docs: sync README/DESIGN.md against actual repo state - #17
Merged
Conversation
- Researched catalogue: README said 8/20, actual is 23 (18/20 measurement titles + 5 more). Updated every count/list that still referenced 8. - Disclosed the "+ Suggest a movie" auto-publish path (writes straight to content/researched/, no review step -- a deliberate, documented risk per webapp/app.py's own module docstring). README previously implied every path into content/researched/ went through human review, which stopped being true on 2026-08-19. - CI has been failing on every push since 2026-08-19 -- test_auto_publish.py/ test_groq_retry.py import fastapi/groq/httpx without pytest.importorskip, and tests.yml never installs those, so collection is interrupted before any test runs. README was claiming "18/26 passing in CI." Not fixed here (product/CI-config change, out of scope for a docs-only pass) -- just documented. - Corrected the test count (26 -> 35, verified by running the suite locally: 27 pass with neither optional dep installed, 31 with scikit-learn, 35 with both plus sentence-transformers). - Noted Groq's 2026-08-18 model decommission next to the LLMJudge comparison numbers that cite the now-gone model names (numbers are historically accurate, just no longer what running the command today would use). - docs/DESIGN.md: added D17 for the auto-publish decision, closed the "Pending" bullet that still said this wasn't built. No product or measurement-track code touched. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…nstalled tests.yml only installs pydantic/pytest/pyyaml, so test_auto_publish.py and test_groq_retry.py failed collection outright since 2026-08-19, interrupting the whole suite. Both now use pytest.importorskip instead of a bare import, matching how every other optional-dependency test in this repo already handles it. Verified both ways: skips cleanly with CI's actual minimal install (18 passed, 4 skipped), and runs for real with fastapi/httpx/groq present (27 passed, 2 skipped -- the sentence-transformers-only ones).
serpeigd
marked this pull request as ready for review
August 30, 2026 22:38
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
2 participants
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.
Docs-only sync. Verified against the working tree, not the changelog — folds in and supersedes PR #16 (same finding, verified independently against the current tip of
main).content/researched/, no review step — a deliberate, documented risk perwebapp/app.py's own module docstring). README previously implied every path intocontent/researched/went through human review, which stopped being true on 2026-08-19.test_auto_publish.py/test_groq_retry.pyimportfastapi/groq/httpxwithoutpytest.importorskip, andtests.ymlnever installs those, so collection is interrupted before any test runs. README was claiming "18/26 passing in CI." Not fixed here (product/CI-config change, out of scope for a docs-only pass) — just documented.scikit-learntoo, 35 withsentence-transformersas well).llama-3.3-70b-versatile/llama-3.1-8b-instantboth gone) next to theLLMJudgecomparison numbers that cite those model names — numbers stay historically accurate, just flagged as not what running the command today would use (current default:openai/gpt-oss-120b/openai/gpt-oss-20b).docs/DESIGN.md: added D17 for the auto-publish decision, closed the "Pending" bullet that still said this wasn't built.No product or measurement-track code touched.
CI note: because of the pre-existing collection failure above, this PR's own CI run is expected to fail (unrelated to this diff — a docs-only change can't fix it, and fixing
tests.yml/addingpytest.importorskipis a code change outside this task's scope). Per the standing "merge only when CI is green" rule, this won't be self-merged; flagging for a decision on the workflow fix.Generated by Claude Code