Make the array-api-strict suite green and promote the strict CI job

[mwcraig]

Follow-up to #1000/#1001, finishing the strict-backend items on the array-API migration list.

Commit 1 — make the strict suite green. The one remaining strict failure in ccdproc's own code was flat_correct checking the flat's mask with the mask.any() method, which spec-only namespaces like array-api-strict don't provide; it now uses xp.any. The stale backend_xfail on test_flat_correct_masked_flat_with_immutable_array (its reason described an older failure mode) is removed and the test passes everywhere. The two other strict failures are astropy-upstream and get accurate backend_xfail marks in line with the rest of the suite:

• test_unit_mismatch_behaves_as_expected: astropy's arithmetic mixin builds the Quantity with data << unit, and array-api-strict rejects the astropy unit operand (see Array API: units/Quantity handling with non-numpy arrays #936).
• test_rebin_ccddata[True-True]: astropy's CCDData.mask setter coerces the mask with np.asarray, which fails for an array on the non-default device.

Commit 2 — promote the strict CI job. With the suite green (839 passed, 46 xfailed, 0 failed locally), the expected-failures carve-out is retired: py313-strict moves into the regular ci-tests matrix where a failure fails CI, and the continue-on-error step, outcome-artifact plumbing, and the strict_status.yml reporting workflow are removed.

Verified locally on the full suite for every backend: numpy (890 passed), dask (879 passed, escape-baseline ratchet enforced), jax (885 passed, ratchet enforced), array-api-strict (839 passed, 46 xfailed).

— Written by Claude at @mwcraig's direction.

🤖 Generated with Claude Code

https://claude.ai/code/session_01LVxnTLrWKStBdPwcmDNhxA

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.96%. Comparing base (c91b593) to head (fd5aab8).
⚠️ Report is 6 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1005   +/-   ##
=======================================
  Coverage   97.96%   97.96%           
=======================================
  Files           9        9           
  Lines        1912     1912           
=======================================
  Hits         1873     1873           
  Misses         39       39           

Flag	Coverage Δ
dask	97.16% <100.00%> (ø)
jax	97.32% <100.00%> (ø)
numpy	97.85% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

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

[Copilot]

Pull request overview

Fixes the remaining project-owned array-API-strict failure and promotes the strict backend to required CI.

Changes:

• Uses xp.any for backend-compatible mask checks.
• Marks two Astropy limitations as expected strict-backend failures.
• Moves strict testing into the main CI matrix and removes legacy reporting.

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
CHANGES.rst	Documents strict compatibility and CI promotion.
ccdproc/core.py	Makes flat-mask detection array-API compatible.
ccdproc/tests/test_ccdproc.py	Updates strict-backend expected failures.
ccdproc/tests/test_rebin.py	Marks an Astropy mask-setter incompatibility.
.github/workflows/ci_tests.yml	Makes strict test failures fail CI.
.github/workflows/strict_status.yml	Removes obsolete status reporting.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[mwcraig]

Confirmed — backend_xfail is applied non-strict (conftest.py adds pytest.mark.xfail(reason=..., strict=False)), so with the function-level marker a genuine regression in the currently passing [False-False] case would be reported as an expected failure rather than failing the job. Will scope the marker to the failing case via the suite's existing per-parameter pattern (as with _GAIN_APPLY_STRICT_XFAIL in test_cosmicray.py):

@pytest.mark.parametrize(
    "mask_data, uncertainty",
    [
        (False, False),
        pytest.param(
            True,
            True,
            marks=pytest.mark.backend_xfail(
                "array-api-strict",
                reason="astropy's CCDData.mask setter coerces the mask with "
                "np.asarray, which fails for an array-api-strict array on "
                "the non-default device",
            ),
        ),
    ],
)
def test_rebin_ccddata(mask_data, uncertainty):

— Written by Claude at @mwcraig's direction.

[mwcraig]

Shorten

[mwcraig]

Proposed shorter entry:

- The ``py313-strict`` CI job is now part of the regular CI matrix and fails
  CI on a test failure; the expected-failures carve-out and the
  ``strict_status.yml`` reporting workflow are retired. [#1005]

— Written by Claude at @mwcraig's direction.

[mwcraig]

Shorten

[mwcraig]

Proposed shorter entry:

- ``flat_correct`` now checks the flat's mask with ``xp.any`` instead of the
  ``mask.any()`` method, which spec-only array-API namespaces do not
  provide. [#1005]

— Written by Claude at @mwcraig's direction.