Repository navigation
The coordinated re-run: settings layer, refined peaks, the whole solved camera - #21
Merged
Merged
Conversation
…ement on
- settings.get reads environment, then a profile (configs/<name>.toml via FIGLIB_PROFILE,
or a module main()'s default), then the default. Nothing is written to os.environ, so
import order no longer changes results: the setdefault blocks in coverage, bias,
fig_bearing and terrain_range become default_profile("calibrated") in their __main__.
- provenance.record stores the whole resolved set, defaults included, with the profile's
name and hash; configs/ counts toward the dirty-tree check.
- FIGLIB_REFINE_KM defaults to 0.01: solve re-finds its peak off the 0.4 km grid unless
set to 0. Every published number moves with the coordinated re-run that follows.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Measured before deciding, with refinement on. Core corpus: bearings move a median 0.00 deg and at most 0.30 (sm-n at Steele); seven fires move by 0.03 km or less; no count changes. Recent corpus (NOTES, 2026-09-25): it removes the bm-lens miss (bm-w at Bernardo +4.6 -> +1.6 deg, bm-e at Junction +4.6 -> +1.8), and Bernardo loses the 0.63 km that miss had bought it. The model is the one the stars measured, so it is the calibrated one. compare_geolocation gains the _full column; stars.fig_geolocation draws it as "after". Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…erywhere Every stage in rerun_results.sh, from d469059 (the recent block run twice; the first four entries were dirty from a docstring edit mid-run, and are kept). Confirmed-tier upper medians, grid -> refined: 1.90 -> 1.87 (published), 1.70 -> 1.68 (fisheye), 2.00 -> 2.02 (star azimuth), 2.02 with the whole solved camera. All 26: 2.58 -> 2.80. Rainbow 1.42 -> 2.17 and Palisades 1.82 -> 1.08 are refinement: their grid cells sat off flat ridge peaks. README, figures, likelihood-over-time and NOTES follow the new numbers. Co-Authored-By: Claude Opus 5.5 <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.
The one re-run that moves every README number, from a single commit (d469059), tagged
results-2026-09-26once this merges.What changed
settings.py. All 13FIGLIB_*switches are resolved in one place, in this order: environment, then profile (configs/*.tomlviaFIGLIB_PROFILE, or a modulemain()'s default), then default. Nothing writes intoos.environany more, so import order can't change results. The foursetdefaultblocks are gone;tests/test_settings.pycovers this.FIGLIB_REFINE_KM=0.01; set it to 0 for the bare grid node).FIGLIB_POSE_FULLis on in the calibrated profiles. It was measured first. On the core corpus it moves no fire by more than 0.03 km. On the recent corpus it removes the bm-lens miss of about 3°.rerun_results.shruns every stage that reaches the README, from a committed tree.Numbers that moved (confirmed-tier upper medians, center bearing)
solve.The full account is in NOTES.md, "The coordinated re-run" (2026-09-26).
Builds on #19 and #20, both merged (merge commits).
🤖 Generated with Claude Code