Repository navigation
analytic_gaussian: pin leg A seed 140 zero-iteration convergence (PyAutoFit#1711) - #115
Merged
Merged
Conversation
…utoFit#1711) Adds an EP-only leg on seed 140 of the exactly-Gaussian leg A, where the mu prior factor converges at its start point (NO_CHANGE, iter=0). Asserts closed-form parity row by row including widths, flag health (every factor ends with a SUCCESS or only accepted zero-iteration NO_CHANGE updates) and that the library STALE FACTORS warning did not fire. Adds ~2 s. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Regression for PyAutoFit#1711. On the analytic_gaussian wave-1 ensemble, leg A seed 140 ended with the mu prior factor recording only
NO_CHANGEupdates: its Laplace optimiser met the gradient tolerance at the start point (iter=0), the projection was accepted and the message moved to the exact prior, and mu matches the closed form to a = 6e-5, b = 2e-5. This leg pins that as a converged (not stale) state: the EP column alone on seed 140 underprojection="moments", judged row by row against the closed form at the leg-A tolerance (widths included), plus two flag-health checks (every factor ends with ≥1 SUCCESS or only accepted zero-iteration NO_CHANGE rows withupdatedtrue;STALE FACTORSabsent from the log andep_diagnostics.results). Adds ~2 s to a script already insmoke_tests.txt.The checks hold against the released PyAutoFit too (
ep_history.csvalready wroteupdated=Trueon NO_CHANGE rows post-projection), so no release gate applies; PyAutoFit#1712 aligns the Newton status with that.Scripts Changed
scripts/graphical/analytic_gaussian.py— new docstring section and a final "Leg A, seed 140" EP-only leg:_WarningRecorderlogging handler,ep_history_rowsreader, parity rows + flag-health + STALE-absence checks; CI-cap note updatedUpstream PR
PyAutoLabs/PyAutoFit#1712
Test Plan
scripts/graphical/analytic_gaussian.pyrun locally against the worktree PyAutoFit: seed-140 legcells passed: 8/8, both flag checks PASS, EP 1.9 s; whole scriptPARITY: PASS (49/49), wall 158 s (< 300 s cap)Generated by the PyAutoLabs agent workflow.
🤖 Generated with Claude Code