ci: drop archived pytest-pydocstyle; xfail test_xipm_theo (numpy>=2.5)#70
Merged
Conversation
Two pre-existing CI breakages, both dependency drift, neither a source bug: 1. pytest-pydocstyle is archived upstream and its pytest_collect_file hook still declares the legacy `path` arg that pytest 9.1 removed from the hookspec, so pytest aborted at startup with PluginValidationError. Since the plugin will never be updated, remove it rather than pin pytest back. Docstring linting, if wanted, should move to a maintained tool (ruff's pydocstyle `D` rules). 2. With pytest running again, test_cosmo::test_xipm_theo fails on Python 3.12 (numpy 2.5.0): pyccl 3.3.4 calls float() on a non-0-d array in boltzmann.py, which numpy>=2.5 turns into a hard TypeError. xfail it (conditioned on numpy>=2.5, non-strict) as a stopgap. The real fix and the broader "where should cosmology code live" question are tracked in #71. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BXY93SBtDvLLrxWX8Vh4KW
cailmdaley
force-pushed
the
ci/pin-pytest-9.1
branch
from
00bfd0b to
105777a
Compare
cailmdaley
changed the title
cailmdaley
added a commit
that referenced
this pull request
Jul 3, 2026
…ry into cs_util.cosmo Move the cosmology theory-curve machinery from sp_validation.cosmology into cs_util.cosmo so cs_util is the single home for the collaboration's cosmology primitives (per a conversation with M. Kilbinger). Additive: no existing cs_util.cosmo symbol changes, so shear_psf_leakage (get_cosmo_default, xipm_theo) is unaffected. Adds: - PLANCK18 fiducial-cosmology dict - get_cosmo: flexible Planck18-default cosmology builder (individual, CCL, CAMB, or CosmoCov parameter formats) - get_theo_c_ell, c_ell_to_xi, get_theo_xi: angular power spectra and shear correlation functions, CCL and CAMB backends - _ccl_to_camb / _camb_to_ccl / _cosmocov_to_ccl parameter converters - tests/test_cosmology.py, ported from sp_validation Caps scipy < 1.18 (mirroring sp_validation): scipy 1.18's FITPACK-to-C port changed a RectBivariateSpline return shape that breaks camb's BBN Y_He predictor in set_cosmology (and pyccl's Boltzmann spline path likewise). With the cap the full cosmo suite passes for real on py3.10 AND py3.12 (numpy 2.5): 31/31. This also removes the numpy>=2.5 xfail on test_xipm_theo -- a stopgap for the same root cause (#70/#71) -- which now passes. Bumps version 0.2.1 -> 0.2.2. No new dependency (camb, pyccl already required). Part of #75; the sp_validation delete + repoint follows in a companion PR. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013NCc4BHtZ4Lx6n9zuUwEeP
cailmdaley
added a commit
that referenced
this pull request
Jul 6, 2026
…ry into cs_util.cosmo (#76) Move the cosmology theory-curve machinery from sp_validation.cosmology into cs_util.cosmo so cs_util is the single home for the collaboration's cosmology primitives (per a conversation with M. Kilbinger). Additive: no existing cs_util.cosmo symbol changes, so shear_psf_leakage (get_cosmo_default, xipm_theo) is unaffected. Adds: - PLANCK18 fiducial-cosmology dict - get_cosmo: flexible Planck18-default cosmology builder (individual, CCL, CAMB, or CosmoCov parameter formats) - get_theo_c_ell, c_ell_to_xi, get_theo_xi: angular power spectra and shear correlation functions, CCL and CAMB backends - _ccl_to_camb / _camb_to_ccl / _cosmocov_to_ccl parameter converters - tests/test_cosmology.py, ported from sp_validation Caps scipy < 1.18 (mirroring sp_validation): scipy 1.18's FITPACK-to-C port changed a RectBivariateSpline return shape that breaks camb's BBN Y_He predictor in set_cosmology (and pyccl's Boltzmann spline path likewise). With the cap the full cosmo suite passes for real on py3.10 AND py3.12 (numpy 2.5): 31/31. This also removes the numpy>=2.5 xfail on test_xipm_theo -- a stopgap for the same root cause (#70/#71) -- which now passes. Bumps version 0.2.1 -> 0.2.2. No new dependency (camb, pyccl already required). Part of #75; the sp_validation delete + repoint follows in a companion PR. Claude-Session: https://claude.ai/code/session_013NCc4BHtZ4Lx6n9zuUwEeP Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Two pre-existing CI breakages on
develop, both dependency drift — neither a source change.1. pytest-pydocstyle is archived and breaks under pytest 9.1
CI aborted at pytest startup:
pytest-pydocstyle'spytest_collect_filestill declares the legacypatharg that pytest 9.1 removed from the hookspec. The plugin's upstream is archived, so it will never be fixed — remove it rather than pin pytest back. Docstring linting, if wanted, should move to a maintained tool (ruff's pydocstyleDrules), ideally folded into the project-wide ruff effort.2. test_xipm_theo fails under numpy ≥ 2.5
With pytest running again,
test_cosmo.py::test_xipm_theofails on Python 3.12 (numpy 2.5.0): pyccl 3.3.4 callsfloat()on a non-0-d array inboltzmann.py:119, which numpy ≥ 2.5 turns into a hardTypeError. 3.10 passes (numpy 2.2.6). xfail it (conditioned on numpy ≥ 2.5, non-strict) as a stopgap.The real fix — and the broader question of where cosmology code should live (cs_util.cosmo vs sp_validation's inline re-implementation; note cs_util.cosmo is consumed by shear_psf_leakage) — is tracked in #71.
Verified locally: full suite green under latest pytest with the plugin removed; the numpy ≥ 2.5 xfail path exercised.
— Claude on behalf of Cail