex._map_paper sets Paper.id = data["doi"], which is the full URL https://doi.org/10.1007/…. The stored canonical id is the bare 10.1007/… (post the M-1 canonical-id migration). So INSERT … ON CONFLICT (id) finds no id match, attempts a fresh INSERT, and trips the separate unique constraint on openalex_id (already held by the canonical-id row).
• Confirmed: openalex.fetch_paper("10.1007/s00454-019-00132-8").id == "https://doi.org/10.1007/s00454-019-00132-8" (≠ stored bare id).
• Impact: any re-ingest of a DOI paper aborts. This blocks roadmap R-A (GROBID on PDFs) and R-C (.tex ingest) — both re-ingest existing papers. Also implies new DOI ingests store full-URL ids, inconsistent with the bare-id corpus (latent split-identity risk).
• Proposed fix (handle with care — M-1 incident territory): normalize the DOI in _map_paper to the canonical bare, lower-cased form (strip https://doi.org/), matching what the M-1 migration produced. Add a re-ingest idempotency test over a DOI paper. Did not fix inline this session — the canonical-id derivation is exactly where the M-1 migration incident occurred, so it needs its own focused change + a live re-ingest check.
• Workaround used for DQ-2: targeted UPDATE … SET abstract, abstract_emb instead of full re-ingest.

---

Roadmap — data-acquisition levers (post-audit, all free sources)

These are forward-looking acquisition improvements, distinct from the DQ-1..DQ-4 audit findings (which assessed existing data). They came out of a strategy discussion on 2026-06-16: for this Math/CS corpus the valuable data (references, abstracts, full text) is almost entirely open (arXiv, OpenAlex, Crossref, S2), so the quality lever is combining more free sources, not buying paywall access. Paywall/uni-login was explicitly considered and deferred (see R-E). Each item is self-contained so a cold session can pick it up.

R-A — Run GROBID reference extraction on PDFs · CORE DONE 2026-06-17 (full sweep optional)

Resolution (what was done) — citing coverage 27/29 → 29/29 (100%); edges 920 → 1022:

• References-only backfill, not full re-ingest. Added scripts/ra_grobid_backfill.py: per paper, run grobid.extract_references(pdf), normalize cited-ids to the canonical bare form (DOIs de-URL'd + lower-cased; arXiv: stripped), INSERT into citations (ON CONFLICT DO NOTHING). Deliberately avoids ingest_paper(source_path=pdf), whose PDF branch re-runs Nougat + DELETE FROM chunks and would overwrite the DQ-3-verified-clean .txt chunks (and needs Nougat). Idempotent; dry-run by default; 7 tests; ruff/mypy clean.
• GROBID provisioning: the pinned grobid/grobid:0.8.2 is amd64-only, but the Jetson and the dev Mac are both arm64 (this is why GROBID never ran). Ran the lighter CRF image (grobid/grobid:0.8.2-crf) on the Mac under Podman+Rosetta (PDFs are local; citations written over the SSH DB tunnel). Made extract_references timeout configurable (theses need >60s emulated). A separate task is provisioning a native arm64 GROBID on the Jetson (docs/infra/grobid-jetson.md).
• Both depositonce theses closed: …depositonce-20357 (lutz-2024) +96 edges (5 in-corpus); …depositonce-5415 (sechelmann-2016) +6 edges — its 147 MB/173-page scan blew pdfalto's 120s limit, so the 9 bibliography pages were extracted with PyMuPDF first (77 refs, only 6 with a DOI/arXiv). No paper has zero out-edges now.
• Optional remaining (secondary acceptance): sweep the other 27 already-covered papers for extra GROBID refs via python scripts/ra_grobid_backfill.py --write (the 2 book PDFs need the same bib-page extraction as sechelmann).

Original plan (for context):

• ✓ Unblocked (DQ-5 RESOLVED 2026-06-17): DOI re-ingest is now idempotent, so the GROBID re-ingest no longer aborts on papers_openalex_id_key.
• What: the whole corpus was ingested from .txt (PAPERS_DIR=…/papers/txt), so the GROBID PDF path never ran. Re-ingest with a PDF source_path per paper so codex.parsing.grobid.extract_references parses each bibliography.
• Why: extracts references the APIs lack — a complement to DQ-1's S2 supplement. Crucially, the 2 depositonce theses that stayed at zero in DQ-1 (no API references anywhere) almost certainly have a references section in their open-access PDFs → GROBID could finally give them out-edges. Closes the last 2/29.
• How: no new code — ingest already merges + dedups api_citations + pdf_citations (codex/ingest.py:288). GROBID server is already configured (GROBID_URL=http://192.168.178.103:8070 in .env.jetson-ingest). Download free arXiv PDFs, re-ingest with source_path.
• Acceptance: the 2 theses gain references; published papers gain GROBID-parsed refs beyond OpenAlex/S2 coverage; citing coverage 27/29 → 29/29.

R-B — Add Crossref as a third citation + abstract source · MED lever, MED effort

• What: new codex.sources.crossref client hitting Crossref REST /works/{doi} for the reference array and abstract.
• Why: Crossref carries publisher-deposited references for many DOIs (free, polite pool via mailto) — a third leg after OpenAlex/S2. It also recovers abstracts OpenAlex is contractually barred from redistributing (likely the 2 Springer DOIs missing abstracts in DQ-2). cited_ids are DOIs → flow through the existing graph resolver unchanged.
• How: mirror codex/sources/openalex.py (tenacity retry, polite pool). Extend the ingest fallback chain to OpenAlex-empty → S2 → Crossref (same _s2_reference_supplement shape). Add unit tests with mocked HTTP.
• Acceptance: Crossref recovers refs and/or an abstract for ≥1 paper where OpenAlex+S2 are empty.

R-C — Prefer arXiv .tex source over .txt for chunk fidelity · DONE 2026-06-17

Resolution (what was done) — all 13 arXiv papers re-ingested from .tex:

• Multi-file flatten: arxiv.fetch_source now inlines \input/\include (tex.flatten_inputs) so multi-file projects yield the full body, not the primary file's include skeleton (e.g. math/0603097 = 15 files / 12 includes; 2305.10988 = 6 / 5). Also handles legacy single-file bare-gzipped .tex (no tar wrapper) for old math/* papers (commit 121b582).
• Section-aware chunking: new tex.chunk_sections returns (title, chunk) pairs (chunks never span a \section); the ingest .tex path labels the stored section from the real heading instead of running classify_section on a header-less word-window. Quality-gated as pairs so labels stay aligned.
• Unblocked by the arXiv half of DQ-5 (commit f504ca6): _canonical_id strips the 10.48550/arxiv. DOI prefix so an arXiv paper's id is the bare arXiv id — without it, re-ingest tripped papers_openalex_id_key (confirmed live, then fixed; idempotent re-ingest verified).
• Outcome: section column gained signal — ~34/329 arXiv chunks now intro/theorem/proof (was ~all body); math/0603097 {body:31}→{body:29, proof:9}, 2310.17529 gained proof+intro, etc. Math papers with descriptively titled sections stay body (classify_section's fixed vocab) — modest but real (R-12). No paper gutted (chunk counts stable/▲); F-16 gate unchanged.
• Files: codex/parsing/tex.py, codex/sources/arxiv.py, codex/ingest.py, codex/sources/openalex.py, scripts/rc_tex_reingest.py (driver, dry-run default), + tests. Full suite 370 passed; ruff/mypy clean. Branch feat/tex-ingest (commits 1da81c7, 121b582, f504ca6).

Original plan (for context):

• What: re-ingest from arXiv LaTeX source through the existing codex.parsing.tex.latex_to_text path instead of flattened .txt.
• Why: highest-fidelity input — preserves math, structure, and section boundaries. Directly serves DQ-3 (content fidelity) and audit R-12 (the section column is low-signal because word-window .txt chunks rarely start at a real header; .tex headers fix this).
• How: download free arXiv source tarballs, feed .tex to ingest with source_path. Handle multi-file projects (main .tex + includes).
• Acceptance: section column gains signal (fewer body-only); DQ-3 coverage vs source improves; no regression in the F-16 quality gate.

R-D — F-15: warn on low citing-paper coverage, not just paper count · SMALL

• What: the open DQ-1 sub-item: graph_min_corpus_size only flags total paper count. Add a warning when the share of papers with ≥1 out-edge is low (e.g. < 80%), since that is what actually starves PageRank/coupling.
• Why: at 17/29 the graph silently ran on 59% of the corpus with no signal. Post-DQ-1 it's 27/29 (93%) — but the guard should catch future regressions.
• Acceptance: codex graph report surfaces citing-coverage % and warns below threshold.

R-E — Paywall / institutional access · DEFERRED — only for a paywall-only expansion

• Decision (2026-06-16): a uni login does not help the current corpus — every DQ-1/DQ-2 gap is in openly-available works (arXiv preprints, OA theses), and all metadata/citation sources are free APIs. Revisit only if the corpus expands to papers that exist solely behind a paywall with no preprint; then institutional access → full-text PDF → GROBID refs (R-A) + clean chunks (R-C).
• ToS caveat: manual download of individual papers you have legitimate access to is fine; automated bulk download through a uni proxy (EZproxy/Shibboleth) violates most publisher terms and can get the whole institution's access revoked. Do not wire a publisher login into the ingest pipeline.

---

Priority for the next session

1. DQ-1 — DONE 2026-06-16 (live DB at 920 citations, 27/29 citing coverage; remedy wired into ingest).
2. DQ-4 — DONE 2026-06-16 (P0 search paper crash fixed; relevance verified strong; surfaced the zero-vector pollution that motivates DQ-2).
3. DQ-2 — DONE 2026-06-16 (1911.00966 fully recovered; s00454 abstract from S2; book chapter documented; recovery wired into ingest).
4. DQ-3 — DONE 2026-06-16 (content fidelity clean; full coverage, only 18 OCR-junk chunks dropped, no math lost). Assessment axis DQ-1..DQ-4 complete.
5. DQ-5 — DONE 2026-06-17 (DOI id normalized to bare form in openalex._map_paper; live re-ingest of 10.1007/s00454-019-00132-8 idempotent — UPDATE in place, added_at unchanged; suite 348 green). Unblocks R-A/R-C.
6. Roadmap levers (see section above — now all unblocked): R-A (GROBID on arXiv PDFs — closes the last 2 zero-citation theses), R-B (Crossref refs — abstract half already shipped in DQ-2), R-C (.tex ingest — serves DQ-3), R-D (F-15 coverage warning, quick). R-E (paywall) deferred.

DQ-1..DQ-4 are data/assessment work; DQ-5 + roadmap R-A..R-D add new code (canonical-id fix, re-ingest scripts, Crossref references, a .tex path, a graph warning) on top of the already-remediated pipeline (see the AUDIT-* docs and PRs #12–#14). The DQ-1/DQ-2 ingest wiring + Crossref/arXiv abstract recovery already landed this session.

---

Appendix — reproduce the baseline scan

import os, psycopg, statistics
from collections import Counter, defaultdict
from psycopg.rows import dict_row
from codex.config import Settings
from codex.quality import is_quality_chunk, _bib_score

s = Settings()
with psycopg.connect(os.environ["DATABASE_URL"], row_factory=dict_row) as c:
    papers = c.execute("SELECT id, bibkey, title, abstract, year, authors, openalex_id FROM papers").fetchall()
    chunks = c.execute("SELECT paper_id, content, section FROM chunks").fetchall()
    cit = c.execute("SELECT citing_id, count(*) n FROM citations GROUP BY citing_id").fetchall()

# zero-citation papers (DQ-1)
cited = {r["citing_id"] for r in cit}
zero = [p["id"] for p in papers if p["id"] not in cited]
# metadata gaps (DQ-2): p["abstract"] empty / p["bibkey"] None / not p["authors"]
# F-16 re-gate (baseline): [ch for ch in chunks if not is_quality_chunk(ch["content"], settings=s)]
# section dist (R-12): Counter(ch["section"] or "(null)" for ch in chunks)