Repository navigation
Fix extraction push race and churn; pin deps; bump actions - #5
Merged
Merged
Conversation
- extract-pdf-text: queue runs (concurrency group), check out latest main, and rebase-and-retry the push, so merges landing close together no longer fail with a rejected push. - extract_pdf_text.py: skip files whose extracted text is unchanged instead of rewriting every file to bump the extracted: date. - Pin pyyaml==6.0.3 and pypdf==6.21.4 (reproduces the committed _extracted/ text exactly), so output only changes when a PDF does. validate-corpus now installs from requirements.txt too. - actions/checkout v4 -> v5, actions/setup-python v5 -> v6 (Node 24). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_015bzez2N6vdSh6qXV7Y44P9
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up fixes for the corpus workflows, from what showed up after merging #1–#4.
Extraction push race. When #2 and #3 merged seconds apart, two extraction runs started together and the first one's push to
mainwas rejected. Now runs queue (concurrency: extract-pdf-text), check out the latestmainrather than the triggering commit, and rebase-and-retry the push if a merge still lands mid-run.Extracted-text churn.
extract_pdf_text.pyno longer rewrites a file whose extracted text is unchanged. Before, every run touched every file just to bump theextracted:date in the header.pypdfandpyyamlwere unpinned. Now pinned topypdf==6.21.4/pyyaml==6.0.3. I checked that 6.21.4 reproduces the current committed_extracted/text byte-for-byte, so merging this causes no rewrite.validate-corpusnow installs fromrequirements.txttoo.Node 20 deprecation.
actions/checkoutv4 → v5 andactions/setup-pythonv5 → v6 across all three workflows.Tested locally: with the pinned versions, extraction over the whole corpus writes 0 files; a file differing only in its date header is left alone; a file with changed text is rewritten; the validator passes (21 resources); all workflow YAML parses.
Not included: the
CORPUS_BOT_TOKENsecret, which a repo admin needs to add, for inbox PR checks re-running on the bot's commit.Generated by Claude Code