prompt: optional-dependency-skip-guards shipped (PyAutoFit#1512) + correct the test_nautilus drift - #260
Merged
Conversation
…orrect the test_nautilus drift Files the task that closes follow-up 4 of complete/2026/08/uniform-prior-bounds-numpy-path.md, filed 2026-08-18 but never turned into a prompt. test_nautilus.py::test__single_core_builds_no_pool has been reported as "pre-existing on clean main" in at least six completion records since 2026-08-16. It is not: the test runs a real search.fit, which imports the optional `nautilus` sampler, and that ships only in the [optional] extra. CI installs [optional] and the test passes there (latest main: 2024 passed, 3 skipped, 0 failed). Envs without the extras get a hard ModuleNotFoundError. Same cause for the astropy collection errors and the nine aggregator errors. - active/17_optional_dependency_skip_guards.md — the prompt, with the evidence from all three envs and the full correction trail across the records - active.md — registry entry, issue #1511, PR #1512 - three completion records annotated with a dated correction rather than rewritten in place, following the prior-support-clipper.md precedent: multistart-gradient-resume-fom-sanity-check, clipper-usage-in-search-summary, uniform-prior-bounds-numpy-path - dashboard regenerated (in-flight 2 -> 3) lifecycle.py check: OK. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01F1uQdHt11NPBc5cXA5cvme
All three legs pass on b814937 (run 32575550710): unittest 3.12, unittest 3.13 and unittest-nojax. PR mergeable state clean, no review comments. Status lines advanced in active.md and the prompt; dashboard regenerated. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01F1uQdHt11NPBc5cXA5cvme
Shipped 2026-08-22, merge commit 14eb8393. Advances the task to complete/2026/08/optional-dependency-skip-guards.md via lifecycle.py record: prompt folded, active.md section removed, index refreshed, dashboard regenerated. The record keeps the part worth remembering: the fix was one line per file, but the diagnosis had already been written down twice (prior-support-clipper.md on 08-17, uniform-prior-bounds-numpy-path.md follow-up 4 on 08-18) and the false alarm still fired three more times, because a follow-up that lives only in a completion record has no owner. lifecycle.py check: OK. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01F1uQdHt11NPBc5cXA5cvme
Jammy2211
pushed a commit
that referenced
this pull request
Aug 22, 2026
Jammy2211
added a commit
that referenced
this pull request
Aug 22, 2026
…-hontn3 prompt: jax-grad-local-vs-ci-assertions (#260) — route, root cause, fix
Jammy2211
pushed a commit
that referenced
this pull request
Aug 23, 2026
The refresh check on PR #279 was right: my PyAutoBrain checkout was one commit behind main (#260, which changed the board logo markup), so my render differed from the one CI produces. Regenerated against main. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_015qk7hoavMnFyPtW4toYn8K
Jammy2211
pushed a commit
that referenced
this pull request
Aug 24, 2026
PyAutoHands#261 merged (c0e2e53) with CI green on all three matrix legs plus the tenant-firewall step in each; PyAutoMind#280 merged (0cd8544). The task stays ACTIVE on purpose. Issue #260 scopes seven workspace collapses (4 *_workspace_test, then 3 workspace) that have zero commits, so recording this complete would misrepresent it. Registry now carries both merged PR URLs and names the remaining scope. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01UpSFum81Jeq9KZ9wdKtaeZ
Jammy2211
pushed a commit
that referenced
this pull request
Aug 24, 2026
All ten vendored .github/scripts/run_smoke.py copies are now thin delegators over the PyAutoHands runners. 2085 lines -> 1127, and most of what remains is the docstring explaining why each file delegates. Merged, each verified from the CI log rather than the green tick: PyAutoHands#261 --list opt-in lists + the notebook leg PyAutoHands#262 allowlist authoritative (the 13-script finding) PyAutoHands#263 regenerate_notebook resolves by relative path autolens_workspace_test#273, autogalaxy_workspace_test#111, autofit_workspace_test#90, autocti_workspace_test#16 (198 -> 77) autofit_workspace#147, autogalaxy_workspace#226, autolens_workspace#500 (356 -> 119) The record keeps the two findings that mattered most: the no_run-wins rule shipped in #261 would have silently dropped 13 scripts from smoke coverage with CI still green, and regenerate_notebook's bare-filename lookup made the stale-notebook recovery dead for every real (nested) notebook. Both were caught by measuring the real repos before writing a delegator, and both are now pinned by tests that were negative-tested against the old behaviour. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01UpSFum81Jeq9KZ9wdKtaeZ
Jammy2211
added a commit
that referenced
this pull request
Aug 24, 2026
prompt: ship smoke-runner-delegation (#260) → complete
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
1 participant
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.
Mind-side record for PyAutoFit#1511 → #1512 (merged 2026-08-22,
14eb8393).What landed
complete/2026/08/optional-dependency-skip-guards.md— the completion record, written vialifecycle.py record(prompt folded in,active.mdsection removed,complete/index.mdrefreshed)## Correctionblock rather than rewritten in place, following theprior-support-clipper.mdprecedent:multistart-gradient-resume-fom-sanity-check,clipper-usage-in-search-summary,uniform-prior-bounds-numpy-pathdashboard.md/dashboard.htmlregeneratedlifecycle.py check: OK.Why the correction blocks
test_nautilus.py::test__single_core_builds_no_poolwas reported as "pre-existing on cleanmain" in at least six records from 2026-08-16 onward. It never was: the test runs a realsearch.fit, which imports the optionalnautilussampler, and that ships only in the[optional]extra. CI installs[optional]and the test passed there throughout — main was 2024 passed / 3 skipped / 0 failed the whole time.Two records already had it right (
covariance-interpolator-rng-seed.mdnamed the actualModuleNotFoundError;prior-support-clipper.md:218issued the correction) and are left untouched.The bit worth keeping
The fix was one line per file, and it had already been diagnosed correctly twice — on 08-17 in
prior-support-clipper.md, and on 08-18 as follow-up 4 ofuniform-prior-bounds-numpy-path.md, which proposed exactly this fix. Both lived only in completion records, which nothing reads on the way into a new task, so the false alarm fired three more times after the right answer was written down.A follow-up that stays in a record has no owner. That is the finding this record is really for.
🤖 Generated with Claude Code
https://claude.ai/code/session_01F1uQdHt11NPBc5cXA5cvme
Generated by Claude Code