Set panel title/aspect/frameon once per panel, not per render command by timtreis · Pull Request #702 · scverse/spatialdata-plot

timtreis · 2026-06-07T03:49:29Z

Summary

Fixes#695. In PlotAccessor.show, the panel title / set_aspect("equal") / axis("off") block sat inside the for cmd, params in render_cmds: loop (basic.py:~1797), so it ran once per render command instead of once per panel.

For a chained call like render_images().pl.render_shapes(...), ax.set_title(...)/set_aspect/axis("off") executed N times per panel, and the title-length IndexError check re-evaluated each iteration — so a mismatched title length could surface mid-render rather than up front.

Fix

Dedent the block one level so it runs once per panel, after the per-command loop. That's the entire change — Python's indentation moves it out of the inner loop; it already uses only panel-level locals (title, panel_key, cs, i, ax, fig_params).

Verification

Byte-identical to main (RGBA buffers compared via np.array_equal) for a single-command plot and a 3-command chain with title= and frameon=False set — the calls are idempotent, so output is unchanged.
Non-visual suite: 368 passed, 1 skipped.
pre-commit: ruff + mypy pass.

No regression test added: the change removes redundant repeated calls with no observable output difference, so there's no new behavior to lock that the existing multi-panel-title visual tests (CI) don't already cover.

Note / possible follow-up

The issue also suggested validating len(title) == num_panels once before the panel loop. That's left out here to keep the diff minimal — the per-panel IndexError still fires (now once per panel). Happy to add the up-front validation if preferred.

…#695) In `show()`, the title/`set_aspect`/`axis("off")` block sat inside the `for cmd, params in render_cmds:` loop, so for a chained call (e.g. render_images().render_shapes()) these ran once per render command instead of once per panel, and the title-length `IndexError` check re-evaluated each iteration (so it could surface mid-loop rather than up front). Dedent the block one level so it runs once per panel after the per-command loop. Output is unchanged (the calls are idempotent) — verified byte-identical to main for single- and multi-command chains with title and frameon set.

codecov-commenter · 2026-06-14T17:07:46Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.39%. Comparing base (b370b1f) to head (03a3c1c).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #702 +/- ##
==========================================
+ Coverage 76.28% 76.39% +0.11% 
==========================================
Files 14 14 Lines 4327 4326 -1 Branches 1006 1007 +1 ==========================================
+ Hits 3301 3305 +4 + Misses 667 663 -4 + Partials 359 358 -1

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`80.82% <100.00%> (+1.40%)`	⬆️

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Builds on the per-panel title fix (#695): before the panel loop, validate that `title`, when a list, has length 1 or exactly num_panels. This: - surfaces the error before any drawing (not mid-render), - rejects an over-long title list (previously silently truncated), and - makes the per-panel IndexError dead code, so the try/except is removed. The error is now a ValueError (the right type for an invalid argument) raised once up front, instead of an IndexError raised per panel mid-render. No test pinned the old exception type. Adds non-visual regression tests for single- and multi-panel title-count validation. Pre-existing mypy error in utils.py (_resolve_measure_table returning Any) is unrelated and unchanged here; it is fixed separately in #714.

Integrate #702 (per-panel title fix) and #703 (as_points + fast extent) into the show() decomposition. Conflict resolution in src/spatialdata_plot/pl/basic.py: - #702 up-front title-count validation: kept (after num_panels). The adjacent axes/panel-count check was dropped here because the decomposition already relocated it into _plan_panels. - #702 simplified title selection (dropped the per-panel try/except): applied to the extracted _finalize_panel helper. - #703 had no show()-level render-dispatch changes (as_points is param-driven in render.py, handled inside _render_panel already); its only show()-level change, get_extent -> _get_extent_fast, auto-merged into the extent block and consumes _render_panel's `wants` dict. Also sweeps up two pre-existing #703 lint/type nits in utils.py surfaced by the merge: _fast_extent docstring (ruff D205) and _get_extent_fast Any-return (mypy). Verified: ruff + mypy clean; 109 non-visual show/shapes/labels tests pass.

timtreis · 2026-06-07T03:49:29Z

Summary

Fixes#695. In PlotAccessor.show, the panel title / set_aspect("equal") / axis("off") block sat inside the for cmd, params in render_cmds: loop (basic.py:~1797), so it ran once per render command instead of once per panel.

For a chained call like render_images().pl.render_shapes(...), ax.set_title(...)/set_aspect/axis("off") executed N times per panel, and the title-length IndexError check re-evaluated each iteration — so a mismatched title length could surface mid-render rather than up front.

Fix

Dedent the block one level so it runs once per panel, after the per-command loop. That's the entire change — Python's indentation moves it out of the inner loop; it already uses only panel-level locals (title, panel_key, cs, i, ax, fig_params).

Verification

Byte-identical to main (RGBA buffers compared via np.array_equal) for a single-command plot and a 3-command chain with title= and frameon=False set — the calls are idempotent, so output is unchanged.
Non-visual suite: 368 passed, 1 skipped.
pre-commit: ruff + mypy pass.

No regression test added: the change removes redundant repeated calls with no observable output difference, so there's no new behavior to lock that the existing multi-panel-title visual tests (CI) don't already cover.

Note / possible follow-up

The issue also suggested validating len(title) == num_panels once before the panel loop. That's left out here to keep the diff minimal — the per-panel IndexError still fires (now once per panel). Happy to add the up-front validation if preferred.

…#695) In `show()`, the title/`set_aspect`/`axis("off")` block sat inside the `for cmd, params in render_cmds:` loop, so for a chained call (e.g. render_images().render_shapes()) these ran once per render command instead of once per panel, and the title-length `IndexError` check re-evaluated each iteration (so it could surface mid-loop rather than up front). Dedent the block one level so it runs once per panel after the per-command loop. Output is unchanged (the calls are idempotent) — verified byte-identical to main for single- and multi-command chains with title and frameon set.

codecov-commenter · 2026-06-14T17:07:46Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.39%. Comparing base (b370b1f) to head (03a3c1c).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #702 +/- ##
==========================================
+ Coverage 76.28% 76.39% +0.11% 
==========================================
Files 14 14 Lines 4327 4326 -1 Branches 1006 1007 +1 ==========================================
+ Hits 3301 3305 +4 + Misses 667 663 -4 + Partials 359 358 -1

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`80.82% <100.00%> (+1.40%)`	⬆️

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Builds on the per-panel title fix (#695): before the panel loop, validate that `title`, when a list, has length 1 or exactly num_panels. This: - surfaces the error before any drawing (not mid-render), - rejects an over-long title list (previously silently truncated), and - makes the per-panel IndexError dead code, so the try/except is removed. The error is now a ValueError (the right type for an invalid argument) raised once up front, instead of an IndexError raised per panel mid-render. No test pinned the old exception type. Adds non-visual regression tests for single- and multi-panel title-count validation. Pre-existing mypy error in utils.py (_resolve_measure_table returning Any) is unrelated and unchanged here; it is fixed separately in #714.

Integrate #702 (per-panel title fix) and #703 (as_points + fast extent) into the show() decomposition. Conflict resolution in src/spatialdata_plot/pl/basic.py: - #702 up-front title-count validation: kept (after num_panels). The adjacent axes/panel-count check was dropped here because the decomposition already relocated it into _plan_panels. - #702 simplified title selection (dropped the per-panel try/except): applied to the extracted _finalize_panel helper. - #703 had no show()-level render-dispatch changes (as_points is param-driven in render.py, handled inside _render_panel already); its only show()-level change, get_extent -> _get_extent_fast, auto-merged into the extent block and consumes _render_panel's `wants` dict. Also sweeps up two pre-existing #703 lint/type nits in utils.py surfaced by the merge: _fast_extent docstring (ruff D205) and _get_extent_fast Any-return (mypy). Verified: ruff + mypy clean; 109 non-visual show/shapes/labels tests pass.

timtreis · 2026-06-07T03:49:29Z

Summary

Fixes#695. In PlotAccessor.show, the panel title / set_aspect("equal") / axis("off") block sat inside the for cmd, params in render_cmds: loop (basic.py:~1797), so it ran once per render command instead of once per panel.

For a chained call like render_images().pl.render_shapes(...), ax.set_title(...)/set_aspect/axis("off") executed N times per panel, and the title-length IndexError check re-evaluated each iteration — so a mismatched title length could surface mid-render rather than up front.

Fix

Dedent the block one level so it runs once per panel, after the per-command loop. That's the entire change — Python's indentation moves it out of the inner loop; it already uses only panel-level locals (title, panel_key, cs, i, ax, fig_params).

Verification

Byte-identical to main (RGBA buffers compared via np.array_equal) for a single-command plot and a 3-command chain with title= and frameon=False set — the calls are idempotent, so output is unchanged.
Non-visual suite: 368 passed, 1 skipped.
pre-commit: ruff + mypy pass.

No regression test added: the change removes redundant repeated calls with no observable output difference, so there's no new behavior to lock that the existing multi-panel-title visual tests (CI) don't already cover.

Note / possible follow-up

The issue also suggested validating len(title) == num_panels once before the panel loop. That's left out here to keep the diff minimal — the per-panel IndexError still fires (now once per panel). Happy to add the up-front validation if preferred.

…#695) In `show()`, the title/`set_aspect`/`axis("off")` block sat inside the `for cmd, params in render_cmds:` loop, so for a chained call (e.g. render_images().render_shapes()) these ran once per render command instead of once per panel, and the title-length `IndexError` check re-evaluated each iteration (so it could surface mid-loop rather than up front). Dedent the block one level so it runs once per panel after the per-command loop. Output is unchanged (the calls are idempotent) — verified byte-identical to main for single- and multi-command chains with title and frameon set.

codecov-commenter · 2026-06-14T17:07:46Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.39%. Comparing base (b370b1f) to head (03a3c1c).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #702 +/- ##
==========================================
+ Coverage 76.28% 76.39% +0.11% 
==========================================
Files 14 14 Lines 4327 4326 -1 Branches 1006 1007 +1 ==========================================
+ Hits 3301 3305 +4 + Misses 667 663 -4 + Partials 359 358 -1

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`80.82% <100.00%> (+1.40%)`	⬆️

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Builds on the per-panel title fix (#695): before the panel loop, validate that `title`, when a list, has length 1 or exactly num_panels. This: - surfaces the error before any drawing (not mid-render), - rejects an over-long title list (previously silently truncated), and - makes the per-panel IndexError dead code, so the try/except is removed. The error is now a ValueError (the right type for an invalid argument) raised once up front, instead of an IndexError raised per panel mid-render. No test pinned the old exception type. Adds non-visual regression tests for single- and multi-panel title-count validation. Pre-existing mypy error in utils.py (_resolve_measure_table returning Any) is unrelated and unchanged here; it is fixed separately in #714.

Integrate #702 (per-panel title fix) and #703 (as_points + fast extent) into the show() decomposition. Conflict resolution in src/spatialdata_plot/pl/basic.py: - #702 up-front title-count validation: kept (after num_panels). The adjacent axes/panel-count check was dropped here because the decomposition already relocated it into _plan_panels. - #702 simplified title selection (dropped the per-panel try/except): applied to the extracted _finalize_panel helper. - #703 had no show()-level render-dispatch changes (as_points is param-driven in render.py, handled inside _render_panel already); its only show()-level change, get_extent -> _get_extent_fast, auto-merged into the extent block and consumes _render_panel's `wants` dict. Also sweeps up two pre-existing #703 lint/type nits in utils.py surfaced by the merge: _fast_extent docstring (ruff D205) and _get_extent_fast Any-return (mypy). Verified: ruff + mypy clean; 109 non-visual show/shapes/labels tests pass.

timtreis · 2026-06-07T03:49:29Z

Summary

Fixes#695. In PlotAccessor.show, the panel title / set_aspect("equal") / axis("off") block sat inside the for cmd, params in render_cmds: loop (basic.py:~1797), so it ran once per render command instead of once per panel.

For a chained call like render_images().pl.render_shapes(...), ax.set_title(...)/set_aspect/axis("off") executed N times per panel, and the title-length IndexError check re-evaluated each iteration — so a mismatched title length could surface mid-render rather than up front.

Fix

Dedent the block one level so it runs once per panel, after the per-command loop. That's the entire change — Python's indentation moves it out of the inner loop; it already uses only panel-level locals (title, panel_key, cs, i, ax, fig_params).

Verification

Byte-identical to main (RGBA buffers compared via np.array_equal) for a single-command plot and a 3-command chain with title= and frameon=False set — the calls are idempotent, so output is unchanged.
Non-visual suite: 368 passed, 1 skipped.
pre-commit: ruff + mypy pass.

No regression test added: the change removes redundant repeated calls with no observable output difference, so there's no new behavior to lock that the existing multi-panel-title visual tests (CI) don't already cover.

Note / possible follow-up

The issue also suggested validating len(title) == num_panels once before the panel loop. That's left out here to keep the diff minimal — the per-panel IndexError still fires (now once per panel). Happy to add the up-front validation if preferred.

…#695) In `show()`, the title/`set_aspect`/`axis("off")` block sat inside the `for cmd, params in render_cmds:` loop, so for a chained call (e.g. render_images().render_shapes()) these ran once per render command instead of once per panel, and the title-length `IndexError` check re-evaluated each iteration (so it could surface mid-loop rather than up front). Dedent the block one level so it runs once per panel after the per-command loop. Output is unchanged (the calls are idempotent) — verified byte-identical to main for single- and multi-command chains with title and frameon set.

codecov-commenter · 2026-06-14T17:07:46Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.39%. Comparing base (b370b1f) to head (03a3c1c).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #702 +/- ##
==========================================
+ Coverage 76.28% 76.39% +0.11% 
==========================================
Files 14 14 Lines 4327 4326 -1 Branches 1006 1007 +1 ==========================================
+ Hits 3301 3305 +4 + Misses 667 663 -4 + Partials 359 358 -1

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`80.82% <100.00%> (+1.40%)`	⬆️

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Builds on the per-panel title fix (#695): before the panel loop, validate that `title`, when a list, has length 1 or exactly num_panels. This: - surfaces the error before any drawing (not mid-render), - rejects an over-long title list (previously silently truncated), and - makes the per-panel IndexError dead code, so the try/except is removed. The error is now a ValueError (the right type for an invalid argument) raised once up front, instead of an IndexError raised per panel mid-render. No test pinned the old exception type. Adds non-visual regression tests for single- and multi-panel title-count validation. Pre-existing mypy error in utils.py (_resolve_measure_table returning Any) is unrelated and unchanged here; it is fixed separately in #714.

Integrate #702 (per-panel title fix) and #703 (as_points + fast extent) into the show() decomposition. Conflict resolution in src/spatialdata_plot/pl/basic.py: - #702 up-front title-count validation: kept (after num_panels). The adjacent axes/panel-count check was dropped here because the decomposition already relocated it into _plan_panels. - #702 simplified title selection (dropped the per-panel try/except): applied to the extracted _finalize_panel helper. - #703 had no show()-level render-dispatch changes (as_points is param-driven in render.py, handled inside _render_panel already); its only show()-level change, get_extent -> _get_extent_fast, auto-merged into the extent block and consumes _render_panel's `wants` dict. Also sweeps up two pre-existing #703 lint/type nits in utils.py surfaced by the merge: _fast_extent docstring (ruff D205) and _get_extent_fast Any-return (mypy). Verified: ruff + mypy clean; 109 non-visual show/shapes/labels tests pass.

timtreis · 2026-06-07T03:49:29Z

Summary

Fixes#695. In PlotAccessor.show, the panel title / set_aspect("equal") / axis("off") block sat inside the for cmd, params in render_cmds: loop (basic.py:~1797), so it ran once per render command instead of once per panel.

For a chained call like render_images().pl.render_shapes(...), ax.set_title(...)/set_aspect/axis("off") executed N times per panel, and the title-length IndexError check re-evaluated each iteration — so a mismatched title length could surface mid-render rather than up front.

Fix

Dedent the block one level so it runs once per panel, after the per-command loop. That's the entire change — Python's indentation moves it out of the inner loop; it already uses only panel-level locals (title, panel_key, cs, i, ax, fig_params).

Verification

Byte-identical to main (RGBA buffers compared via np.array_equal) for a single-command plot and a 3-command chain with title= and frameon=False set — the calls are idempotent, so output is unchanged.
Non-visual suite: 368 passed, 1 skipped.
pre-commit: ruff + mypy pass.

No regression test added: the change removes redundant repeated calls with no observable output difference, so there's no new behavior to lock that the existing multi-panel-title visual tests (CI) don't already cover.

Note / possible follow-up

The issue also suggested validating len(title) == num_panels once before the panel loop. That's left out here to keep the diff minimal — the per-panel IndexError still fires (now once per panel). Happy to add the up-front validation if preferred.

…#695) In `show()`, the title/`set_aspect`/`axis("off")` block sat inside the `for cmd, params in render_cmds:` loop, so for a chained call (e.g. render_images().render_shapes()) these ran once per render command instead of once per panel, and the title-length `IndexError` check re-evaluated each iteration (so it could surface mid-loop rather than up front). Dedent the block one level so it runs once per panel after the per-command loop. Output is unchanged (the calls are idempotent) — verified byte-identical to main for single- and multi-command chains with title and frameon set.

codecov-commenter · 2026-06-14T17:07:46Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.39%. Comparing base (b370b1f) to head (03a3c1c).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #702 +/- ##
==========================================
+ Coverage 76.28% 76.39% +0.11% 
==========================================
Files 14 14 Lines 4327 4326 -1 Branches 1006 1007 +1 ==========================================
+ Hits 3301 3305 +4 + Misses 667 663 -4 + Partials 359 358 -1

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`80.82% <100.00%> (+1.40%)`	⬆️

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Builds on the per-panel title fix (#695): before the panel loop, validate that `title`, when a list, has length 1 or exactly num_panels. This: - surfaces the error before any drawing (not mid-render), - rejects an over-long title list (previously silently truncated), and - makes the per-panel IndexError dead code, so the try/except is removed. The error is now a ValueError (the right type for an invalid argument) raised once up front, instead of an IndexError raised per panel mid-render. No test pinned the old exception type. Adds non-visual regression tests for single- and multi-panel title-count validation. Pre-existing mypy error in utils.py (_resolve_measure_table returning Any) is unrelated and unchanged here; it is fixed separately in #714.

Integrate #702 (per-panel title fix) and #703 (as_points + fast extent) into the show() decomposition. Conflict resolution in src/spatialdata_plot/pl/basic.py: - #702 up-front title-count validation: kept (after num_panels). The adjacent axes/panel-count check was dropped here because the decomposition already relocated it into _plan_panels. - #702 simplified title selection (dropped the per-panel try/except): applied to the extracted _finalize_panel helper. - #703 had no show()-level render-dispatch changes (as_points is param-driven in render.py, handled inside _render_panel already); its only show()-level change, get_extent -> _get_extent_fast, auto-merged into the extent block and consumes _render_panel's `wants` dict. Also sweeps up two pre-existing #703 lint/type nits in utils.py surfaced by the merge: _fast_extent docstring (ruff D205) and _get_extent_fast Any-return (mypy). Verified: ruff + mypy clean; 109 non-visual show/shapes/labels tests pass.

timtreis · 2026-06-07T03:49:29Z

Summary

Fixes#695. In PlotAccessor.show, the panel title / set_aspect("equal") / axis("off") block sat inside the for cmd, params in render_cmds: loop (basic.py:~1797), so it ran once per render command instead of once per panel.

For a chained call like render_images().pl.render_shapes(...), ax.set_title(...)/set_aspect/axis("off") executed N times per panel, and the title-length IndexError check re-evaluated each iteration — so a mismatched title length could surface mid-render rather than up front.

Fix

Dedent the block one level so it runs once per panel, after the per-command loop. That's the entire change — Python's indentation moves it out of the inner loop; it already uses only panel-level locals (title, panel_key, cs, i, ax, fig_params).

Verification

Byte-identical to main (RGBA buffers compared via np.array_equal) for a single-command plot and a 3-command chain with title= and frameon=False set — the calls are idempotent, so output is unchanged.
Non-visual suite: 368 passed, 1 skipped.
pre-commit: ruff + mypy pass.

No regression test added: the change removes redundant repeated calls with no observable output difference, so there's no new behavior to lock that the existing multi-panel-title visual tests (CI) don't already cover.

Note / possible follow-up

The issue also suggested validating len(title) == num_panels once before the panel loop. That's left out here to keep the diff minimal — the per-panel IndexError still fires (now once per panel). Happy to add the up-front validation if preferred.

…#695) In `show()`, the title/`set_aspect`/`axis("off")` block sat inside the `for cmd, params in render_cmds:` loop, so for a chained call (e.g. render_images().render_shapes()) these ran once per render command instead of once per panel, and the title-length `IndexError` check re-evaluated each iteration (so it could surface mid-loop rather than up front). Dedent the block one level so it runs once per panel after the per-command loop. Output is unchanged (the calls are idempotent) — verified byte-identical to main for single- and multi-command chains with title and frameon set.

codecov-commenter · 2026-06-14T17:07:46Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.39%. Comparing base (b370b1f) to head (03a3c1c).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #702 +/- ##
==========================================
+ Coverage 76.28% 76.39% +0.11% 
==========================================
Files 14 14 Lines 4327 4326 -1 Branches 1006 1007 +1 ==========================================
+ Hits 3301 3305 +4 + Misses 667 663 -4 + Partials 359 358 -1

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`80.82% <100.00%> (+1.40%)`	⬆️

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Builds on the per-panel title fix (#695): before the panel loop, validate that `title`, when a list, has length 1 or exactly num_panels. This: - surfaces the error before any drawing (not mid-render), - rejects an over-long title list (previously silently truncated), and - makes the per-panel IndexError dead code, so the try/except is removed. The error is now a ValueError (the right type for an invalid argument) raised once up front, instead of an IndexError raised per panel mid-render. No test pinned the old exception type. Adds non-visual regression tests for single- and multi-panel title-count validation. Pre-existing mypy error in utils.py (_resolve_measure_table returning Any) is unrelated and unchanged here; it is fixed separately in #714.

Integrate #702 (per-panel title fix) and #703 (as_points + fast extent) into the show() decomposition. Conflict resolution in src/spatialdata_plot/pl/basic.py: - #702 up-front title-count validation: kept (after num_panels). The adjacent axes/panel-count check was dropped here because the decomposition already relocated it into _plan_panels. - #702 simplified title selection (dropped the per-panel try/except): applied to the extracted _finalize_panel helper. - #703 had no show()-level render-dispatch changes (as_points is param-driven in render.py, handled inside _render_panel already); its only show()-level change, get_extent -> _get_extent_fast, auto-merged into the extent block and consumes _render_panel's `wants` dict. Also sweeps up two pre-existing #703 lint/type nits in utils.py surfaced by the merge: _fast_extent docstring (ruff D205) and _get_extent_fast Any-return (mypy). Verified: ruff + mypy clean; 109 non-visual show/shapes/labels tests pass.

timtreis · 2026-06-07T03:49:29Z

Summary

Fixes#695. In PlotAccessor.show, the panel title / set_aspect("equal") / axis("off") block sat inside the for cmd, params in render_cmds: loop (basic.py:~1797), so it ran once per render command instead of once per panel.

For a chained call like render_images().pl.render_shapes(...), ax.set_title(...)/set_aspect/axis("off") executed N times per panel, and the title-length IndexError check re-evaluated each iteration — so a mismatched title length could surface mid-render rather than up front.

Fix

Dedent the block one level so it runs once per panel, after the per-command loop. That's the entire change — Python's indentation moves it out of the inner loop; it already uses only panel-level locals (title, panel_key, cs, i, ax, fig_params).

Verification

Byte-identical to main (RGBA buffers compared via np.array_equal) for a single-command plot and a 3-command chain with title= and frameon=False set — the calls are idempotent, so output is unchanged.
Non-visual suite: 368 passed, 1 skipped.
pre-commit: ruff + mypy pass.

No regression test added: the change removes redundant repeated calls with no observable output difference, so there's no new behavior to lock that the existing multi-panel-title visual tests (CI) don't already cover.

Note / possible follow-up

The issue also suggested validating len(title) == num_panels once before the panel loop. That's left out here to keep the diff minimal — the per-panel IndexError still fires (now once per panel). Happy to add the up-front validation if preferred.

…#695) In `show()`, the title/`set_aspect`/`axis("off")` block sat inside the `for cmd, params in render_cmds:` loop, so for a chained call (e.g. render_images().render_shapes()) these ran once per render command instead of once per panel, and the title-length `IndexError` check re-evaluated each iteration (so it could surface mid-loop rather than up front). Dedent the block one level so it runs once per panel after the per-command loop. Output is unchanged (the calls are idempotent) — verified byte-identical to main for single- and multi-command chains with title and frameon set.

codecov-commenter · 2026-06-14T17:07:46Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.39%. Comparing base (b370b1f) to head (03a3c1c).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #702 +/- ##
==========================================
+ Coverage 76.28% 76.39% +0.11% 
==========================================
Files 14 14 Lines 4327 4326 -1 Branches 1006 1007 +1 ==========================================
+ Hits 3301 3305 +4 + Misses 667 663 -4 + Partials 359 358 -1

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`80.82% <100.00%> (+1.40%)`	⬆️

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Builds on the per-panel title fix (#695): before the panel loop, validate that `title`, when a list, has length 1 or exactly num_panels. This: - surfaces the error before any drawing (not mid-render), - rejects an over-long title list (previously silently truncated), and - makes the per-panel IndexError dead code, so the try/except is removed. The error is now a ValueError (the right type for an invalid argument) raised once up front, instead of an IndexError raised per panel mid-render. No test pinned the old exception type. Adds non-visual regression tests for single- and multi-panel title-count validation. Pre-existing mypy error in utils.py (_resolve_measure_table returning Any) is unrelated and unchanged here; it is fixed separately in #714.

Integrate #702 (per-panel title fix) and #703 (as_points + fast extent) into the show() decomposition. Conflict resolution in src/spatialdata_plot/pl/basic.py: - #702 up-front title-count validation: kept (after num_panels). The adjacent axes/panel-count check was dropped here because the decomposition already relocated it into _plan_panels. - #702 simplified title selection (dropped the per-panel try/except): applied to the extracted _finalize_panel helper. - #703 had no show()-level render-dispatch changes (as_points is param-driven in render.py, handled inside _render_panel already); its only show()-level change, get_extent -> _get_extent_fast, auto-merged into the extent block and consumes _render_panel's `wants` dict. Also sweeps up two pre-existing #703 lint/type nits in utils.py surfaced by the merge: _fast_extent docstring (ruff D205) and _get_extent_fast Any-return (mypy). Verified: ruff + mypy clean; 109 non-visual show/shapes/labels tests pass.

timtreis · 2026-06-07T03:49:29Z

Summary

Fixes#695. In PlotAccessor.show, the panel title / set_aspect("equal") / axis("off") block sat inside the for cmd, params in render_cmds: loop (basic.py:~1797), so it ran once per render command instead of once per panel.

For a chained call like render_images().pl.render_shapes(...), ax.set_title(...)/set_aspect/axis("off") executed N times per panel, and the title-length IndexError check re-evaluated each iteration — so a mismatched title length could surface mid-render rather than up front.

Fix

Dedent the block one level so it runs once per panel, after the per-command loop. That's the entire change — Python's indentation moves it out of the inner loop; it already uses only panel-level locals (title, panel_key, cs, i, ax, fig_params).

Verification

Byte-identical to main (RGBA buffers compared via np.array_equal) for a single-command plot and a 3-command chain with title= and frameon=False set — the calls are idempotent, so output is unchanged.
Non-visual suite: 368 passed, 1 skipped.
pre-commit: ruff + mypy pass.

No regression test added: the change removes redundant repeated calls with no observable output difference, so there's no new behavior to lock that the existing multi-panel-title visual tests (CI) don't already cover.

Note / possible follow-up

The issue also suggested validating len(title) == num_panels once before the panel loop. That's left out here to keep the diff minimal — the per-panel IndexError still fires (now once per panel). Happy to add the up-front validation if preferred.

…#695) In `show()`, the title/`set_aspect`/`axis("off")` block sat inside the `for cmd, params in render_cmds:` loop, so for a chained call (e.g. render_images().render_shapes()) these ran once per render command instead of once per panel, and the title-length `IndexError` check re-evaluated each iteration (so it could surface mid-loop rather than up front). Dedent the block one level so it runs once per panel after the per-command loop. Output is unchanged (the calls are idempotent) — verified byte-identical to main for single- and multi-command chains with title and frameon set.

codecov-commenter · 2026-06-14T17:07:46Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.39%. Comparing base (b370b1f) to head (03a3c1c).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #702 +/- ##
==========================================
+ Coverage 76.28% 76.39% +0.11% 
==========================================
Files 14 14 Lines 4327 4326 -1 Branches 1006 1007 +1 ==========================================
+ Hits 3301 3305 +4 + Misses 667 663 -4 + Partials 359 358 -1

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`80.82% <100.00%> (+1.40%)`	⬆️

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Builds on the per-panel title fix (#695): before the panel loop, validate that `title`, when a list, has length 1 or exactly num_panels. This: - surfaces the error before any drawing (not mid-render), - rejects an over-long title list (previously silently truncated), and - makes the per-panel IndexError dead code, so the try/except is removed. The error is now a ValueError (the right type for an invalid argument) raised once up front, instead of an IndexError raised per panel mid-render. No test pinned the old exception type. Adds non-visual regression tests for single- and multi-panel title-count validation. Pre-existing mypy error in utils.py (_resolve_measure_table returning Any) is unrelated and unchanged here; it is fixed separately in #714.

Integrate #702 (per-panel title fix) and #703 (as_points + fast extent) into the show() decomposition. Conflict resolution in src/spatialdata_plot/pl/basic.py: - #702 up-front title-count validation: kept (after num_panels). The adjacent axes/panel-count check was dropped here because the decomposition already relocated it into _plan_panels. - #702 simplified title selection (dropped the per-panel try/except): applied to the extracted _finalize_panel helper. - #703 had no show()-level render-dispatch changes (as_points is param-driven in render.py, handled inside _render_panel already); its only show()-level change, get_extent -> _get_extent_fast, auto-merged into the extent block and consumes _render_panel's `wants` dict. Also sweeps up two pre-existing #703 lint/type nits in utils.py surfaced by the merge: _fast_extent docstring (ruff D205) and _get_extent_fast Any-return (mypy). Verified: ruff + mypy clean; 109 non-visual show/shapes/labels tests pass.

timtreis force-pushed the fix/issue-695-title-per-panel branch 2 times, most recently from b05fb5b to 6bce242Compare June 14, 2026 16:59

timtreis force-pushed the fix/issue-695-title-per-panel branch from 6bce242 to 03a3c1cCompare June 14, 2026 17:10

timtreis merged commit cf0073a into mainJun 14, 2026
7 of 8 checks passed

timtreis deleted the fix/issue-695-title-per-panel branch June 14, 2026 17:17

timtreis added the release-changed label Aug 10, 2026

timtreis force-pushed the fix/issue-695-title-per-panel branch 2 times, most recently from b05fb5b to 6bce242Compare June 14, 2026 16:59

timtreis force-pushed the fix/issue-695-title-per-panel branch from 6bce242 to 03a3c1cCompare June 14, 2026 17:10

timtreis merged commit cf0073a into mainJun 14, 2026
7 of 8 checks passed

timtreis deleted the fix/issue-695-title-per-panel branch June 14, 2026 17:17

timtreis added the release-changed label Aug 10, 2026

timtreis force-pushed the fix/issue-695-title-per-panel branch 2 times, most recently from b05fb5b to 6bce242Compare June 14, 2026 16:59

timtreis force-pushed the fix/issue-695-title-per-panel branch from 6bce242 to 03a3c1cCompare June 14, 2026 17:10

timtreis merged commit cf0073a into mainJun 14, 2026
7 of 8 checks passed

timtreis deleted the fix/issue-695-title-per-panel branch June 14, 2026 17:17

timtreis added the release-changed label Aug 10, 2026

timtreis force-pushed the fix/issue-695-title-per-panel branch 2 times, most recently from b05fb5b to 6bce242Compare June 14, 2026 16:59

timtreis force-pushed the fix/issue-695-title-per-panel branch from 6bce242 to 03a3c1cCompare June 14, 2026 17:10

timtreis merged commit cf0073a into mainJun 14, 2026
7 of 8 checks passed

timtreis deleted the fix/issue-695-title-per-panel branch June 14, 2026 17:17

timtreis added the release-changed label Aug 10, 2026

timtreis force-pushed the fix/issue-695-title-per-panel branch 2 times, most recently from b05fb5b to 6bce242Compare June 14, 2026 16:59

timtreis force-pushed the fix/issue-695-title-per-panel branch from 6bce242 to 03a3c1cCompare June 14, 2026 17:10

timtreis merged commit cf0073a into mainJun 14, 2026
7 of 8 checks passed

timtreis deleted the fix/issue-695-title-per-panel branch June 14, 2026 17:17

timtreis added the release-changed label Aug 10, 2026

timtreis force-pushed the fix/issue-695-title-per-panel branch 2 times, most recently from b05fb5b to 6bce242Compare June 14, 2026 16:59

timtreis force-pushed the fix/issue-695-title-per-panel branch from 6bce242 to 03a3c1cCompare June 14, 2026 17:10

timtreis merged commit cf0073a into mainJun 14, 2026
7 of 8 checks passed

timtreis deleted the fix/issue-695-title-per-panel branch June 14, 2026 17:17

timtreis added the release-changed label Aug 10, 2026

timtreis force-pushed the fix/issue-695-title-per-panel branch 2 times, most recently from b05fb5b to 6bce242Compare June 14, 2026 16:59

timtreis force-pushed the fix/issue-695-title-per-panel branch from 6bce242 to 03a3c1cCompare June 14, 2026 17:10

timtreis merged commit cf0073a into mainJun 14, 2026
7 of 8 checks passed

timtreis deleted the fix/issue-695-title-per-panel branch June 14, 2026 17:17

timtreis added the release-changed label Aug 10, 2026

timtreis force-pushed the fix/issue-695-title-per-panel branch 2 times, most recently from b05fb5b to 6bce242Compare June 14, 2026 16:59

timtreis force-pushed the fix/issue-695-title-per-panel branch from 6bce242 to 03a3c1cCompare June 14, 2026 17:10

timtreis merged commit cf0073a into mainJun 14, 2026
7 of 8 checks passed

timtreis deleted the fix/issue-695-title-per-panel branch June 14, 2026 17:17

timtreis added the release-changed label Aug 10, 2026

Uh oh!

Conversation

timtreis commented Jun 7, 2026

Summary

Fix

Verification

Note / possible follow-up

Uh oh!

codecov-commenter commented Jun 14, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

timtreis commented Jun 7, 2026

Summary

Fix

Verification

Note / possible follow-up

Uh oh!

codecov-commenter commented Jun 14, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

timtreis commented Jun 7, 2026

Summary

Fix

Verification

Note / possible follow-up

Uh oh!

codecov-commenter commented Jun 14, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

timtreis commented Jun 7, 2026

Summary

Fix

Verification

Note / possible follow-up

Uh oh!

codecov-commenter commented Jun 14, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

codecov-commenter commented Jun 14, 2026•
edited
Loading

codecov-commenter commented Jun 14, 2026•
edited
Loading

codecov-commenter commented Jun 14, 2026•
edited
Loading

codecov-commenter commented Jun 14, 2026•
edited
Loading

codecov-commenter commented Jun 14, 2026•
edited
Loading

codecov-commenter commented Jun 14, 2026•
edited
Loading