[trissim]

Give image pixels the space, not the padding

Two requested layout repairs in one change:

• Figure3C grows upward from18% to23% of canvas height (+27.8%), consuming5 percentage points of the A/B-to-C gap. Its statistics, lower labels and page footprint stay fixed.
• Figure6G becomes a compact vertical raw/first/final comparison beside H. H's unchanged four matched crops grow from11.8% to20.9% of canvas height (about77% larger in each dimension,3.1× image area). G's native image size stays approximately unchanged, with proper13pt figure-owned column labels instead of scaled embedded headings.

Owners and before/after

Class structure is unchanged: BenchmarkFigure(FigureSheet) and AnalysisResultsFigure(VolumeResultsFigure), with VolumeResultsFigure(SupplementFigure) retaining responsibility for the retinal/body layout. FigureSheet still validates retained artwork hashes, preserves aspect and records editorial crops. No new renderer, helper, registry, data source or compatibility system.

Before: G was a reduced wide composite with large side gutters; H's tall crops were centred in short wide boxes. After: the same first retinal row is separated into its three complete retained image tiles (including20µm bars), stacked at left; H uses the right-hand strip at the native crop aspect. The four H crop coordinates and10-voxel bars are unchanged. E/F move slightly upward to make room; E's F1 annotation moves to its heading and uses compact decimal formatting, without changing either score.

No image selection, contrast, labels, scale calibration, quantitative result, measurement/timing source, text/caption or scientific interpretation changes. The old placements are replaced in place; the historical source artwork remains immutable scientific evidence.

Verification

• Existing submission_benchmark_summary() and submission_summary() entrypoints, existing papers-only environment only; no scientific/model/benchmark reruns.
• Inspected full generated figures; corrected heading collisions found in the first preview before acceptance.
• Source check passed: AST imports/declarations/signatures/bases unchanged in all3generators; retained timing projection CSV, both source artworks and analysis_results.csv byte-identical to main.
• Real candidate build passed:834observed inputs,343matching artifacts,153exit0commands. Main29pages, supplement22, reader50.
• Personally inspected actual manuscript pages17(Figure3) and21(Figure6) at150dpi. C's gap is reduced; G's three views and H's four larger views retain their bars and fit without heading/caption collisions. Combined-reader pages independently render pixel-identically to those manuscript pages.
• Accepted run:run-20261030T233628-2724705c; exact reader SHA256:50e6e776fdbc4358abe073062e54a1372bdb0b225575eb1786a6e711c1c7f872. No rebuild is needed for publication.
• Source PR scope:11files,+1178/−1091, primarily regenerated SVGs. Existing class structure is retained. Evidence:/home/ts/.cache/agent-scratch/openhcs-paper/spacing-gh-{source-proof.json,inspect.log,visual-acceptance.md} and spacing-gh-pages/package.json.

Exact commands and results

Using the existing /home/ts/code/projects/papers-only/.venv/bin/python, without creating or syncing an environment:

PYTHONPATH=/home/ts/code/projects/papers-only:.:paper/figures python -c 'from build_slas_benchmark_reference import submission_benchmark_summary; from build_slas_task_only import submission_summary; submission_benchmark_summary(); submission_summary()'
PYTHONPATH=/home/ts/code/projects/papers-only python paper/build_paper.py build --candidate
PYTHONPATH=/home/ts/code/projects/papers-only python spacing_inspect.py

All exited0. The one-shot inspection verified all343manifest artifacts/153commands, located Figures3/6, rendered manuscript/reader pages17/21 at150dpi, and proved the corresponding page images identical before personal visual inspection. Its source was removed after that one run. No full test suite, local NRA gate or benchmark execution.

Merged, exact-byte published, checkout retired

Merged as ce73ee93eb9d60fe6215723547654dc9abeb2940; final-head hosted structural-ratchet and nra-r1 both succeeded. Independently matched all834 merged-main inputs and tool identities, then published the343 exact inspected artifacts without rebuilding. Reader remains file:///home/ts/code/projects/openhcs/paper/current/openhcs_review.pdf, state current, changed []; its SHA256 equals the accepted value above.

Preserved470 user untracked files/modes and all343 prior-release artifacts. Both one-shot helper sources are removed; the clean goal checkout and duplicate candidate were retired through supported goal-checkout remove after remote and actual process/mount borrower checks. Canonical accepted build, original inspected manifest and visual acceptance remain in paper/build/run-20261030T233628-2724705c. Concise publication/retirement evidence is retained beside spacing-gh-publication.json and spacing-gh-retirement.json in the stated scratch directory. Canonical git diff --stat origin/main...HEAD is empty. Available disk is back to approximately11GiB.

Assumed: G continues to show the first retained retinal region, and H the exact existing Z36 body views; no new field is selected.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota. Try requesting a review at a lower effort level.

An organization owner can choose the repository organization as the billing source for reviews requested by members with a paid Copilot license. Learn about Copilot code review billing.