test: force headless Matplotlib backend for local + CI test runs

[petercorke]

Problem

Running pytest locally pops up real Matplotlib windows, and a couple of
animation tests hang outright. CI already sets MPLBACKEND=Agg in the
workflow env, but that never applied to local runs — several modules
(geom2d, geom3d, spline, animate) import matplotlib.pyplot at
module load time, so whichever backend is active at that point sticks for
the whole session, and locally that's the platform's interactive backend.

Fix

1. tests/conftest.py (new): os.environ.setdefault("MPLBACKEND", "Agg")
   before anything else runs. conftest.py is guaranteed to load before
   pytest imports any test module, so this closes the gap for local runs
   without touching the CI workflow (redundant there now, not conflicting).
   setdefault(), not a hard override, so MPLBACKEND=MacOSX pytest ...
   still shows a real plot when you actually want to eyeball one.

2. Forcing Agg everywhere surfaced a real, previously-masked hang:
   tranimate(..., wait=True) busy-loops on plt.pause() waiting for
   FuncAnimation's timer callback to deregister itself — driven by the
   GUI event loop, which Agg doesn't have. Confirmed with a hard-alarm
   timeout: the affected test ate a full 12s budget with zero output vs.
   completing normally under a real backend. These tests were already
   skipped in CI via os.environ.get("CI") == "true", which happened to
   also mask this locally-only-when-Agg-is-forced hang — but only in CI;
   locally (pre-Agg) they ran fine because a real backend genuinely pumps
   the event loop.

   Broadened the three affected skip conditions (test_animate in
   test_transforms3d_plot.py; the two animate=True test_visualize
   tests and the inline guard in test_spline.py) to additionally skip
   when plt.get_backend() == "agg" — the actual root cause, rather than
   the CI-env-var proxy for it. Purely additive, no existing condition
   removed, so CI behaviour is unchanged. Everything else (plain,
   non-animated plotting tests) now safely runs headless both locally and
   in CI — strictly more coverage than before.

Testing

• Full suite, CI unset (simulating a local run): 339 passed, 3 skipped,
  1.6s, no windows, no hangs.
• Full suite, CI=true (simulating CI): 338 passed, 4 skipped — matches
  prior CI behaviour.
• black --check clean at the pinned 23.10.0.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[petercorke]

Merge-order suggestion across the currently open PRs, based on a full file-overlap check across all of them. Nothing here is a conflict with anything already merged — these are parallel PRs sharing some source files with each other, so a rebase will likely be needed wherever paths cross, regardless of order.

Recommended order:

1. Fixes to base.plot_box() #179 (fix/base-plot-box) — already approved, CI green, ready now
2. Fix/twist pitch clean #196 (fix/twist-pitch-clean) — isolated to twist.py
3. test: force headless Matplotlib backend for local + CI test runs #197 (fix/headless-tests-locally) — isolated to test files only (this one)
4. docs: fix Sphinx docstring formatting and reST syntax errors #195 (docs/sphinx-docstring-fixes) — touches ~18 files, overlaps most of the PRs below; docstring-only/low-risk, so merging it here means the functional PRs below each only rebase against it once
5. feat(pose,quaternion): add __imatmul__ (@=) operator #199 (feat/imatmul) — only conflicts with refactor: consistent printline() across pose and quaternion classes #200
6. fix(base): export tr2pos2, pos2tr2, tr2adjoint2; rename tradjoint2 #194 (fix/export-base-functions) — conflicts with perf: defer Matplotlib import until something actually plots #198 and refactor: consistent printline() across pose and quaternion classes #200
7. perf: defer Matplotlib import until something actually plots #198 (perf/lazy-matplotlib-import) — conflicts with fix(base): export tr2pos2, pos2tr2, tr2adjoint2; rename tradjoint2 #194 and refactor: consistent printline() across pose and quaternion classes #200
8. refactor: consistent printline() across pose and quaternion classes #200 (refactor/printline) — touches the most shared surface (conflicts with fix(base): export tr2pos2, pos2tr2, tr2adjoint2; rename tradjoint2 #194, perf: defer Matplotlib import until something actually plots #198, feat(pose,quaternion): add __imatmul__ (@=) operator #199); merging it last means it absorbs one final rebase instead of three others rebasing against it