build(env): raise the shared baseline to PyMC 6.2 - #79
Merged
Conversation
Move the canonical conda-forge core to current conda-forge releases and cut v0.9.0. PyMC 6.2.0 requires pytensor >=3.2.2,<3.3 (verified against the 6.2.0 sdist metadata), so the floor moves from 6.1.0 to 6.2.0. The three documented caps all still hold and are re-verified here: - numba stays <=0.65.1 because pytensor-base 3.2.3 still declares `numba >=0.58,<=0.65.1`, even though numba 0.66.0 is released. - numpy stays <2.5 because numba pins it there — and note numba 0.66.0 does too, so lifting the numba cap alone would not unlock numpy 2.5. - pyarrow stays >=24.0.0 because libarrow 25.x needs libabseil >=20260526 while every jaxlib 0.10.2 build still pins libabseil <20260108. Remaining floors track current releases: pandas 3.0.5, matplotlib 3.11.1, preliz 0.27.1, h5py 3.16.0, h5netcdf 1.8.1, python-graphviz 0.21, plus the research add-ons (lightgbm 4.7.0, polars 1.43.1, duckdb 1.5.5, xarray 2026.7.0) and the pip tail (ruff 0.16.1, setuptools 83.0.0, hatchling 1.31.0, jupytext 1.19.5). Verified with `conda create --dry-run` on both linux-64 and osx-arm64: the environment solves to pymc 6.2.0 / pytensor 3.2.3 / numba 0.65.1 / numpy 2.4.6 / pyarrow 24.0.0. `dse-check-env` reports no drift and the 276 tests pass. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This was referenced Jul 31, 2026
Merged
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Note
Drafted by a LLM-based AI tool (Claude Code/Opus 5).
Raises the canonical conda-forge core to current releases, moves the PyMC floor to 6.2, and cuts
dse-research-utilsv0.9.0.PyMC 6.2
pymc 6.2.0requirespytensor>=3.2.2,<3.3(from the 6.2.0 sdist metadata), so the floor moves6.1.0→6.2.0and the comment now records the real constraint rather than a paraphrase.The three documented caps all still hold
Each was re-verified against live conda-forge metadata rather than carried forward on trust:
numba<=0.65.1pytensor-base 3.2.3still declaresnumba >=0.58,<=0.65.1, even though numba 0.66.0 is outnumpy<2.5numpy >=1.22.3,<2.5too, so lifting the numba cap alone would not unlock numpy 2.5pyarrow>=24.0.0libarrow 25.0.0needslibabseil >=20260526; everyjaxlib 0.10.2build still pinslibabseil >=20260107.1,<20260108The numpy comment previously implied the cap would move with the numba cap. It won't, so the comment now says so explicitly.
Floors moved to current releases
Core:
pandas 3.0.5,matplotlib 3.11.1,preliz 0.27.1,h5py 3.16.0,h5netcdf 1.8.1,python-graphviz 0.21. Add-ons:lightgbm 4.7.0,polars 1.43.1,duckdb 1.5.5,xarray 2026.7.0. Pip tail:ruff 0.16.1,setuptools 83.0.0,hatchling 1.31.0,jupytext 1.19.5,pyyaml 6.0.3.Unchanged because they are already at the current release:
python 3.14.6,scipy 1.18.0,scikit-learn 1.9.0,statsmodels 0.14.6,nutpie 0.16.11,numpyro 0.21.0,jax 0.10.2,arviz* 1.2.0,graphviz 14.1.2,xgboost 3.3.0,shap 0.52.0,pyreadstat 1.3.5,pytest 9.1.1.Verification
conda create --dry-runon bothlinux-64andosx-arm64: solves topymc 6.2.0/pytensor 3.2.3/numba 0.65.1/numpy 2.4.6/pyarrow 24.0.0/libabseil 20260107.1.dse-check-env environment.yml— no drift.pytest— 276 passed.npm run spellcheckandnpm run format:check— clean.Follow-up
Tag
v0.9.0after merge. The three downstream repos have companion PRs that pip-installdse-research-utilsfrom that tag, so they stay red until the tag exists.