fix: add bounded OpenAlex and arXiv HTML fallbacks

[djimit]

Summary\n\n- add bounded retry/backoff for transient OpenAlex 429/5xx responses\n- honor Retry-After with a capped delay\n- fall back to official arXiv HTML when PDF/full text is unavailable\n- sanitize HTML to usable research text and select preferred full-text candidates explicitly\n- add focused regression coverage\n\n## Validation\n\n- TAP version 13

Subtest: abstractFromInvertedIndex reconstructs OpenAlex abstracts

ok 1 - abstractFromInvertedIndex reconstructs OpenAlex abstracts

duration_ms: 0.561208
type: 'test'
...

Subtest: fetchOpenAlexWorks bounds provider calls with an abort signal

ok 2 - fetchOpenAlexWorks bounds provider calls with an abort signal

duration_ms: 8.071958
type: 'test'
...

Subtest: fetchOpenAlexWorks retries a rate-limited provider once with bounded backoff

ok 3 - fetchOpenAlexWorks retries a rate-limited provider once with bounded backoff

duration_ms: 2.869834
type: 'test'
...

Subtest: PaperRank slug and limit parsing are stable

ok 4 - PaperRank slug and limit parsing are stable

duration_ms: 0.457334
type: 'test'
...

Subtest: computePageRank gives cited papers more local graph prestige

ok 5 - computePageRank gives cited papers more local graph prestige

duration_ms: 0.2395
type: 'test'
...

Subtest: extractEvidenceSpans returns marker offsets and surrounding source text

ok 6 - extractEvidenceSpans returns marker offsets and surrounding source text

duration_ms: 0.357125
type: 'test'
...

Subtest: extractEvidenceSpans requires marker boundaries

ok 7 - extractEvidenceSpans requires marker boundaries

duration_ms: 0.143875
type: 'test'
...

Subtest: extractPaperContentText normalizes common alphaXiv content payload shapes

ok 8 - extractPaperContentText normalizes common alphaXiv content payload shapes

duration_ms: 0.070209
type: 'test'
...

Subtest: fetchArxivPaperContent converts official HTML into bounded text

ok 9 - fetchArxivPaperContent converts official HTML into bounded text

duration_ms: 1.037958
type: 'test'
...

Subtest: extractFullTextSections preserves canonical section offsets

ok 10 - extractFullTextSections preserves canonical section offsets

duration_ms: 0.548917
type: 'test'
...

Subtest: evaluatePaperRubric answers checklist items from extracted sections

ok 11 - evaluatePaperRubric answers checklist items from extracted sections

duration_ms: 0.361083
type: 'test'
...

Subtest: buildFullTextAccessPlan records legal source-specific candidates

ok 12 - buildFullTextAccessPlan records legal source-specific candidates

duration_ms: 1.325709
type: 'test'
...

Subtest: normalizeOpenAlexWorks keeps repository PDFs from all OpenAlex locations

ok 13 - normalizeOpenAlexWorks keeps repository PDFs from all OpenAlex locations

duration_ms: 0.168208
type: 'test'
...

Subtest: normalizeOpenAlexWorks drops non-http provider URLs and canonicalizes DOI links

ok 14 - normalizeOpenAlexWorks drops non-http provider URLs and canonicalizes DOI links

duration_ms: 0.154083
type: 'test'
...

Subtest: resolvePaperAccess writes Markdown-safe provider titles and links

ok 15 - resolvePaperAccess writes Markdown-safe provider titles and links

duration_ms: 2.157542
type: 'test'
...

Subtest: normalizeOpenAlexWorks extracts arXiv ids from secondary OpenAlex access locations

ok 16 - normalizeOpenAlexWorks extracts arXiv ids from secondary OpenAlex access locations

duration_ms: 0.109958
type: 'test'
...

Subtest: fetchEuropePmcPaperContent discovers PMCID from DOI before fetching XML

ok 17 - fetchEuropePmcPaperContent discovers PMCID from DOI before fetching XML

duration_ms: 0.896833
type: 'test'
...

Subtest: runPaperRank full-text enrichment discovers Europe PMC content from DOI-only OpenAlex papers

ok 18 - runPaperRank full-text enrichment discovers Europe PMC content from DOI-only OpenAlex papers

duration_ms: 11.351916
type: 'test'
...

Subtest: scorePapers separates impact, graph, methodology, and reproducibility

ok 19 - scorePapers separates impact, graph, methodology, and reproducibility

duration_ms: 0.652542
type: 'test'
...

Subtest: scorePapers excludes graph prestige when local citation edges are absent

ok 20 - scorePapers excludes graph prestige when local citation edges are absent

duration_ms: 0.655
type: 'test'
...

Subtest: expandCitationNeighborhood adds outgoing references and incoming citations

ok 21 - expandCitationNeighborhood adds outgoing references and incoming citations

duration_ms: 0.526333
type: 'test'
...

Subtest: generatePaperCritiques creates span-grounded review questions

ok 22 - generatePaperCritiques creates span-grounded review questions

duration_ms: 0.900208
type: 'test'
...

Subtest: generateFieldMap identifies clusters and paper roles

ok 23 - generateFieldMap identifies clusters and paper roles

duration_ms: 0.522875
type: 'test'
...

Subtest: generateRankSensitivity records rank movement across weighting profiles

ok 24 - generateRankSensitivity records rank movement across weighting profiles

duration_ms: 1.164708
type: 'test'
...

Subtest: generateScoreCalibration evaluates researcher read-order preferences

ok 25 - generateScoreCalibration evaluates researcher read-order preferences

duration_ms: 0.929292
type: 'test'
...

Subtest: generateReproductionEvidenceLedger records completed reproduction notes separately

ok 26 - generateReproductionEvidenceLedger records completed reproduction notes separately

duration_ms: 0.627875
type: 'test'
...

Subtest: buildModelSynthesisPacket creates a bounded auditable model handoff

ok 27 - buildModelSynthesisPacket creates a bounded auditable model handoff

duration_ms: 1.332708
type: 'test'
...

Subtest: resolvePaperAccess writes bounded paper-access artifacts

ok 28 - resolvePaperAccess writes bounded paper-access artifacts

duration_ms: 1.137917
type: 'test'
...

Subtest: resolvePaperAccess escapes provider full-text source labels in Markdown reports

ok 29 - resolvePaperAccess escapes provider full-text source labels in Markdown reports

duration_ms: 0.681084
type: 'test'
...

Subtest: enrichPapersWithFullText redacts fetcher failure messages

ok 30 - enrichPapersWithFullText redacts fetcher failure messages

duration_ms: 0.347375
type: 'test'
...

Subtest: resolvePaperAccess rejects unrelated OpenAlex hits for arXiv identifiers

ok 31 - resolvePaperAccess rejects unrelated OpenAlex hits for arXiv identifiers

duration_ms: 1.867541
type: 'test'
...

Subtest: resolvePaperAccess enriches arXiv fallback records from the arXiv API

ok 32 - resolvePaperAccess enriches arXiv fallback records from the arXiv API

duration_ms: 1.688042
type: 'test'
...

Subtest: resolvePaperAccess accepts OpenAlex hits whose arXiv id is only in secondary locations

ok 33 - resolvePaperAccess accepts OpenAlex hits whose arXiv id is only in secondary locations

duration_ms: 0.770292
type: 'test'
...

Subtest: resolvePaperAccess checks multiple OpenAlex candidates for arXiv identifiers

ok 34 - resolvePaperAccess checks multiple OpenAlex candidates for arXiv identifiers

duration_ms: 0.525542
type: 'test'
...

Subtest: resolvePaperAccess treats title-like arXiv-shaped numbers as title search text

ok 35 - resolvePaperAccess treats title-like arXiv-shaped numbers as title search text

duration_ms: 0.462083
type: 'test'
...

Subtest: resolvePaperAccess checks multiple OpenAlex candidates for title searches

ok 36 - resolvePaperAccess checks multiple OpenAlex candidates for title searches

duration_ms: 0.895791
type: 'test'
...

Subtest: resolvePaperAccess rejects unrelated OpenAlex title search results

ok 37 - resolvePaperAccess rejects unrelated OpenAlex title search results

duration_ms: 0.471541
type: 'test'
...

Subtest: resolvePaperAccess treats title-like DOI substrings as title search text

ok 38 - resolvePaperAccess treats title-like DOI substrings as title search text

duration_ms: 0.687916
type: 'test'
...

Subtest: resolvePaperAccess keeps explicit DOI inputs on the DOI lookup path

ok 39 - resolvePaperAccess keeps explicit DOI inputs on the DOI lookup path

duration_ms: 0.734875
type: 'test'
...

Subtest: resolvePaperAccess keeps explicit OpenAlex IDs on the OpenAlex lookup path

ok 40 - resolvePaperAccess keeps explicit OpenAlex IDs on the OpenAlex lookup path

duration_ms: 0.5
type: 'test'
...

Subtest: resolvePaperAccess keeps explicit PMID inputs on the PMID lookup path

ok 41 - resolvePaperAccess keeps explicit PMID inputs on the PMID lookup path

duration_ms: 0.539375
type: 'test'
...

Subtest: resolvePaperAccess keeps explicit PMCID inputs on the PMCID lookup path

ok 42 - resolvePaperAccess keeps explicit PMCID inputs on the PMCID lookup path

duration_ms: 0.410416
type: 'test'
...

Subtest: runPaperRank writes durable report, data, graph, and provenance artifacts

ok 43 - runPaperRank writes durable report, data, graph, and provenance artifacts

duration_ms: 9.19375
type: 'test'
...

Subtest: runPaperRank writes Markdown-safe topic headings and shell snippets

ok 44 - runPaperRank writes Markdown-safe topic headings and shell snippets

duration_ms: 3.451
type: 'test'
...

Subtest: runPaperRank escapes script-breaking graph artifact data

ok 45 - runPaperRank escapes script-breaking graph artifact data

duration_ms: 2.103
type: 'test'
...

Subtest: runPaperRank redacts model synthesis failure messages from artifacts

ok 46 - runPaperRank redacts model synthesis failure messages from artifacts

duration_ms: 2.175875
type: 'test'
...

Subtest: feynman rank works end to end through the CLI with a fixture source

ok 47 - feynman rank works end to end through the CLI with a fixture source

duration_ms: 1708.422625
type: 'test'
...

Subtest: feynman rank default output is concise and decision-first

ok 48 - feynman rank default output is concise and decision-first

duration_ms: 961.6535
type: 'test'
...

Subtest: feynman rank writes default outputs under --cwd

ok 49 - feynman rank writes default outputs under --cwd

duration_ms: 1631.02725
type: 'test'
...

Subtest: feynman paper works end to end through the CLI with a fixture source

ok 50 - feynman paper works end to end through the CLI with a fixture source

duration_ms: 1587.384083
type: 'test'
...

Subtest: feynman paper default output names the best access route

ok 51 - feynman paper default output names the best access route

duration_ms: 965.927875
type: 'test'
...

Subtest: feynman paper writes default outputs under --cwd

ok 52 - feynman paper writes default outputs under --cwd

duration_ms: 1700.056875
type: 'test'
...

Subtest: feynman rank accepts a preference file through the CLI

ok 53 - feynman rank accepts a preference file through the CLI

duration_ms: 1703.431166
type: 'test'
...

Subtest: feynman rank accepts reproduction notes through the CLI

ok 54 - feynman rank accepts reproduction notes through the CLI

duration_ms: 1992.54175
type: 'test'
...
1..54

tests 54

suites 0

pass 54

fail 0

cancelled 0

skipped 0

todo 0

duration_ms 12411.798834 — 54 passed\n-

@advaitpaliwal/feynman@0.3.49 typecheck
tsc --noEmit && npm run typecheck:workbench-web

@advaitpaliwal/feynman@0.3.49 typecheck:workbench-web
tsc -p workbench-web/tsconfig.json --noEmit — passed\n- — passed\n\nThe full suite is not claimed green because unrelated provider/model environment tests require external credentials or native runtime assumptions.

[vercel]

@djimit is attempting to deploy a commit to the Advait Paliwal's projects Team on Vercel.

A member of the Team first needs to authorize it.