Overview

test: isolate EP output directories for parallel runs

Plan

Override the graphical suite configuration fixture with per-test tmp_path output and balanced config push/pop. Ensure explicit output-directory consumers and teardown use the isolated location. Run graphical tests serially and under xdist with identical collected counts; preserve production identifiers and behavior.

Tier: glance — merge mode: human /prm; this session stops at PR-open by explicit agreement.

Detailed implementation plan

Affected repository: PyAutoLabs/PyAutoFit. Files: test_autofit/graphical/conftest.py and explicit shared-output call sites if needed.
Override the graphical suite configuration fixture with per-test tmp_path output and balanced config push/pop. Ensure explicit output-directory consumers and teardown use the isolated location. Run graphical tests serially and under xdist with identical collected counts; preserve production identifiers and behavior.
Run focused regression tests and the full PyAutoFit suite; compare failures against main rather than claiming unverified green. No public signature changes planned.

Branch survey: main clean; no existing PyAutoFit claim in current Mind active.md; no open PyAutoFit PRs or matching open issue found. Branch: feature/ep-test-output-isolation. These four separately reviewed tasks are coordinated in one user-authorized group, with isolated worktrees and sequential edits.

Authorization and Heart RED

User request: "Find 4 grouped tasks and do them". After the exact Heart RED reasons were disclosed and development-only authorization requested, the user replied "Yes I authorise". Applies to selecting, implementing and testing these four tasks and opening reviewable PRs; no merge or release. Tests and author diff review will be recorded at PR-open, not claimed in advance.

Published RED reasons:

• install verification FAILED (testpypi; checks F)
• release validation FAILED (stage integrate)
• PyAutoHeart/workflows/Release Integrate: failure
• manifest_drift/checks/workspace checkouts (manifest ↔ disk): red
• Release readiness: release validation FAILED (stage integrate)
• Release readiness: install verification FAILED (testpypi; checks F)
• PyAutoArray/test_autoarray/inversion/inversion/interferometer/test_interferometer.py::test__interferometer_sparse_operator__mapped_reconstructed_data_dict_matc…
• Unit-test timing: 1 test regressions (>3× baseline), 11 slow (>1.5×)
• Release validation: NOT release_ready — v2026.10.25.1.dev83701 profile=release (2026-10-25T07:21:05+00:00)
• validation_report/stages/integrate: fail
• verify_install FAILED: verify_install FAILED
• F: colab gate: finufft (declared by autoarray, absent on Colab, imported)
• Install verification: FAILED (testpypi; F) (2026-10-25T07:42:48.134987+00:00)
• verify_install: red
• PyAutoBrain: red
• PyAutoLens: red
• autolens_profiling: red
• Worktree drift: 1 orphan / 0 missing / 3 dirty

Original prompt

EP test suite is not xdist-safe: tests share one on-disk output directory

Type: bug
Target: PyAutoFit
Repos:

• PyAutoFit
  Difficulty: small
  Autonomy: safe
  Priority: medium
  Status: formalised
  Consequence: glance
  Witness: 'python -m pytest test_autofit/graphical -q -n auto' passes with the same count as the serial run (279 on main e331e33, 281 after PyAutoFit#1608).
  Review-minutes: 3
  Unattended: ready

EP test suite is not xdist-safe: tests share one on-disk output directory
Type: bug
Target: PyAutoFit
Difficulty: small
Autonomy: safe
Priority: medium
Witness: 'python -m pytest test_autofit/graphical -q -n auto' passes with the same count as the serial run (279 on main e331e33, 281 after PyAutoFit#1608).

Measured 2026-09-11 on clean main e331e33: 'python -m pytest test_autofit/graphical -q -n auto' gives 17 failed / 268 passed, while the same suite serial gives 279 passed. Signature: FileNotFoundError on test_autofit/output//mean_field_evolution.png inside PIL Image.save (e.g. test_autofit/graphical/gaussian/test_optimizer.py::test_default; also hierarchical/test_optimise.py::test_optimise, regression/test_linear_regression.py::test_exact_updates and test_laplace, stochastic/test_regression.py::test_stochastic_linear_regression, gaussian/test_optimizer.py::TestDynesty::test_null_paths). Several EP tests write their visualisation output under the same search-identifier path in test_autofit/output/, so one xdist worker's teardown removes the directory another worker is still writing into. The remote-session doctrine says to run suites with -n auto, so this suite silently cannot follow it and PyAutoFit#1608 had to be judged serially.

Fix: give every EP test that writes output its own tmp_path-rooted output directory (a conftest fixture pointing the conf output path at tmp_path, or a unique_tag per test), so the suite is parallel-safe.

Related: complete/2026/10/prior-stable-test-ids.md covers the other xdist breakage (collection-id mismatch in test_prior_properties), a separate cause.

Implementation and validation — 2026-10-25

PR: #1728

Full library suite: 3913 passed, 1 skipped, 1 xfailed. Graphical suite: 329 passed serially and 329 passed with four xdist workers; unchanged main reproduced 13 failures / 316 passed under xdist. Python 3.12.14. Author diff review passed; production code and output identifier generation untouched.

Published vitals: organism RED / PyAutoFit scope GREEN. Local readiness STALE. Existing development-only authorization applies; no merge or release. Awaiting required CI and human review.