Fix broken links from link-checker report (#751)

Merged
LU
4mo ago
+429-510

Fixes the link issues from the weekly link-checker report in #751. All replacement URLs and DOIs were verified live before being applied (DOI Handle API / Crossref for DOIs; browser-UA requests, Wayback availability API, and a real-browser check for ambiguous cases).

What was fixed

Broken links (~150)

  • Malformed URLs (%29 paren auto-linking, jammed double links, trailing spaces/dots) fixed in the markdown following the pattern from #715.
  • Moved sites re-pointed to their successors: Stanford DMP → guides.library.stanford.edu/dmps, ICMJE recommendations, i4replication.org/papers, SAGE research-tools directory, GIDA CARE principles, slovakrn.org, understood.org, gida-global.org/careprinciples, psychologytoday /intl/→/us/, and more.
  • Genuinely dead pages (domain gone / deleted, verified) replaced with Wayback Machine snapshots: replicationmarkets, FOSTER portal pages, remotestudentexchange, pictureascientist, riotscience, annameier, karolinaurbanska, Everything Hertz (site currently unreachable even in a real browser), citeblackwomencollective, criticalethnicstudiesjournal, and others.
  • /reversals/: dead CiteSeerX, Sci-Hub, EBSCO-session, silverchair-token and Elsevier-OAuth links replaced with verified doi.org links for the cited papers; two no-DOI items (Biomedicine article, Behind the Shock Machine) got Wayback snapshots.
  • Removed unfilled USERNAME/PERSON-ID placeholder social links from 9 author pages (the checker's "in 12 pages" entry).
  • /booklet/: dead themify-icons@0.1.2 CDN css replaced with the maintained @icon/themify-icons package (verified css+fonts load).
  • Fixed FORRT YouTube handle (@forrtproject → @forrtproject9995).
  • Open Research Toolkit (educators-corner 008): all 14 module videos were deleted from YouTube; links now point to the author's archived recordings on OSF (osf.io/a4ftw, ORT_Recorded_Modules).

Broken DOIs (34)

  • Truncated/typo'd DOIs corrected (e.g. 10.1177/251524592513361 → 10.1177/25152459251336130, sciadv.adp604 → adp6040, Lakens s41586 → d41586), junk suffixes stripped (?error=cookies…, /full, anchors, %20, en-dash), placeholder your_doi_placeholder resolved to the actual PsyArXiv DOI (10.31234/osf.io/suzej), and the Albornoz "Framing power" chapter re-cited to its registered ELPUB 2018 version.
  • Five `,authors entries are checker parsing artifacts of the cluster-page JS embedding — the stored DOIs are clean and valid; no source change.

Publisher URLs → doi.org (91)

  • Converted per the approach in #759, incl. ScienceDirect PII→DOI via Crossref alternative-id and psycnet fulltext IDs resolved by context. One skipped (Pegado & Grainger 2019 psycnet PDF — no confident DOI match).

Upstream sources fixed too (per review guidance: sheet, not script)

Generated content was fixed at its true source so the daily regeneration keeps the fixes:

  • "FORRT Database 2.0" (curated resources, gid 1924034107): 67 cells corrected (URL/DOI column + abstracts with paren-linked URLs).
  • "FORRT Clusters - Website Source" (Publications tab): 8 cells corrected (4 DOIs + matching APA references).
    No generator scripts were modified.

Left unchanged (verified false positives)

Bot-blocking, not breakage: WSJ share links (anti-bot interstitial; articles live), Unsplash 401s, ORCID (valid via public API), dimensions.ai, alexandria.unisg.ch, par.nsf.gov, oasis.geneseo.edu, oer.deepwebaccess.com, gcbialyzer, badscience.net, firstmonday, scipost, preclinicaltrials.eu, hdl.handle.net, dpjedi.org, phil.vt.edu PDF, ron.dotsch.org (http). The /welcome/ TIMEOUT entries and the collapsed 403 section were skipped as checker noise per the report itself.

Needs maintainer action (could not be fixed from the repo)

  • FRED data-curator application form (/replication-hub/fred_data_curators_role/): linked as a private /edit URL and the /viewform variant also requires sign-in — the form needs to be published publicly and the link updated.
  • Syllabus submission Google Form (/syllabus/, also embedded): public viewform URL redirects to sign-in — form is closed/restricted; reopen or remove the embed.
  • Three deleted Google Docs/Sheets resources in curated resources (hackathon doc 1x-Q6pE…, open-research good-practice doc 1Vblnwym…, preregistration sheet 1vlp5GN…): 410-deleted, no usable archive; the sheet rows should probably be retired.
  • The Economist "Unlikely Results" video (curated resource): deleted from YouTube, no archive or re-upload found.
  • One /reversals/ reference (Patson et al. 2016, dead EBSCO session URL) could not be confidently matched to a DOI; left unchanged rather than risk a wrong citation.

Verification: hugo --renderToMemory builds cleanly; all edited JSON files re-validated; every replacement URL/DOI checked live.

Closes #751

🤖 Generated with Claude Code