Repository navigation
fix(scale): retain measured median-passing probes after convergence - #137
Closed
austinjiann wants to merge 1 commit into
Closed
austinjiann wants to merge 1 commit into
austinjiann wants to merge 1 commit into
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
When visibility changes during geometric-mean convergence, the final probe can fail the median band even though an earlier measured probe passed it. Retain the earlier passing probe closest to unit geometric mean only after normal convergence and only when the final medians fail. Preserve its scale, rows, distance bands and figure together, with selection provenance.
The city replay exposed this: scale 0.722908 passed every median, while the converged 0.735632 failed one. The current playable city preview uses a separately measured explicit scale of 0.726089; this PR does not silently replace its placement. Depth spread still fails and quality remains unaccepted. No thresholds, retry counts, or provider calls change.
Validation: 22 focused tests and 2,764 full Python tests passed; Ruff and format checks passed. Regression cases cover overwritten diagnostic filenames, consistent evidence, unchanged explicit scales, nonconvergence, final-probe preference, and independent spread failure.
Stacked on #136. No footage, generated assets or evidence committed.