Skip to content

v1.16.1.0

Choose a tag to compare

@jorenham jorenham released this 01 Aug 02:41
· 323 commits to master since this release
v1.16.1.0
8244740

This is the first scipy-stubs release that officially supports SciPy 1.16.1. In these last 4 weeks of hard work, we've made many improvements to the accuracy and correctness of many annotations, and improved compatibility with older NumPy versions.
If you encounter any regressions or other issues, please report them at https://github.com/scipy/scipy-stubs/issues.

This release requires Python 3.11-3.14, NumPy 1.25.2 or greater, and optype 0.13.1 or greater.

What's Changed

πŸ‘½οΈ SciPy 1.16.1 Changes

✨ Improvements

  • 🏷️ linalg: stub more of the internal scipy machinery by @jorenham in #710
  • ✨ linalg: support sparse arrays in clarkson_woodruff_transform by @jorenham in #711
  • ✨ stub scipy._cyutility by @jorenham in #712
  • 🏷️ _lib: stub _ccallback_c and _array_api[_compat_vendor] by @jorenham in #713
  • 🏷️ cluster: stub internal cython submodules by @jorenham in #714
  • 🏷️ integrate: stub integrate._rules.* (private api) by @jorenham in #715
  • 🏷️ integrate: stub the remaining private C-extension modules by @jorenham in #716
  • 🏷️ interpolate: stub the remaining private C- and cython modules by @jorenham in #717
  • 🏷️ io.matlab: stub private cython modules by @jorenham in #718
  • 🏷️ sparse: added stubs for all formats in kron and kronsum by @JulVandenBroeck in #719
  • 🏷️ ndimage: stub the remaining private modules by @jorenham in #720
  • 🏷️ optimize: stub the remaining private modules by @jorenham in #722
  • 🏷️ signal: stub the remaining private modules by @jorenham in #723
  • 🏷️ sparse: stub the remaining private modules by @jorenham in #724
  • 🏷️ spatial: stub the remaining private modules by @jorenham in #725
  • 🏷️ special: stub the remaining private modules by @jorenham in #726
  • 🏷️ stats: stub the remaining private modules by @jorenham in #727
  • ✨ stats: improved [differential_]entropy annotations by @jorenham in #738
  • ✨ stats: improved Covariance generic type parameter inference by @jorenham in #739
  • ✨ ndimage: improved fourier_* annotations by @jorenham in #740
  • ✨ linalg: improved basic solve* annotations by @jorenham in #741
  • ✨ linalg: improved inv annotations by @jorenham in #743
  • ✨ linalg: improved det annotations by @jorenham in #744
  • πŸ§‘β€πŸ’» vscode tasks for basedpyright, mypy, and stubtest by @jorenham in #754
  • ✨ linalg: improved eigh return type inference for float64 by @jorenham in #755
  • 🏷️ sparse: overhauled stubs for diags_array, eye_array, and their matrix equivalents by @JulVandenBroeck in #733
  • ✨ integrate: improve the odeint signature by @jorenham in #763
  • ✨ integrate: improved solve_ivp annotations by @jorenham in #765
  • ✨ signal: improved convolution and correlation functions by @jorenham in #767
  • ✨ signal: improved filter function annotations by @jorenham in #768
  • Add hstack and vstack overloads for different sparse formats by @JulVandenBroeck in #774

πŸ› Fixes

  • πŸ› sparse.linalg: allow passing arrays to expm_multiply by @jorenham in #708
  • πŸ› sparse.linalg: fix several LinearOperator issues by @jorenham in #709
  • 🏷️ sparse: improved init stubs of sparse arrays with shape as positional argument. by @JulVandenBroeck in #704
  • πŸ› sparse: csc array/matrix constructor fixes by @jorenham in #753
  • 🩹 fix missing type arguments on numpy<2.3 by @jorenham in #779
  • 🚨 fftpack: fix stubtest error on numpy<2 by @jorenham in #786
  • βœ… fix failing type-tests on numpy==2.2.* by @jorenham in #787
  • βœ… fix failing type-tests on numpy==2.1.* by @jorenham in #788
  • βœ… fix failing type-tests on numpy<2.1 by @jorenham in #789
  • πŸ’š fix CI numpy testing matrix by @jorenham in #778
  • πŸ› interpolate: Fix incorrect return type for make_lsq_spline by @elael in #776
  • 🩹 optimize: fix compatibility issues with ABCPolyBase on numpy==2.1 by @jorenham in #780
  • 🩹 fix compatibility issues with np.[c]longdouble on numpy<2.2 by @jorenham in #781
  • πŸ—‘οΈ special.comb: remove expired @deprecated non-integer exact=True overloads by @jorenham in #793

πŸ“ Documentation

  • add a "who's using scipy-stubs" section to the readme by @jorenham in #732
  • add scanpy as dependent project by @jorenham in #747
  • CONTRIBUTING.md fixes by @jorenham in #760
  • add xarray, radioactivedecay, neurogym, and foamlib to the list of dependents by @jorenham in #772

⬆️ Dependencies

Other Changes

  • πŸ”– prepare for further development by @jorenham in #701
  • πŸ”¨ dev script for listing all (private) unstubbed modules by @jorenham in #706
  • 🎨 use 88 as max line length for scipts/*.py by @jorenham in #721
  • πŸ“Œ pin CI action version tags by @jorenham in #728
  • ♻️ always use optype.numpy.compat for abstract scalar types by @jorenham in #730
  • ♻️ reduce reliance on "fake" *._typing modules by @jorenham in #731
  • ♻️ explicit onp.AnyFloat64DType | None by @jorenham in #735
  • 🎨 workaround for path-dependent pyright bug by shuffling disjoint overloads by @jorenham in #745
  • 🎨 linalg: simplified _decomp_svd type aliases by @jorenham in #749
  • πŸ€– correct TypeVarTuple naming in copilot instructions by @jorenham in #764
  • 🐴 work around false positive overload-overlap mypy errors on numpy>=2.2 by @jorenham in #782
  • 🀑 work around false positive reportOverlappingOverload pyright errors on numpy<2.1 by @jorenham in #783
  • 🐴 special: work around false positive overload-overlap mypy errors on numpy<2.1 by @jorenham in #785
  • πŸ“„ explicit license-files by @jorenham in #795
  • 🧱 uv_build by @jorenham in #796

πŸŽ‰ New Contributors

Full Changelog: v1.16.0.2...v1.16.1.0