Repository navigation
House style for every figure, and a design-review checklist - #154
Open
syednafizreja wants to merge 1 commit into
Open
syednafizreja wants to merge 1 commit into
syednafizreja wants to merge 1 commit into
Conversation
One style for all ShellBench figures (src/shellbench/style.py, shellbench.mplstyle): Elsevier column widths and 7 pt lettering, Arial, validated colour order, vector PDF/SVG plus 600 dpi PNG without date stamps, and an automated review. figures.py redraws every figure from results/ (python -m shellbench.figures; --paper for journal copies). docs/visual_standard.md holds the rules and the checklist. No analysis or number changed. src/shellbench/ml/ (frozen at 5c11ce8) is untouched; scripts/surrogate_figures.py calls it and adds results/surrogate_predictions.csv for the parity figure.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Requested by Researcher007 · project thread
What
The author's rule of 9 Nov 2026: every ShellBench output has to look and work like careful, expert work. This PR sets one house style and redraws every figure in it. No analysis or result changed.
src/shellbench/style.pyandshellbench.mplstyle: Elsevier column widths (90/140/190 mm) and 7 pt lettering, Arial (Liberation Sans on Linux), the validated dataviz colour order, vector PDF and SVG plus a 600 dpi PNG with no date stamps, ownership footer, andreview(), an automated design check (width, text size, overflow, axis labels, no title in paper copies).src/shellbench/figures.py: one builder per figure, fed only fromresults/.python -m shellbench.figuresredraws all seven;--paperwrites footer-free copies tofigures/paper/.docs/visual_standard.md: the standard, with sources, and the checklist every figure, CLI output, README and social card must pass.tests/test_style.py: 23 tests.figures.py.Surrogate freeze
src/shellbench/ml/is untouched (frozen at 5c11ce8).scripts/surrogate_figures.pycalls the frozen functions in the order ofconformal.main, writes the same CSVs, and addsresults/surrogate_predictions.csvso the parity figure can be redrawn without re-training. Re-run in the cloud:surrogate_accuracy.csvcame out byte-identical. For the 11 Nov DOI run, use this script in place ofpython -m shellbench.ml.conformal.Checks
ccxin the cloud).results/unchanged apart from the new predictions file.surrogate_accuracy.csv; the waterfall's saving equalsS_binanchor_slab_vs_shell.csv.Generated by Claude Code