v1.16.0.1

This is the second scipy-stubs release that supports SciPy 1.16. In these last 2 weeks of hard work, we've also made many improvements to the accuracy, correctness, and documentation of scipy-stubs.
If you encounter any regressions or other issues, please report them at https://github.com/scipy/scipy-stubs/issues.

This release requires Python 3.11-3.13 and NumPy 1.25.2 or greater.

What's Changed

👽️ SciPy Changes

• scipy 1.16 changes for optimize.fmin_l_bfgs_b by @jorenham in #643

✨ Improvements

• stats: Annotate the private functions and consts in _continuous_distns by @jorenham in #650
• cluster.vq: improved dtype annotations by @jorenham in #651
• interpolate: Generic Krogh and barycentric interpolator types by @jorenham in #667
• optimize: Generic Bounds type by @jorenham in #675
• sparse.linalg: Improved linear solver function annotations by @jorenham in #678
• fft: improved *fft, *fft2 and *fftn annotations by @jorenham in #681
• fft: improved [i]fftshift annotations by @jorenham in #682
• fft: sync (private) signatures of _realtransforms and _realtransforms_backend by @jorenham in #684

🐛 Fixes

• stats.qmc: Add missing seed constructor parameters by @jorenham in #647
• optimize: accept tuple functions in fmin_l_bfgs_b by @jorenham in #649
• cluster.vq: Add missing seed parameter to kmeans and kmeans2 by @jorenham in #652
• interpolate: Add missing seed kwarg to the BarycentricInterpolator constructor by @jorenham in #654
• linalg: Add missing seed kwarg to clarkson_woodruff_transform by @jorenham in #655
• optimize: Add missing seed kwargs to the global optimize functions by @jorenham in #656
• sparse: Add missing random_state kwarg to random_array and rand[om] by @jorenham in #657
• sparse.linalg: Add missing random_state kwarg to svds by @jorenham in #658
• spatial.distance: Add missing seed kwarg to directed_hausdorff by @jorenham in #659
• spatial.transform: Add missing random_state kwarg to Rotation.random by @jorenham in #660
• stats: Add missing random_state kwargs by @jorenham in #661
• sparse: fix sparse array/matrix indexing by @jorenham in #664
• signal: simplify get_window input type by @jorenham in #665
• sparse: Transposition duality of CSR and CSC arrays by @jorenham in #669
• sparse.linalg: Fix factorize return type annotations by @jorenham in #679
• interpolate: consistent RBFInterpolator generic type parameter ordering by @jorenham in #685

⬆️ Dependencies

• require scipy>=1.16.0<1.17 by @jorenham in #642
• bump ruff to 0.12.1 by @jorenham in #680
• bump basedpyright to 1.29.5 by @jorenham in #691

📝 Documentation

• generic type reference docs by @jorenham in #673
• support versions of generic types by @jorenham in #690

Other Changes

• 🏷️ io: loadmat return dict with gradual value types by @jorenham in #676
• 💬 update string literals for SciPy 1.16.0 by @jorenham in #683
• 🔧 update release note config by @jorenham in #686
• 🧑‍💻 update editorconfig and vscode settings by @jorenham in #687
• 🧑‍💻 dprint formatter for markdown, json, toml and yaml by @jorenham in #688
• ⬆️ Bump crate-ci/typos from 1.33.1 to 1.34.0 in the actions group by @dependabot in #689

Full Changelog: v1.16.0.0...v1.16.0.1

v1.16.0.0

This is the first scipy-stubs release that supports SciPy 1.16. In these last 6 weeks of hard work, we've also made many improvements to the accuracy, correctness, and testing coverage of scipy-stubs.
If you encounter any regressions or other issues, please report them at https://github.com/scipy/scipy-stubs/issues.

This release requires Python 3.11-3.13 and NumPy 1.25.2 or greater.

⭐ Highlights

• Support for SciPy 1.16 (currently 1.16.0rc2) (release notes)
• Support for NumPy 2.3 (currently 2.3.0) (release notes)
• scipy.linalg: Low-level BLAS and LAPACK function annotations
• scipy.sparse: Many sparse array and matrix improvements, including preliminary shape-typing support
• Improved type-testing coverage, which uncovered several bugs that have now been fixed

See below for the complete list of PR's that are included in this scipy-stubs release.

What's Changed

👽️ SciPy 1.16.0 Changes

• scipy 1.16 changes for _lib by @jorenham in #525
• scipy 1.16 changes for cluster.vq.whiten by @jorenham in #526
• scipy 1.16 changes for interpolate.dfitpack by @jorenham in #527
• scipy 1.16 change for interpolate.RectBivariateSpline by @jorenham in #528
• scipy 1.16 change for interpolate.make_smoothing_spline by @jorenham in #529
• scipy 1.16 addition for io.matlab by @jorenham in #530
• scipy 1.16 changes for linalg.sqrtm by @jorenham in #531
• scipy 1.16 new function ndimage.vectorized_filter by @jorenham in #537
• scipy 1.16 batching support in scipy.linalg by @jorenham in #535
• scipy 1.16 changes for optimize._differentiable_functions by @jorenham in #538
• scipy 1.16 changes for optimize.least_squares by @jorenham in #539
• scipy 1.16 changes for optimize._numdiff by @jorenham in #540
• scipy 1.16 changes for optimize._trustregion_constr by @jorenham in #541
• scipy 1.16 changes for optimize.nnls by @jorenham in #542
• scipy 1.16 changes for optimize.slsqp by @jorenham in #544
• scipy 1.16 additions for signal._short_time_fft by @jorenham in #546
• scipy 1.16 new function signal.firwin_2d by @jorenham in #548
• scipy 1.16 changes for signal.__all__ by @jorenham in #549
• scipy 1.16 changes for signal.windows by @jorenham in #550
• signal: dedupe the @deprecated public submodule function definitions by @jorenham in #551
• scipy 1.16 deprecations for linalg.{logm, signm, sqrtm} by @jorenham in #553
• scipy 1.16 new internal module _lib._sparse by @jorenham in #555
• scipy 1.16 changes for sparse._base by @jorenham in #556
• scipy 1.16 removals for sparse by @jorenham in #557
• scipy 1.16 changes for sparse.csgraph._shortest_path by @jorenham in #558
• scipy 1.16 changes for sparse._sputils by @jorenham in #559
• scipy 1.16 expired deprecations for sparse._dok by @jorenham in #560
• scipy 1.16 additions for spatial.transform._rotation by @jorenham in #561
• scipy 1.16 new class spatial.transform.RigidTransform by @jorenham in #562
• scipy 1.16 changes for special._support_alternative_backends by @jorenham in #563
• scipy 1.16 new function stats.quantile by @jorenham in #564
• scipy 1.16 changes for stats.make_distribution by @jorenham in #565
• scipy 1.16 new class stats._distribution_infrastructure by @jorenham in #566
• scipy 1.16 new class stats.Binomial by @jorenham in #567
• scipy 1.16 removal of stats._mvn by @jorenham in #569
• scipy 1.16 changes for stats.tukey_hsd by @jorenham in #570
• scipy 1.16 changes for stats.f_oneway by @jorenham in #573
• scipy 1.16 removal of stats.trapz by @jorenham in #574
• scipy 1.16 changes for stats.multivariate_normal.[log]cdf by @jorenham in #575
• scipy 1.16 changes for stats.gaussian_kde.integrate_box by @jorenham in #576
• scipy 1.16 changes for stats.boxcox_llf by @jorenham in #577
• scipy 1.16 changes for stats.{siegel,theil}slopes by @jorenham in #580
• scipy 1.16 changes for stats.gstd by @jorenham in #581
• scipy 1.16 changes for stats.power_divergence and stats.chisquare by @jorenham in #582
• scipy 1.16 expired deprecation for stats.linregress by @jorenham in #583
• scipy 1.16 changes for stats.linregress by @jorenham in #584
• scipy 1.16 changes for stats.pointbiserialr by @jorenham in #585
• scipy 1.16 changes for stats.{kendall,weighted}tau by @jorenham in #586
• scipy 1.16 changes for version by @jorenham in #587

✨ Improvements

• signal: generic ShortTimeFFT type-param and improved annotations by @jorenham in #547
• linalg: improved _matfuncs return dtypes by @jorenham in #554
• stats: improved hypothesis test function return type annotations by @jorenham in #571
• linalg: low-level BLAS function annotations by @jorenham in #594
• linalg: low-level LAPACK function annotations by @jorenham in #595
• optimize: stub _slsqplib cpython extension module by @jorenham in #596
• linalg: stub _decomp_interpolative cpython extension module by @jorenham in #597
• stats: stub the _rcont subpackage by @jorenham in #598
• sparse: improved annotations and code cleanup by @jorenham in #602
• sparse: type testing & sparse matrix and array improvements by @jorenham in #605
• constants: type-tests & improved annotations by @jorenham in #610
• sparse: constructor function type-tests by @jorenham in #619
• sparse: improved integer matrix/array method annotations by @jorenham in #624
• sparse: fix and improve get_index_type return type by @jorenham in #625
• sparse: simplified gradual load_npz return type by @jorenham in #626
• preliminary support for Python 3.14 by @jorenham in #631
• sparse: Improved shape-typing support for CSR and DOK arrays by @jorenham in #634
• fft: improved [i]fht() annotations by @jorenham in #640
• fft: improved discrete sine- and cosine transform return types by @jorenham in #641

🐛 Fixes

• signal: accept numbers as args to width and plateau_size for find_peaks() by @jorenham in #519
• fix numpy<2.2 compatability with np.bool_ by @jorenham in #579
• prefer gradual shape-type defaults for generics by @jorenham in #601
• sparse: covariant sparse matrix/array types by @jorenham in #623
• sparse: fix and improve get_index_type return type by @jorenham in #625
• sparse: distinguish between different values for argument which in eigs and eigsh by @lcwllmr in #633
• optimize: fix false rejection of identity functions in minimize by @jorenham in https://github.com/scipy/scipy-stubs/p...

v1.15.3.0 (scipy 1.15.3)

⭐ Highlights

• Support for SciPy 1.15.3 and NumPy 2.2.5

What's Changed

🐛 Fixes

• Correct key name in _OptionsMILP by @it176131 in #511

⬆️ Dependencies

• upgrade development dependencies by @jorenham in #508
• bump astral-sh/setup-uv to v6 by @jorenham in #509
• update the dev dependencies by @jorenham in #514
• upgrade scipy to 1.15.3 by @jorenham in #516

Other Changes

• 🔖 prepare for further development by @jorenham in #507
• Remove License Trove Classifier from pyproject.toml by @it176131 in #513
• 🛂 explicit CI permissions by @jorenham in #515

New Contributors

• @it176131 made their first contribution in #511

Full Changelog: v1.15.2.2...v1.15.3.0

v1.15.2.2 (scipy 1.15.2)

What's Changed

✨ Improvements

• ndimage: improved filter and measurement function annotations by @jorenham in #499
• optimize: narrowed OptimizeResult.{fun,x} attrs for minimize[_scalar] by @jorenham in #504

🐛 Fixes

• adjust pyright config for optype.numpy compatibility by @jorenham in #491
• interpolate: accept ndarray for points in RegularGridInterpolator by @jorenham in #501
• optimize: accept a scalar for x0 in minimize by @jorenham in #502
• optimize: accept a scalar for x0 in fmin* by @jorenham in #503

📝 Documentation

• prune outdated README.md bits by @jorenham in #505

⬆️ Dependencies

• upgrade optype to 0.9.2 by @jorenham in #481
• bump ruff to 0.11.0 by @jorenham in #484
• bump numpy to 2.2.4 by @jorenham in #487
• update development dependencies by @jorenham in #489
• bump optype to 0.9.3 by @jorenham in #490
• bump ruff to 0.11.4 by @jorenham in #506

Other Changes

• 🔖 prepare for further development by @jorenham in #480
• 👷 ensure pyright compatibility by @jorenham in #482
• 🔧 adjust basedmypy config for mypy compat by @jorenham in #483
• 🔧 fix vscode settings syntax by @jorenham in #485
• 🔧 remove files from [tool.mypy] by @jorenham in #486
• 💚 fix mypy_primer failure and mypy/basedmypy compat by @jorenham in #500

Full Changelog: v1.15.2.1...v1.15.2.2

v1.15.2.1 (scipy 1.15.2)

What's Changed

✨ Improvements

• interpolate: improved support for complex splines by @jorenham in #477

🐛 Fixes

• optimize: allow calling minimize_scalar without a bracket by @jorenham in #467
• transform: accept proper Euler angles in Rotation.as_euler by @KiSchnelle in #470
• interpolate: accept scalar input in __call__ by @jorenham and @ev-br in #476 and #478

📝 Documentation

• add a "typed" badge by @jorenham in #468

Dependencies

• ⬆️ bump ruff to 0.9.7 by @jorenham in #460
• ⬆️ basedpyright 1.28.1 and ruff 0.9.9 by @jorenham in #466

Other Changes

• 🔖 prepare for further development by @jorenham in #458
• 💄 add-trailing-comma by @jorenham in #461
• 🔥 remove the codemods by @jorenham in #462
• 🔧 fix the add-trailing-comma hook to also work on .pyi by @jorenham in #463
• 🔥 remove pre-commit by @jorenham in #469

New Contributors

• @KiSchnelle made their first contribution in #470
• @ev-br made their first contribution in #478

Full Changelog: v1.15.2.0...v1.15.2.1

v1.15.2.0

⭐ Highlights

• The repo has been transferred from jorenham/scipy-stubs to scipy/scipy-stubs by @rgommers
• Support for SciPy 1.15.2 and NumPy 2.2.3

What's Changed

✨ Improvements

• special: improved dtype handling in the "basic" functions by @jorenham in #419
• signal: generic lti, and dlti types by @jorenham in #425
• optimize: improved linprog and milp annotations by @jorenham in #427
• optimize: method-specific overloads and options for minimize_scalar by @jorenham in #428
• optimize: Generic InverseJacobian and related types by @jorenham in #429
• signal: improvements related to czt and zoom_fft by @jorenham in #430
• sparse: generic sparray by @jorenham in #435
• stats.zmap: accept complex input and basic shape-typing by @jorenham in #457

🐛 Fixes

• stats: fix rv_discrete sample constructor by @jorenham in #418
• signal: add plain number option for find_peaks argument prominence by @drewsilcock in #448
• stats: make optimizer argument optional in rv_continuous.fit() by @drewsilcock in #451

📝 Documentation

• Use SciPy's Code of Conduct by @jorenham in #439
• Streamlined contributing guide by @jorenham in #445

⬆️ Dependencies

• bump sp-repo-review to 2025.1.22 by @jorenham in #426
• bump basedpyright to 1.27.1, ruff to 0.9.6 by @jorenham in #454
• Optype 0.9.1 by @jorenham in #449
• NumPy 2.2.3 by @jorenham in #455
• SciPy 1.15.2 by @jorenham in #456

Other Changes

• 🔖 prepare for further development by @jorenham in #415
• 🎨 optype 0.9 simplifications by @jorenham in #424
• 🔨 more tox, less poe by @jorenham in #431
• ⬆️ bump ruff to 0.9.3 by @jorenham in #432
• 🎨 sparse: prefer optype.numpy over scipy._typing by @jorenham in #434
• 💸 remove the sponsor button by @jorenham in #437
• 🔧 scipy.org as homepage and add scipy-dev as maintainer by @jorenham in #440
• 🔧 .toml auto-format config for vscode users by @jorenham in #444
• 🚚 change jorenham/scipy-stubs refs to scipy/scipy-stubs by @jorenham in #446
• 🔨 use uv as source of python versions by @jorenham in #452

New Contributors

• @drewsilcock made their first contribution in #448

Full Changelog: v1.15.1.0...v1.15.2.0

v1.15.1.0 (scipy 1.15.1)

This is the first scipy-stubs release that supports SciPy 1.15.1.

What's Changed

Improvements

• 🎨 remove outdated # TODO's by @jorenham in #408
• ✨ spatial: generic [c]KDTree and other improvements by @jorenham in #413

Dependencies

• ⬆️ bump ruff to 0.8.5 by @jorenham in #396
• ⬆️ bump scipy to 1.15.0 (final) by @jorenham in #409
• ⬆️ bump dev deps to latest by @jorenham in #410
• ⬆️ bump ruff to 0.9.0 by @jorenham in #412
• 📌 SciPy 1.15.1 by @jorenham in #414

Other Changes

• 🎨 fix and run the FixTrailingComma codemod by @jorenham in #397

Full Changelog: v1.15.0.0...v1.15.1.0

v1.15.0.0 (scipy 1.15.0)

This is the first scipy-stubs release that supports SciPy 1.15.0.

What's Changed

SciPy 1.15.0 compatability

• ⬆️ increase the minimum scipy version to 1.15.0rc1 by @jorenham in #327
• 🚧 temporarily ignore all stubtest errors by @jorenham in #326
• ⚰️ misc: remove all code by @jorenham in #328
• ✨ differentiate: preliminary (yet complete) stubs by @jorenham in #329
• 🏷️ _typing: SPEC-7 compliant (private) ToRNG type alias (previously Seed) by @jorenham in #333
• 👽️ fft[pack]: 1.15.0 support by @jorenham in #336
• 👽️ integrate: 1.15.0 support by @jorenham in #337
• 👽️ interpolate: 1.15.0 support by @jorenham in #338
• 👽️ constants: 1.15.0 support by @jorenham in #330
• 👽️ _lib: 1.15.0 support by @jorenham in #331
• 👽️ cluster: 1.15.0 support by @jorenham in #334
• 👽️ io: 1.15.0 support by @jorenham in #341
• 👽️ linalg.interpolative: 1.15.0 support by @jorenham in #343
• 👽️ linalg: 1.15.0 support by @jorenham in #344
• 👽️ version: 1.15.0 remove commit_count by @jorenham in #345
• 👽️ ndimage: 1.15.0 support by @jorenham in #346
• 👽️ optimize: SPEC 7 & scipy 1.15.0 support by @jorenham in #347
• 👽️ sparse: 1.15.0 support by @jorenham in #349
• 👽️ special: 1.15.0 support by @jorenham in #351
• 👽️ signal: 1.15.0 support by @jorenham in #348
• 👽️ spatial: 1.15.0 support and many improvements by @jorenham in #350
• 👽️ stats: 1.15.0 support and several improvements by @jorenham in #353
• 👽️ sparse: accept axes: tuple[int[:], int[:]] in coo_array.tensordot by @jorenham in #357
• 👽️ stats: 1.15.0 new distribution infrastructure by @jorenham in #360
• 💬 interpolate: fix scipy version in interpnd deprecation warning messages by @jorenham in #366
• 💬 special: mention the deprecation version in lpn and [c]lpmn by @jorenham in #367
• 🗑️ interpolate: deprecate the dfitpack functions, fix & improve the sp* functions by @jorenham in #372
• 📌 pin scipy<1.16 by @jorenham in #371
• 👽️ scipy==1.15.0rc2 by @jorenham in #376

Improvements

• ✨ more precise __config__ and __init__ annotations by @jorenham in #362
• ✨ _lib: improved annotations by @jorenham in #364
• ✨ special: several improved ufuncs by @jorenham in #374
• ✨ special: ufunc improvements by @jorenham in #375
• ✨ signal: complete _filter_design, fix and improve _czt by @jorenham in #378
• ✨ signal: complete _signaltools by @jorenham in #379
• ✨ signal: complete _fir_filter_design.* en max_len_seq by @jorenham in #384
• ✨ signal: complete _ltisys.* by @jorenham in #385
• 🎨 optype 0.8 simplifications by @jorenham in #382
• 🎨 stats: simplify _distn_infrastructure with optype 0.8 by @jorenham in #387
• 🎨 _typing: DRY Falsy and Truthy type aliases by @jorenham in #390

Fixes

• 🐛 sparse: add the missing sparse array/matrix dunder methods by @jorenham in #391

Documentation

• 📝 several README.md improvements by @jorenham in #365
• 📝 better describe the supported numpy versions by @jorenham in #368

Dependencies

• 📌 upgrade optype to 0.8.0 by @jorenham in #380
• ⬆️ update pre-commit hooks by @pre-commit-ci in #339
• ⬆️ bump basedpyright to 1.23.1 by @jorenham in #356
• ⬆️ bump ruff to 0.8.4 and disallow mdformat==0.7.20 by @jorenham in #358
• ⬆️ bump mdformat to 0.7.21 by @jorenham in #359
• ⬆️ bump basedmypy to 2.9.0 by @jorenham in #393
• Bump astral-sh/setup-uv from 4 to 5 in the actions group by @dependabot in #361
• Bump DavidAnson/markdownlint-cli2-action from 18 to 19 in the actions group by @dependabot in #386

Other Changes

• 🧑‍💻 remove mypy from pre-commit (too slow) by @jorenham in #355
• 👷 increase the timeout of the CI typetest step by @jorenham in #342
• 👷 PyPI trusted publishing workflow by @jorenham in #383
• 👷 exclude unsupported test matrix version combinations by @jorenham in #388
• 🔧 configure pre-commit ci to use gitmoji style by @jorenham in #340
• 🔧 de-duplicate the scipy version requirements by @jorenham in #389
• 🔧 regex'd .mypyignore by @jorenham in #335
• 🔧 .mypyignore cleanup by @jorenham in #392

Full Changelog: v1.14.1.6...v1.15.0.0

v1.14.1.6

Highlights

• scipy.interpolate: The public API is now 100% annotated (by humans)
• scipy.sparse: The public API is now 100% annotated (by humans)
  • the sparse array- and matrix- types optionally accept a generic type parameter for their scalar type, e.g. coo_array[np.float64].
  • {coo, csr, dok}_array additionally accept a 2nd optional type parameter for their 1-d (tuple[int]) or 2-d (tuple[int, int]) shape-type.
• fixed some typing issues on older numpy versions

What's Changed

Improvements

• 🎨 interpolate: remove duplicated function and classes by @jorenham in #267
• ✨ interpolate: stub _dfitpack by @jorenham in #270
• ✨ spatial: improve _qhull by @jorenham in #271
• ✨ interpolate: complete and improve interpnd by @jorenham in #272
• ✨ interpolate: complete _bsplines and _ndbspline by @jorenham in #273
• ✨ interpolate: complete RBFInterpolator and Rbf by @jorenham in #274
• ✨ interpolate: complete _fitpack_impl (1/2) by @jorenham in #275
• ✨ interpolate: complete _fitpack_impl (2/2) by @jorenham in #276
• ✨ interpolate: complete _fitpack2 by @jorenham in #277
• ✨ interpolate: complete _polyint by @jorenham in #279
• ✨ signal: complete frequency response functions in _filter_design by @jorenham in #287
• 🎨 sparse: code cleanup by @jorenham in #294
• ✨ sparse.linalg: complete the LinearOperators in _interface by @jorenham in #295
• 🎨 sparse explicit imports in __init__ by @jorenham in #296
• ✨ sparse.linalg: complete the matrix operation functions by @jorenham in #298
• ✨ sparse.linalg: complete the matrix norm functions by @jorenham in #300
• ✨ sparse.linalg: complete LaplacianNd & accept in dtypes in LinearOperator by @jorenham in #301
• ✨ sparse.linalg: complete _dsolve by @jorenham in #302
• ✨ sparse.linalg: complete _eigen.arpack by @jorenham in #303
• ✨ sparse.linalg: complete lobpcg by @jorenham in #305
• ✨ sparse.linalg: complete svds by @jorenham in #306
• ✨ sparse: complete the sparse arrays and matrices by @jorenham in #307
• ✨ sparse: complete eye[_array] by @jorenham in #309
• ✨ sparse: complete identity by @jorenham in #310
• ✨ sparse: complete kron[sum] by @jorenham in #311
• ✨ sparse: complete {h,v}stack by @jorenham in #313
• ✨ sparse: complete [sp]diags[_array] by @jorenham in #314
• ✨ sparse: complete {load,save}_npz by @jorenham in #315
• ✨ sparse: complete _sputils.* (private) by @jorenham in #316
• ✨ sparse: complete _spfuncs.* (private) by @jorenham in #317
• ✨ sparse: complete find and tri{l,u} by @jorenham in #318
• ✨ sparse: complete rand[om[_array]] by @jorenham in #319
• ✨ sparse: complete bmat and block_{array,diag} by @jorenham in #320
• ✨ sparse: adopt the new generic sparse arrays in linalg and csgraph by @jorenham in #321
• 🎨 sparse: various linalg.LinearOperator tweaks and improvements by @jorenham in #322
• ✨ sparse: complete linalg.ls{m,q}r by @jorenham in #323
• ✨ sparse: complete linalg._isolve.utils.* (private) by @jorenham in #324
• ✨ sparse: complete the remaining sparse.linalg functions by @jorenham in #325

Fixes

• ✨ interpolate: complete _rgi by @jorenham in #278
• 🐛 fix typing errors on numpy<1.24 by @jorenham in #292
• ⬆️ bump basedmypy to 2.8.1 by @jorenham in #293
• 👽️ work around several mypy bugs with numpy>=2.2.0 by @jorenham in #288

Documentation

• 🌕 interpolate: mark as 100% annotated by @jorenham in #280
• 📝 correct the documented lowest supported numpy version by @jorenham in #281
• 📝 installation instructions for humans by @jorenham in #282
• 🎤 add podcast to README.md by @jorenham in #299
• 📝 update the coverage progress for scipy.sparse by @jorenham in #304
• 📝 mention gitmoji in CONTRIBUTING.md by @jorenham in #308

Dependencies

• ⬆️ bump basedpyright to 1.22.1 and basedmypy to 2.8.0 by @jorenham in #255
• dynamic minimum requirements by @miloth in #259
• ⬆️ bump ruff to 0.8.2 by @jorenham in #265
• ⬆️ bump the dev deps by @jorenham in #312

Other Changes

• 🧑‍💻 uv-compatible mypy vscode config by @jorenham in #256
• 💡 add a comment to prevent breaking the #installation permalink by @jorenham in #266
• 🔧 fix caching issue in poe {mypy,stubtest,typetest} by @jorenham in #269
• 🔧 stricter basedpyright config by @jorenham in #283
• 👷 include .mypyignore in sdist build by @jorenham in #284
• ➖ remove broken mdformat-pyproject by @miloth in #262
• 💚 towards fixing the typetest-oldest-supported-numpy CI job by @jorenham in #289
• 👷 fix and simplify the CI workflow by @jorenham in #290
• Dynamic testing matrix for combinations of python and numpy versions. by @wolph in #286

New Contributors

• @wolph made their first contribution in #286

AI generated Podcast

SciPy.Type.Hints.mp4

Full Changelog: v1.14.1.5...v1.14.1.6

v1.14.1.5

Highlights

• scipy-stubs is now also available on conda-forge, thanks to @lucascolley
• scipy-stubs now uses uv for package- and project-management (previously poetry)
• scipy.linalg has been significantly improved
  • functions with either real or complex output now use overloads for more accurate return types
  • almost all functions are now aware of the shape-type of input array-likes and output ndarrays.
• scipy.optimize is now 100% annotated (by hand)
• scipy.signal now accurately type-hints the "waveform" functions sawtooth, square, chirp, sweep_poly, unit_impulse and gausspulse
• scipy.special now accurately type-hints 80% of its ufuncs using type-signature-specific overloaded protocols.

What's Changed

Improvements

• *: switch to the shape-typed optype.numpy array-likes by @jorenham in #192
• special: type-signature-specific annotations for 80% of the ufuncs by @jorenham in #189
• special: update the development progress phase to 🌔 by @jorenham in #194
• signal: Add type stubs to _waveforms.pyi. by @pavyamsiri in #195
• datasets: Improved type-hints with TypedDicts and Literals by @jorenham in #203
• prefer optype.numpy for annotating arrays by @jorenham in #204
• optimize: complete _constraints and _trustregion_constr.* by @jorenham in #205
• optimize: complete the root-finding functions by @jorenham in #207
• optimize: complete fixed_point by @jorenham in #208
• optimize: complete curve_fit, leastsq, and fsolve by @jorenham in #209
• optimize: complete HessianUpdateStrategy, BFGS, and SR1 by @jorenham in #210
• optimize: complete basinhopping by @jorenham in #211
• optimize: fix and complete direct by @jorenham in #212
• optimize: complete _differentiable_functions (private) by @jorenham in #214
• optimize: complete _lsq (private subpackage) by @jorenham in #215
• Codemod for pos-only self parameters by @jorenham in #218
• optimize: complete differential_evolution by @jorenham in #219
• optimize: complete dual_annealing by @jorenham in #220
• optimize: fix and improve minimize and basinhopping by @jorenham in #221
• optimize: complete shgo, improve basinhopping by @jorenham in #222
• optimize: complete line_search by @jorenham in #223
• optimize: complete _dcsrch (private) by @jorenham in #224
• optimize: complete linear_sum_assignment by @jorenham in #225
• optimize: complete quadratic_assignment by @jorenham in #227
• signal: Replace the final Untyped in _short_time_fft.pyi. by @pavyamsiri in #226
• optimize: complete _shgo_lib (private) by @jorenham in #228
• optimize: complete _trlib and _trustregion* (private) by @jorenham in #236
• optimize: complete _numdiff (private) by @jorenham in #237
• optimize: complete _group_columns (private) by @jorenham in #238
• optimize: remove _remove_redundancy (private) by @jorenham in #239
• optimize: complete fmin_cobyla by @jorenham in #240
• optimize: complete fmin_l_bfgs_b and LbfgsInvHessProduct by @jorenham in #241
• optimize: complete fmin_slsqp by @jorenham in #242
• linalg._basic: add overloads for shapes and dtypes by @jorenham in #244
• linalg._decomp*: add overloads for shapes and dtypes by @jorenham in #245
• linalg: improved special matrix constructor signatures by @jorenham in #246
• linalg: improved norm signature by @jorenham in #247
• linalg: improve the shape- & scalar-type overloads of various functions by @jorenham in #252

Documentation

• add docs to the codegen/ directory by @jorenham in #200
• improved README.md by @jorenham in #206
• document that scipy-stubs is installable with conda by @jorenham in #230
• 📝 update the scipy coverage phase table by @jorenham in #243
• Add a highlights section in README.md by @jorenham in #249
• update the README.md phase table to better reflect the Untyped counts by @jorenham in #253
• prevent lock update with tools and uv setup guide by @miloth in #254

Dependencies

• bump mdformat and mdformat-pyproject by @jorenham in #187
• bump ruff to 0.8.0 by @jorenham in #191
• bump basedpyright to 1.22.0 by @jorenham in #193
• bump optype to 0.7.2 by @jorenham in #197
• Bump poethepoet from 0.31.0 to 0.31.1 in the actions group by @dependabot in #202
• configure (based)pyright to ignore .venv when opened by @jorenham in #213
• bump ruff to 0.8.1 by @jorenham in #229
• bump optype to 0.7.3 by @jorenham in #232

Other Changes

• prepare for further development by @jorenham in #186
• move libcst config out of the root dir by @jorenham in #188
• optimize: remove private testing module _tstutils by @jorenham in #196
• custom auto-generated release notes by @jorenham in #199
• enable all ruff rules by default by @jorenham in #201
• restore the libcst codemod config by @jorenham in #217
• test with the oldest supported numpy version in the CI by @jorenham in #231
• use the new 0-d optype array aliases to simplify the code by @jorenham in #234
• use the new optype strict-shape array-likes by @jorenham in #235
• security policy update by @jorenham in #251
• use uv as package manager by @miloth in #174

New Contributors

• @miloth made their first contribution in #174

Full Changelog: v1.14.1.4...v1.14.1.5