SciPy 1.8.0 Release Notes#

SciPy 1.8.0 is the culmination of 6 months of hard work. It contains many new features, numerous bug-fixes, improved test coverage and better documentation. There have been a number of deprecations and API changes in this release, which are documented below. All users are encouraged to upgrade to this release, as there are a large number of bug-fixes and optimizations. Before upgrading, we recommend that users check that their own code does not use deprecated SciPy functionality (to do so, run your code with python -Wd and check for DeprecationWarning s). Our development attention will now shift to bug-fix releases on the 1.8.x branch, and on adding new features on the master branch.

This release requires Python 3.8+ and NumPy 1.17.3 or greater.

For running on PyPy, PyPy3 6.0+ is required.

Highlights of this release#

  • A sparse array API has been added for early testing and feedback; this work is ongoing, and users should expect minor API refinements over the next few releases.

  • The sparse SVD library PROPACK is now vendored with SciPy, and an interface is exposed via scipy.sparse.svds with solver='PROPACK'. It is currently default-off due to potential issues on Windows that we aim to resolve in the next release, but can be optionally enabled at runtime for friendly testing with an environment variable setting of USE_PROPACK=1.

  • A new scipy.stats.sampling submodule that leverages the UNU.RAN C library to sample from arbitrary univariate non-uniform continuous and discrete distributions

  • All namespaces that were private but happened to miss underscores in their names have been deprecated.

New features#

scipy.fft improvements#

Added an orthogonalize=None parameter to the real transforms in scipy.fft which controls whether the modified definition of DCT/DST is used without changing the overall scaling.

scipy.fft backend registration is now smoother, operating with a single registration call and no longer requiring a context manager.

scipy.integrate improvements#

scipy.integrate.quad_vec introduces a new optional keyword-only argument, args. args takes in a tuple of extra arguments if any (default is args=()), which is then internally used to pass into the callable function (needing these extra arguments) which we wish to integrate.

scipy.interpolate improvements#

scipy.interpolate.BSpline has a new method, design_matrix, which constructs a design matrix of b-splines in the sparse CSR format.

A new method from_cubic in BSpline class allows to convert a CubicSpline object to BSpline object.

scipy.linalg improvements#

scipy.linalg gained three new public array structure investigation functions. scipy.linalg.bandwidth returns information about the bandedness of an array and can be used to test for triangular structure discovery, while scipy.linalg.issymmetric and scipy.linalg.ishermitian test the array for exact and approximate symmetric/Hermitian structure.

scipy.optimize improvements#

scipy.optimize.check_grad introduces two new optional keyword only arguments, direction and seed. direction can take values, 'all' (default), in which case all the one hot direction vectors will be used for verifying the input analytical gradient function and 'random', in which case a random direction vector will be used for the same purpose. seed (default is None) can be used for reproducing the return value of check_grad function. It will be used only when direction='random'.

The scipy.optimize.minimize TNC method has been rewritten to use Cython bindings. This also fixes an issue with the callback altering the state of the optimization.

Added optional parameters target_accept_rate and stepwise_factor for adapative step size adjustment in basinhopping.

The epsilon argument to approx_fprime is now optional so that it may have a default value consistent with most other functions in scipy.optimize.

scipy.signal improvements#

Add analog argument, default False, to zpk2sos, and add new pairing option 'minimal' to construct analog and minimal discrete SOS arrays. tf2sos uses zpk2sos; add analog argument here as well, and pass it on to zpk2sos.

savgol_coeffs and savgol_filter now work for even window lengths.

Added the Chirp Z-transform and Zoom FFT available as scipy.signal.CZT and scipy.signal.ZoomFFT.

scipy.sparse improvements#

An array API has been added for early testing and feedback; this work is ongoing, and users should expect minor API refinements over the next few releases. Please refer to the scipy.sparse docstring for more information.

maximum_flow introduces optional keyword only argument, method which accepts either, 'edmonds-karp' (Edmonds Karp algorithm) or 'dinic' (Dinic’s algorithm). Moreover, 'dinic' is used as default value for method which means that Dinic’s algorithm is used for computing maximum flow unless specified. See, the comparison between the supported algorithms in this comment.

Parameters atol, btol now default to 1e-6 in scipy.sparse.linalg.lsmr to match with default values in scipy.sparse.linalg.lsqr.

Add the Transpose-Free Quasi-Minimal Residual algorithm (TFQMR) for general nonsingular non-Hermitian linear systems in scipy.sparse.linalg.tfqmr.

The sparse SVD library PROPACK is now vendored with SciPy, and an interface is exposed via scipy.sparse.svds with solver='PROPACK'. For some problems, this may be faster and/or more accurate than the default, ARPACK. PROPACK functionality is currently opt-in–you must specify USE_PROPACK=1 at runtime to use it due to potential issues on Windows that we aim to resolve in the next release.

sparse.linalg iterative solvers now have a nonzero initial guess option, which may be specified as x0 = 'Mb'.

The trace method has been added for sparse matrices.

scipy.spatial improvements#

scipy.spatial.transform.Rotation now supports item assignment and has a new concatenate method.

Add scipy.spatial.distance.kulczynski1 in favour of scipy.spatial.distance.kulsinski which will be deprecated in the next release.

scipy.spatial.distance.minkowski now also supports 0<p<1.

scipy.special improvements#

The new function scipy.special.log_expit computes the logarithm of the logistic sigmoid function. The function is formulated to provide accurate results for large positive and negative inputs, so it avoids the problems that would occur in the naive implementation log(expit(x)).

A suite of five new functions for elliptic integrals: scipy.special.ellipr{c,d,f,g,j}. These are the Carlson symmetric elliptic integrals, which have computational advantages over the classical Legendre integrals. Previous versions included some elliptic integrals from the Cephes library (scipy.special.ellip{k,km1,kinc,e,einc}) but was missing the integral of third kind (Legendre’s Pi), which can be evaluated using the new Carlson functions. The new Carlson elliptic integral functions can be evaluated in the complex plane, whereas the Cephes library’s functions are only defined for real inputs.

Several defects in scipy.special.hyp2f1 have been corrected. Approximately correct values are now returned for z near exp(+-i*pi/3), fixing #8054. Evaluation for such z is now calculated through a series derived by López and Temme (2013) that converges in these regions. In addition, degenerate cases with one or more of a, b, and/or c a non-positive integer are now handled in a manner consistent with mpmath’s hyp2f1 implementation, which fixes #7340. These fixes were made as part of an effort to rewrite the Fortran 77 implementation of hyp2f1 in Cython piece by piece. This rewriting is now roughly 50% complete.

scipy.stats improvements#

scipy.stats.qmc.LatinHypercube introduces two new optional keyword-only arguments, optimization and strength. optimization is either None or random-cd. In the latter, random permutations are performed to improve the centered discrepancy. strength is either 1 or 2. 1 corresponds to the classical LHS while 2 has better sub-projection properties. This construction is referred to as an orthogonal array based LHS of strength 2. In both cases, the output is still a LHS.

scipy.stats.qmc.Halton is faster as the underlying Van der Corput sequence was ported to Cython.

The alternative parameter was added to the kendalltau and somersd functions to allow one-sided hypothesis testing. Similarly, the masked versions of skewtest, kurtosistest, ttest_1samp, ttest_ind, and ttest_rel now also have an alternative parameter.

Add scipy.stats.gzscore to calculate the geometrical z score.

Random variate generators to sample from arbitrary univariate non-uniform continuous and discrete distributions have been added to the new scipy.stats.sampling submodule. Implementations of a C library UNU.RAN are used for performance. The generators added are:

  • TransformedDensityRejection

  • DiscreteAliasUrn

  • NumericalInversePolynomial

  • DiscreteGuideTable

  • SimpleRatioUniforms

The binned_statistic set of functions now have improved performance for the std, min, max, and median statistic calculations.

somersd and _tau_b now have faster Pythran-based implementations.

Some general efficiency improvements to handling of nan values in several stats functions.

Added the Tukey-Kramer test as scipy.stats.tukey_hsd.

Improved performance of scipy.stats.argus rvs method.

Added the parameter keepdims to scipy.stats.variation and prevent the undesirable return of a masked array from the function in some cases.

permutation_test performs an exact or randomized permutation test of a given statistic on provided data.

Deprecated features#

Clear split between public and private API#

SciPy has always documented what its public API consisted of in its API reference docs, however there never was a clear split between public and private namespaces in the code base. In this release, all namespaces that were private but happened to miss underscores in their names have been deprecated. These include (as examples, there are many more):

  • scipy.signal.spline

  • scipy.ndimage.filters

  • scipy.ndimage.fourier

  • scipy.ndimage.measurements

  • scipy.ndimage.morphology

  • scipy.ndimage.interpolation

  • scipy.sparse.linalg.solve

  • scipy.sparse.linalg.eigen

  • scipy.sparse.linalg.isolve

All functions and other objects in these namespaces that were meant to be public are accessible from their respective public namespace (e.g. scipy.signal). The design principle is that any public object must be accessible from a single namespace only; there are a few exceptions, mostly for historical reasons (e.g., stats and stats.distributions overlap). For other libraries aiming to provide a SciPy-compatible API, it is now unambiguous what namespace structure to follow. See gh-14360 for more details.

Other deprecations#

NumericalInverseHermite has been deprecated from scipy.stats and moved to the scipy.stats.sampling submodule. It now uses the C implementation of the UNU.RAN library so the result of methods like ppf may vary slightly. Parameter tol has been deprecated and renamed to u_resolution. The parameter max_intervals has also been deprecated and will be removed in a future release of SciPy.

Backwards incompatible changes#

  • SciPy has raised the minimum compiler versions to GCC 6.3 on linux and VS2019 on windows. In particular, this means that SciPy may now use C99 and C++14 features. For more details see here.

  • The result for empty bins for scipy.stats.binned_statistic with the builtin 'std' metric is now nan, for consistency with np.std.

  • The function scipy.spatial.distance.wminkowski has been removed. To achieve the same results as before, please use the minkowski distance function with the (optional) w= keyword-argument for the given weight.

Other changes#

Some Fortran 77 code was modernized to be compatible with NAG’s nagfor Fortran compiler (see, e.g., PR 13229).

threadpoolctl may now be used by our test suite to substantially improve the efficiency of parallel test suite runs.

Authors#

  • @endolith

  • adamadanandy +

  • akeemlh +

  • Anton Akhmerov

  • Marvin Albert +

  • alegresor +

  • Andrew Annex +

  • Pantelis Antonoudiou +

  • Ross Barnowski +

  • Christoph Baumgarten

  • Stephen Becker +

  • Nickolai Belakovski

  • Peter Bell

  • berberto +

  • Georgii Bocharov +

  • Evgeni Burovski

  • Matthias Bussonnier

  • CJ Carey

  • Justin Charlong +

  • Hood Chatham +

  • Dennis Collaris +

  • David Cottrell +

  • cruyffturn +

  • da-woods +

  • Anirudh Dagar

  • Tiger Du +

  • Thomas Duvernay

  • Dani El-Ayyass +

  • Castedo Ellerman +

  • Donnie Erb +

  • Andreas Esders-Kopecky +

  • Livio F +

  • Isuru Fernando

  • Evelyn Fitzgerald +

  • Sara Fridovich-Keil +

  • Mark E Fuller +

  • Ralf Gommers

  • Kevin Richard Green +

  • guiweber +

  • Nitish Gupta +

  • h-vetinari

  • Matt Haberland

    1. Hariharan +

  • Charles Harris

  • Jonathan Helgert +

  • Trever Hines

  • Nadav Horesh

  • Ian Hunt-Isaak +

  • ich +

  • Itrimel +

  • Jan-Hendrik Müller +

  • Jebby993 +

  • Yikun Jiang +

  • Evan W Jones +

  • Nathaniel Jones +

  • Jeffrey Kelling +

  • Malik Idrees Hasan Khan +

  • Paul Kienzle

  • Sergey B Kirpichev

  • Kadatatlu Kishore +

  • Andrew Knyazev

  • Ravin Kumar +

  • Peter Mahler Larsen

  • Eric Larson

  • Antony Lee

  • Gregory R. Lee

  • Tim Leslie

  • lezcano +

  • Xingyu Liu

  • Christian Lorentzen

  • Lorenzo +

  • Smit Lunagariya +

  • Lv101Magikarp +

  • Yair M +

  • Cong Ma

  • Lorenzo Maffioli +

  • majiang +

  • Brian McFee +

  • Nicholas McKibben

  • John Speed Meyers +

  • millivolt9 +

  • Jarrod Millman

  • Harsh Mishra +

  • Boaz Mohar +

  • naelsondouglas +

  • Andrew Nelson

  • Nico Schlömer

  • Thomas Nowotny +

  • nullptr +

  • Teddy Ort +

  • Nick Papior

  • ParticularMiner +

  • Dima Pasechnik

  • Tirth Patel

  • Matti Picus

  • Ilhan Polat

  • Adrian Price-Whelan +

  • Quentin Barthélemy +

  • Sundar R +

  • Judah Rand +

  • Tyler Reddy

  • Renal-Of-Loon +

  • Frederic Renner +

  • Pamphile Roy

  • Bharath Saiguhan +

  • Atsushi Sakai

  • Eric Schanet +

  • Sebastian Wallkötter

  • serge-sans-paille

  • Reshama Shaikh +

  • Namami Shanker

  • siddhantwahal +

  • Walter Simson +

  • Gagandeep Singh +

  • Leo C. Stein +

  • Albert Steppi

  • Kai Striega

  • Diana Sukhoverkhova

  • Søren Fuglede Jørgensen

  • Masayuki Takagi +

  • Mike Taves

  • Ben Thompson +

  • Bas van Beek

  • Jacob Vanderplas

  • Dhruv Vats +

    1. Vetinari +

  • Thomas Viehmann +

  • Pauli Virtanen

  • Vlad +

  • Arthur Volant

  • Samuel Wallan

  • Stefan van der Walt

  • Warren Weckesser

  • Josh Wilson

  • Haoyin Xu +

  • Rory Yorke

  • Egor Zemlyanoy

  • Gang Zhao +

  • 赵丰 (Zhao Feng) +

A total of 139 people contributed to this release. People with a “+” by their names contributed a patch for the first time. This list of names is automatically generated, and may not be fully complete.

Issues closed for 1.8.0#

  • #592: Statistics Review: variation (Trac #65)

  • #857: A Wrapper for PROPACK (Trac #330)

  • #2009: “Kulsinski” dissimilarity seems wrong (Trac #1484)

  • #2063: callback functions for COBYLA and TNC (Trac #1538)

  • #2358: ndimage.center_of_mass doesnt return all for all labelled objects…

  • #5668: Need zpk2sos for analog filters

  • #7340: SciPy Hypergeometric function hyp2f1 producing infinities

  • #8774: In `optimize.basinhopping`, the target acceptance rate should…

  • #10497: scipy.sparse.csc_matrix.toarray docstring is wrong

  • #10888: Check finite difference gradient approximation in a random direction

  • #10974: Non explicit error message in lobpcg

  • #11452: Normalisation requirement for `Wn` unclear in `scipy.signal.butter`

  • #11700: solve_ivp errors out instead of simply quitting after the solve…

  • #12006: newton: Shouldn’t it take a Jacobian for multivariate problems…

  • #12100: solve_ivp: custom t_eval list and the terminating event

  • #12106: `axis` option for `stats.tmean` do not appear to be working…

  • #12192: `scipy.stats.rv_continuous.moment` does not accept array input

  • #12502: Divide by zero in Jacobian numerical differentiation when equality…

  • #12981: SLSQP constrained minimization error in 1.5.2

  • #12999: Bug in scipy.stats.ks_2samp for two-sided auto and exact modes…

  • #13402: ENH: Faster Max Flow algorithm in scipy.sparse.csgraph

  • #13580: truncnorm gives incorrect means and variances

  • #13642: stats.truncnorm variance works incorrectly when input is an array.

  • #13659: Orthogonal Array for Latin hypercube in `scipy.stats.qmc`

  • #13737: brentq can overflow / underflow

  • #13745: different default atol, btol for lsqr, lsmr

  • #13898: Savitzky-Golay filter for even number data

  • #13902: Different solvers of `svds` return quite different results

  • #13922: Need Exception / Error for Incorrect and/or misleading analog…

  • #14122: Item assignement for spatial.transform.Rotation objects

  • #14140: Likely unnecessary invalid value warning from PchipInterpolator

  • #14152: zpk2sos not working correctly when butterworth band-pass filter…

  • #14165: scipy.optimize.minimize method=’Nelder-Mead’: ‘maxfev’ is not…

  • #14168: Missing “inverse” word in the multidimensional Discrete Cosine/Sine…

  • #14189: Incorrect shape handling in `scipy.stat.multivariate_t.rvs`…

  • #14190: Links in documentation of Dirichlet distribution are a mess

  • #14193: Implementation of scrambled Van der Corput sequence differs from…

  • #14217: Error in documentation for `scipy.stats.gaussian_kde.factor`

  • #14235: Should this be $y$ only, instead of $m_y$?

  • #14236: BUG: discrete isf is wrong at boundary if loc != 0

  • #14277: Broken reference in docstring of scipy.stats.power_divergence

  • #14324: BUG: scipy.stats.theilslopes intercept calculation can produce…

  • #14332: Strange output of `binned_statistic_dd` with `statistic=sum`

  • #14340: Initialize Rotation using list or array of Rotations

  • #14346: scipy.stats.rv_continuous.fit returns wrapper instead of fit…

  • #14360: Making clearer what namespaces are public by use of underscores

  • #14385: csgraph.maximum_flow can cause Python crash for large but very…

  • #14409: Lagrange polynomials and numpy Polynomials

  • #14412: Extra function arguments to `scipy.integrate.quad_vec`

  • #14416: Is the r-value outputted by scipy.stats.linregress always the…

  • #14420: BUG: RBFInterpolator fails when calling it with a slice of a…

  • #14425: Running tests in parallel is not any faster than without pytest-xdist…

  • #14445: BUG: out of bounds indexing issue in `prini.f`

  • #14482: Azure CI jobs do not set exit status for build stage correctly

  • #14491: MAINT: Replace np.rollaxis with np.moveaxis

  • #14501: runtests.py overrides `$PYTHONPATH`

  • #14514: linprog kwargs not recognised

  • #14529: CI: Azure pipelines don’t appear to be running

  • #14535: hess option does not work in minimize function

  • #14551: Cannot create Compressed sparse column matrix of shape N x N-2

  • #14568: `stats.norminvgauss` incorrect implementation?

  • #14585: DOC: toolchain updates and max Python

  • #14607: scipy.sparse.linalg.inv cannot take ndarray as argument despite…

  • #14608: BUG: scipy.stats.multivariate_t distribution math documentation

  • #14623: BUG: Error constructing sparse matrix with indices larger than…

  • #14654: DOC: Linux Devdocs workflow requires installing packages that…

  • #14680: BUG: misleading documentation in scipy.stats.entropy

  • #14683: DOC: OptimizeResult Notes are placed before attribute section,…

  • #14733: BUG: resample_poly does not preserve dtype

  • #14746: site.cfg: [ALL] or [DEFAULT]?

  • #14770: BUG: lpmn ref broken link

  • #14807: BUG: wrong weights of the 7-point gauss rule in QUADPACK: dqk15w.f

  • #14830: do CDF inversion methods have to be public?

  • #14859: BUG: constraint function is overwritten when equal bounds are…

  • #14873: ENH: get the driver used in scipy.linalg.eigh

  • #14879: BUG: TNC output is different if a callback is used.

  • #14891: DOC: `directed_hausdorff` expects 2D array despite docs stating…

  • #14910: `stats.contingency` not listed as public API

  • #14911: MAINT, DOC: CI failure for doc building

  • #14942: DOC: Ambiguous command instruction for running tests in Mac docs

  • #14968: Debug build CI job crashes on `stats._unuran` threading test

  • #14984: BUG: scipy.sparse.linalg.spsolve: runtime memory error caused…

  • #14987: ENH: The knot interval lookup for BSpline.design_matrix is inefficient

  • #15025: Might be j<=i+k?

  • #15033: BUG: scipy.fft.dct type I with norm = “ortho” leads to wrong…

  • #15051: BUG: test failures on aarch in wheel builder repo

  • #15064: MAINT: `interpolation` keyword is renamed to `method` in…

  • #15103: BUG: scipy.stats.chi.mean returns nan for large df due to use…

  • #15186: Fix use of `pytest.warns(None)` for pytest 7.0.0

  • #15206: BUG: Minor issue with suggestions in scipy.sparse DeprecationWarnings…

  • #15224: BUG: 0th power of sparse array/matrix always returns the identity…

  • #15228: BUG: bounded L-BFGS-B doesn’t work with a scalar.

  • #15254: BUG: `DeprecationWarning: distutils Version classes are deprecated`

  • #15267: Windows CI jobs have a build issue with Pythran 0.11

  • #15276: Boost and PROPACK git submodules are too easy to commit changes…

  • #15316: BUG: Failed to install scipy 1.7.x with pypy 3.7 in aarch64

  • #15339: BUG: `highs-ds` returns memoryviews instead of np.arrays for…

  • #15375: BUG: axis argument to scipy.stats.mode does not accept negative…

  • #15517: BUG: Link to mailing list seems broken

Pull requests for 1.8.0#

  • #4607: Add Chirp Z-transform, zoom FFT

  • #10504: ENH: Carlson symmetric elliptic integrals.

  • #11263: MAINT:optimize: Comply with user-specified rel_step

  • #11754: ENH: stats: Updates to `variation`.

  • #11954: ENH: improve ARGUS rv generation in scipy.stats

  • #12143: BUG: Correctly use `axis` in `scipy.stats.tmean`

  • #12146: DOC: add docs to explain behaviour of newton’s mehod on arrays

  • #12197: BUG: fix moments method to support arrays and list

  • #12889: MAINT: deal with cases in `minimize` for `(bounds.lb == bounds.ub).any()

  • #13002: ENH: stats: Tukey’s honestly significant difference test

  • #13096: BUG: optimize: alternative fix for minimize issues with lb==ub

  • #13143: MAINT: deal with cases in `minimize` for `(bounds.lb == bounds.ub).any()…

  • #13229: ENH: modernise some Fortran code, needed for nagfor compiler

  • #13312: ENH: stats: add `axis` and `nan_policy` parameters to functions…

  • #13347: CI: bump gcc from 4.8 to 5.x

  • #13392: MAINT: streamlined kwargs for minimizer in dual_annealing

  • #13419: BUG: Fix group delay singularity check

  • #13471: ENH: LHS based OptimalDesign (scipy.stats.qmc)

  • #13581: MAINT: stats: fix truncnorm stats with array shapes

  • #13839: MAINT: set same tolerance between LSMR and LSQR

  • #13864: Array scalar conversion deprecation

  • #13883: MAINT: move LSAP maximization handling into solver code

  • #13899: ENH: stats: add general permutation hypothesis test

  • #13921: BUG: optimize: fix max function call validation for `minimize`…

  • #13958: ENH: stats: add `alternative` to masked version of T-Tests

  • #13960: ENH: stats: add `alternative` to masked normality tests

  • #14007: BUG: Fix root bracketing logic in Brent’s method (issue #13737)

  • #14024: ENH: Add annotations for `scipy.spatial.cKDTree`

  • #14049: MAINT: Change special.orthogonal.orthopoly1d type hints to ArrayLike

  • #14132: DOC: badge with version of the doc in the navbar

  • #14144: REL: set version to 1.8.0.dev0

  • #14151: BLD: update pyproject.toml - add macOS M1, drop py36

  • #14153: BUG: stats: Implementing boost’s hypergeometric distribution…

  • #14160: ENH: sparse.linalg: Add TFQMR algorithm for non-Hermitian sparse…

  • #14163: BENCH: add benchmark for energy_distance and wasserstein_distance

  • #14173: BUG: Fixed an issue wherein `geometric_slerp` would return…

  • #14174: ENH: Add annotations to `scipy.spatial.geometric_slerp`

  • #14183: DOC: add examples/ update mstats doc of pearsonr in scipy.stats

  • #14186: TST, MAINT: hausdorff test cleanups

  • #14187: DOC: interpolate: rbf has kwargs too.

  • #14191: MAINT:TST:linalg modernize the test assertions

  • #14192: BUG: stats: fix shape handing in multivariate_t.rvs

  • #14197: CI: azure: Fix handling of ‘skip azp’.

  • #14200: DOC: Remove link to alpha in scipy.stats.dirichlet

  • #14201: TST: cleanup in lsqr and lsmr tests

  • #14204: Improve error message for index dimension

  • #14208: MAINT: add invalid=’ignore’ to np.errstate block in PchipInterpolator

  • #14209: ENH: stats: kendalltau: add alternative parameter

  • #14210: BUG: Fix Nelder-Mead logic when using a non-1D x0 and adapative

  • #14211: Fixed doc for gaussian_kde (kde.factor description)

  • #14213: ENH: stats: somersd: add alternative parameter

  • #14214: ENH: Improve the `scipy.spatial.qhull` annotations

  • #14215: ENH: stats: Integrate library UNU.RAN in `scipy.stats` [GSoC…

  • #14218: DOC: clarify `ndimage.center_of_mass` docstring

  • #14219: ENH: sparse.linalg: Use the faster “sqrt” from “math” and be…

  • #14222: MAINT: stats: remove unused ‘type: ignore’ comment

  • #14224: MAINT: Modify to use new random API in benchmarks

  • #14225: MAINT: fix missing LowLevelCallable in `dir(scipy)`

  • #14226: BLD: fix warning for missing dependency, and dev version number

  • #14227: MAINT: fix maybe-uninitialized warnings in lbfgbf.f

  • #14228: BENCH: add more benchmarks for inferential statistics tests

  • #14237: Removes unused variable

  • #14240: ENH: sparse.linalg: Normalize type descriptions

  • #14242: BUG: stats: fix discrete `.isf` to work at boundaries when…

  • #14250: Error in parameter checking in cdfbin.f

  • #14254: BUG: Fixed an issue wherein `SphericalVoronoi` could raise…

  • #14255: BUG: Numerical stability for large N BarycentricInterpolator

  • #14257: MAINT: Fixed deprecated API calls in scipy.optimize

  • #14258: DOC: fix stats.pearsonr example that was failing in CI

  • #14259: CI: pin mypy to 0.902 and fix one CI failure

  • #14260: BLD: optimize: fix some warnings in moduleTNC and minpack.h

  • #14261: BLD: fix include order and build warnings for `optimize/_trlib`

  • #14263: DOC: forward port 1.7.0 relnotes

  • #14268: MAINT: Replaced direct field access in PyArrayObject* with wrapper…

  • #14274: MAINT: more scalar array conversion fixes for optimize

  • #14275: MAINT: Update vendored uarray, required for auto-dispatching

  • #14278: MAINT: two small fixes for implicit scalar-array-conversions

  • #14281: ENH: Annotate the array dtypes of `scipy.spatial.qhull`

  • #14285: DEV: remove scikit-umfpack from environment.yml

  • #14287: TST: Add testing for hyp2f1 for complex values in anticipation…

  • #14291: TST: split combined LSAP input validation tests up

  • #14293: MAINT: remove the last deprecated `PyEval_*` usages

  • #14294: ENH: Annotate array dtypes in `scipy.spatial.ckdtree` and `distance`

  • #14295: MAINT: move LSAP input validation into lsap_module

  • #14297: DOC: Make code block an Item List

  • #14301: MAINT: fix the last build warning in `optimize/_trlib/`

  • #14302: BLD: fix build warnings for `stats/biasedurn`

  • #14305: MAINT: silence warning in odepackmodule.c

  • #14308: ENH: use Pythran to speedup somersd and _tau_b

  • #14309: BLD: fix build warnings for scipy.special

  • #14310: ENH: make epsilon optional in optimize.approx_fprime.

  • #14311: MAINT: Corrected NumPy API usage in scipy.spatial

  • #14312: ENH: Using random directional derivative to check grad

  • #14326: MAINT: Removed redifinition of trace1 in spatial/qhull

  • #14328: MAINT: _lib: add __dealloc__ to MessageStream

  • #14331: ENH: Complement `trace` method of sparse matrices like `csr_matrix/csc_matrix/coo_matrix`

  • #14338: BUG: fix `stats.binned_statistic_dd` issue with values close…

  • #14339: TST: fix `sparse.linalg.spsolve` test with singular input

  • #14341: MAINT: Add missing parenthesis in _nnls.py

  • #14342: ENH: make `savgol_coeffs`, `savgol_filter` work for even…

  • #14344: ENH: scipy.interpolate b-splines (design_matrix)

  • #14350: MAINT: make fit method of rv_continuous pickleable

  • #14358: ENH: Dinic’s algorithm for maximum_flow

  • #14359: ENH: Set fft backend with try_last=True

  • #14362: Use list comprehension

  • #14367: BUG: Check for NULL pointer in `memmove`

  • #14377: Fix behavior of binary morphology with output=input when iterations=1

  • #14378: MAINT: Removing deprecated NumPy C API from `interpolate`

  • #14380: ENH: Fixed intercept computation in theilslopes

  • #14381: BENCH: add benchmark for somersd

  • #14387: MAINT: Removed deprecated NumPy C api from `sparse`

  • #14392: BUG/ENH: rework maximum flow preprocessing

  • #14393: CI: Lint checks failures are reporting success

  • #14403: Fix off by one error in doc string.

  • #14404: DOC: docstring fix for default of n param of interpolate.pade

  • #14406: MAINT: Use numpy_nodepr_api in `spatial`

  • #14411: MAINT: minor cleanups in usage of `compute_uv` keyword of `svd`

  • #14413: DOC:interpolate: Fix the docstring example of “lagrange”

  • #14419: DEP: deprecate private but non-underscored `signal.spline`…

  • #14422: MAINT: csgraph: change Dinic algorithm to iterative implementation

  • #14423: CI: remove printing of skipped and xfailed tests from Azure test…

  • #14426: ENH: Add args argument for callable in quad_vec

  • #14427: MAINT: extra pythran annotation for i686 support

  • #14432: BUG/ENH: more stable recursion for 2-sample ks test exact p-values

  • #14433: ENH: add PROPACK wrapper for improved sparse SVD

  • #14440: MAINT: stats: silence mypy complaints

  • #14441: ENH: TST: add a threadpoolctl hook to limit OpenBLAS parallelism

  • #14442: MAINT: Fix uninitialized warnings in `sparse/linalg/dsolve`

  • #14447: MAINT: rename scipy.ndimage modules

  • #14449: ENH: Cythonize van der corput

  • #14454: MAINT: Begin translation of hyp2f1 for complex numbers into Cython

  • #14456: CI: Lint with flake8 instead of pyflakes + pycodestyle

  • #14458: DOC: clarify meaning of rvalue in stats.linregress

  • #14459: MAINT: Fix uninitialized warnings in `interpolate` and `cluster`

  • #14463: Fix typo in doc overview: “pandas” to “SciPy”

  • #14474: DEP: Deprecate private but non-underscored ndimage.<module> namespace

  • #14477: MAINT: Using Tempita file for bspline (signal)

  • #14479: Added `Inverse` word in `idstn` and `idctn` docstrings

  • #14487: TST: modify flaky test for constrained minimization

  • #14489: MAINT: cleanup of some line_search code

  • #14492: CI: make sure Azure job step fails when building a SciPy wheel…

  • #14496: MAINT: switch to using spmatrix.toarray instead of .todense

  • #14499: DOC: fix toarray/todense docstring

  • #14507: CI: Add lint_diff docs & option to run only on specified files/dirs

  • #14513: DOC: added reference and example in jacobi docstring

  • #14520: BUG: diffev maxfun can be reached partway through population

  • #14524: ENH: Rotation.concatenate

  • #14532: ENH: sparse.linalg: The solution is zero when R.H.S. is zero

  • #14538: CI: Revert “CI: make sure Azure job step fails when building…

  • #14539: DOC: added chebyt and chebyu docstring examples in scipy.special

  • #14546: ENH: Orthogonal Latin Hypercube Sampling to QMC

  • #14547: ENH: __setitem__ method for Rotation class

  • #14549: Small test fixes for pypy + win + mmap

  • #14554: ENH: scipy.interpolate.BSpline from_power_basis

  • #14555: BUG: sparse: fix a DIA.tocsc bug

  • #14556: Fix the link to details of the strongly connected components…

  • #14559: WIP: TST: add tests for Pythran somersd

  • #14561: DOC: added reference and examples in (gen)laguerre docstring…

  • #14564: ENH: Add threaded Van Der Corput

  • #14571: Fix repeated word in _mannwhitneyu.py example

  • #14572: Set min length of the knot array for BSpline.design_matrix

  • #14578: DOC: added examples in spherical Bessel docstrings

  • #14581: MAINT: Refactor `linalg.tests.test_interpolative::TestInterpolativeDecomposition::test_id`

  • #14588: ENH: Added ``kulczynski1`` to ``scipy.spatial.distance``

  • #14592: DOC: clarify parameters of norminvgauss in scipy.stats

  • #14595: Removing unused subroutines in ``scipy/linalg/src/id_dist/src/prini.f``

  • #14601: Fixed inconsistencies between numpy and scipy interp

  • #14602: MAINT: Fix `-Wunused-result` warnings in `sparse/linalg/dsolve`

  • #14603: DEV: initialize all submodules in Gitpod Dockerfile

  • #14609: MAINT: Fix `-Wmaybe-uninitialized` warnings in `optimize/_highs`

  • #14610: MAINT: Ignored ``scipy/signal/bspline_util.c``

  • #14613: MAINT: interpolate: Declare type for a Cython indexing variable.

  • #14619: ENH: stats.unuran: add Polynomial interpolation based numerical…

  • #14620: CI: fix Azure job which uses pre-release wheels + Python 3.7

  • #14625: ENH: optimize min max and median scipy.stats.binned_statistic

  • #14626: MAINT: fix type-narrowing addition in sparse.construct.bmat

  • #14627: MAINT: Bumped tolerances to pass ``special.tests`` on Apple…

  • #14628: DOC: clarify usage of options param in scipy.optimize.linprog

  • #14629: ENH: optimize std in scipy.stats.binned_statistic

  • #14630: DOC: add citation file

  • #14631: Fix unuran builds for older compilers

  • #14633: BUG: scipy.stats._unran: send only strings to include_dirs

  • #14634: DOC: Fix Wikipedia bootstrap link

  • #14635: DOC: stats: fix multivariate_t docs pdf eqn

  • #14637: MAINT: copy discrete dist dict

  • #14643: MAINT: address gh6019, disp for minimize_scalar

  • #14644: DOC: stats: add UNU.RAN references in the tutorial

  • #14649: DOC: clarify SciPy compatibility with Python and NumPy.

  • #14655: MAINT: remove support for Python 3.7 (hence NumPy 1.16)

  • #14656: MAINT: replacing assert_ with assert

  • #14658: DOC: use conda-forge in Ubuntu quickstart

  • #14660: MAINT: refactor “for … in range(len(” statements

  • #14663: MAINT: update leftover Python and NumPy version from pyproject.toml

  • #14665: BLD: fix confusing “import pip” failure that should be caught

  • #14666: MAINT: remove unnecessary seeding and update `check_random_state`

  • #14669: ENH: Refactor GitHub Issue form templates

  • #14673: BLD: fix include order, Python.h before standard headers

  • #14676: BUG: Fixes failing benchmark tests optimize_qap.QuadraticAssignment.track_score

  • #14677: MAINT: github labeler on file paths

  • #14682: DOC: Fix typo in mannwhitneyu docstring

  • #14684: DOC: optimize: fix sporadic linprog doctest failure

  • #14685: MAINT: static typing of entropy

  • #14686: BUG: fix issue in lsqr.py introduced in a recent commit

  • #14689: MAINT: replace IOError alias with OSError or other appropriate…

  • #14692: MAINT: Translation of hyp2f1 for complex numbers into Cython,…

  • #14693: DOC: update OptimizeResult notes

  • #14694: Simplify PythranBuildExt usage

  • #14695: BLD: bump Pythran version to 0.9.12

  • #14697: CI: add `cffi` in the benchmark CI job, and in environment.yml

  • #14699: BUG: Fix TypeError in `stats._discrete_distns`

  • #14700: DOC: update detailed roadmap

  • #14701: ENH:linalg: Add Cythonized get_array_bandwidth, issymmetric,…

  • #14706: BUG: Fix hyp2f1 to return correct values in regions near exp(±iπ/3).

  • #14707: Update constants.py

  • #14708: BENCH: shorten svds benchmark that is timing out in CI

  • #14709: CI: remove labeler sync

  • #14712: MAINT: special: Updates for _cosine.c.

  • #14720: DOC: optimize hess and consistency

  • #14721: MAINT: correct PR template link

  • #14723: DOC: add note on padding to `stats.binned_statistic_2d` docs

  • #14727: ENH: sparse.linalg: Add an useful nonzero initial guess option

  • #14729: DOC: fix documentation for scipy.optimize.brenth

  • #14737: BUG:signal: matching window dtype to input

  • #14739: TST: sparse.linalg: Add test case with 2-D Poisson equations

  • #14743: TST:sparse.linalg: Use the more convenient “assert_normclose”…

  • #14748: DOC: fix matrix representation in scipy.sparse.csgraph

  • #14751: ENH: numpy masked_arrays in refguide-check

  • #14755: BUG: Avoid `solve_ivp` failure when `ts` is empty

  • #14756: MAINT: LinAlgError from public numpy.linalg

  • #14759: BLD: change section name in site.cfg.example from ALL to DEFAULT

  • #14760: TST: suppress jinja2 deprecation warning

  • #14761: CI: remove `pre_release_deps_source_dist` job from Azure CI…

  • #14762: TST: add a seed to the pickling test of RBFInterpolator

  • #14763: MAINT: Make solve_ivp slightly more strict wrt. t_span.

  • #14772: DOC:special: Fix broken links to jburkardt

  • #14787: MAINT: Increase tolerance values to avoid test failures

  • #14789: MAINT: fix a tiny typo in signal/spectral.py

  • #14790: [MRG] BUG: Avoid lobpcg failure when iterations can’t continue

  • #14794: Fix typos in bspline docs (and comments)

  • #14796: MAINT: Allow F401 and F403 in module init files

  • #14798: BUG: correct the test loop in test_arpack.eval_evec

  • #14801: CI, MAINT: pin Cython for azure pre-rel

  • #14805: BUG: optimize: fix max function call validation for minimize…

  • #14808: Fix Bug #14807

  • #14814: MAINT:integrate: add upstream quadpack changes

  • #14817: ENH: stats: add geometric zscore

  • #14820: MAINT: Remove `np.rollaxis` usage with `np.moveaxis` and…

  • #14821: DOC: Updated documentation for interp1d

  • #14822: Add an array API to scipy.sparse

  • #14832: MAINT: py3.10 in more jobs and bump some 3.8 to 3.9

  • #14833: FIX: raise Python OverflowError exception on Boost.Math error

  • #14836: Bug fix: dqc25f.f

  • #14837: DOC: sparse.linalg: Fixed incorrect comments when the initial…

  • #14838: TST: seed a stats test

  • #14841: MAINT: Increase tolerances in tests to avoid Nightly CPython3.10…

  • #14844: DOC: Add refguide_check option details to runtests.rst

  • #14845: DOC: update a type specifier in a docstring in `radau.py`

  • #14848: Typo “copmlex”

  • #14852: DOC: Fix documentation bugs in `lstsq`

  • #14860: minimize: copy user constraints if parameter is factored out….

  • #14865: BUG: stats: Fix a crash in stats.skew

  • #14868: [MRG] BUG: Update lobpcg.py to validate the accuracy and issue…

  • #14871: MAINT: removed a pitfall where a built-in name was being shadowed

  • #14872: DEP: Deprecate private namespaces in `scipy.linalg`

  • #14878: TST: bump rtol for equal_bounds

  • #14881: DEP: Deprecate private namespaces in `scipy.special`

  • #14882: BUG: Convert TNC C module to cython

  • #14883: DOC:linalg: Clarify driver defaults in eigh

  • #14884: BUG: optimize: add missing attributes of `OptimizeResult` for…

  • #14892: DOC: Correct docs for Hausdorff distance

  • #14898: DEP: Deprecate private namespace in `scipy.stats`

  • #14902: MAINT:linalg: Rename func to “bandwidth”

  • #14906: DEP: Deprecate private namespace in `scipy.constants`

  • #14913: DEP: Deprecate private namespace in `scipy.fftpack`

  • #14916: DEP: Deprecate `stats.biasedurn` and make it private

  • #14918: DEP: Deprecate private namespaces in ``scipy.interpolate``

  • #14919: DEP: Deprecate private namespaces in `scipy.integrate`

  • #14920: Fix for complex Fresnel

  • #14923: DEP: Deprecate private namespaces in ``scipy.spatial``

  • #14924: Fix extent for scipy.signal.cwt example

  • #14925: MAINT: Ignore build generated files in ``scipy.stats``

  • #14927: DEP: Deprecate private namespaces in `scipy.misc`

  • #14928: MAINT: fix runtest.py overriding `$PYTHONPATH`: prepend instead

  • #14934: BUG: optimize: add a missing attribute of OptimizeResult in `basinhopping`

  • #14939: DEP: Deprecate private namespaces in ``scipy.sparse``

  • #14941: ENH: optimize: add optional parameters of adaptive step size…

  • #14943: DOC: clarify mac pytest; add blank line

  • #14944: BUG: MultivariateNormalQMC with specific QMCEngine remove unneeded…

  • #14947: DOC: adding example to decimate function

  • #14950: MAINT: Use matmul binary operator in scipy.sparse.linalg

  • #14954: DOC: Add missing params to minres docstring.

  • #14955: BUG: stats: fix broadcasting behavior of argsreduce

  • #14960: Update links for new site

  • #14961: CI: use https protocol for git in CircleCI

  • #14962: DEP: Deprecate private namespaces in `scipy.signal`

  • #14963: MAINT: `integrate.lsoda` missing in .gitignore

  • #14965: DOC: update logo and add favicon.

  • #14966: DEP: Deprecate private namespaces in ``scipy.optimize``

  • #14969: CI: Fixes pyparsing version in doc build

  • #14972: Don’t put space after directive name.

  • #14979: BUG: scipy.sparse.linalg.spsolve: fix memory error caused from…

  • #14988: BLD: update pyproject.toml for Python 3.10

  • #14989: ENH: Speed up knot interval lookup for BSpline.design_matrix

  • #14992: Pythranized version of _matfuncs_sqrtm

  • #14993: MAINT: forward port 1.7.2 relnotes

  • #15004: ENH: Make `get_matfile_version` and other `io.matlab` objects…

  • #15007: DOC: add missing “regularized” to `gammainccinv` documentation

  • #15008: MAINT: restore access to deprecated private namespaces

  • #15010: TST: remove fragile test which checks if g77 is linked

  • #15013: MAINT: Fix use-after-free bug in Py_FindObjects

  • #15018: CI: Work around Sphinx bug

  • #15019: Finite Difference Hessian in Scipy Optimize Solvers (Newton-CG)

  • #15020: ENH: sparse.linalg: Fixed the issue that the initial guess “x0”…

  • #15022: DOC: mitigate newton optimization not converging.

  • #15023: CI: Unpin Sphinx

  • #15027: DOC: linalg: Fix a small condition doc error

  • #15029: DEP: Deprecate private namespaces in `scipy.sparse.linalg`

  • #15034: DOC: use numpydoc format for C function in `_superlumodule.c`

  • #15035: MAINT: simplify UNU.RAN api in stats

  • #15037: New example for gaussian_filter

  • #15040: MAINT: Add test for public API

  • #15041: DOC: Add warning to dct documentation about norm=’ortho’

  • #15045: DOC: update toolchain.rst

  • #15053: TST: Add some test skips to get wheel builder CI green again

  • #15054: MAINT: Remove wminkowski

  • #15055: ENH: allow p>0 for Minkowski distance

  • #15061: MAINT:sparse: expm() fix redundant imports

  • #15062: MAINT:BLD: Open file in text mode for tempita

  • #15066: CI: bump gcc from 4.8 to 6

  • #15067: DOC: Update broken link to SuperLU library.

  • #15078: MAINT: update `stats.iqr` for deprecated `np.percentile`…

  • #15083: MAINT: stats: separate UNU.RAN functionality to its own submodule

  • #15084: MAINT: Include `scipy.io.matlab` in public API

  • #15085: ENH: support creation of analog SOS outputs

  • #15087: TST: Review ``_assert_within_tol`` positional arguments

  • #15095: MAINT: update gitignore to ignore private directories

  • #15099: MAINT: ScalarFunction remember best_x

  • #15100: MAINT: Include `stats.contingency` in public API

  • #15102: ENH: Add orthogonalize argument to DCT/DST

  • #15105: MAINT: Add missing imports in deprecated modules

  • #15107: BUG: Update chi_gen to use scipy.special.gammaln

  • #15109: MAINT: remove NaiveRatioUniforms from scipy.stats

  • #15111: ENH: Add special.log_expit and use it in stats.logistic

  • #15112: DOC: update ‘Wn’ definition in signal.butter

  • #15114: DOC: added Fermi-Dirac distribution by name

  • #15119: DOC: fix symlink to `logistic.sf` in `stats.logistic`

  • #15120: MAINT: Install `sparse.linalg._eigen` tests and fix test failures

  • #15123: MAINT: interpolate: move the `sparse` dependency from cython…

  • #15127: DOC: update linux build instructions to mention C++

  • #15134: DOC: Improve Lomb-Scargle example

  • #15135: ENH: Carlson symmetric elliptic integrals.

  • #15137: DOC: special: Add ‘Examples’ to multigammaln and roots_legendre…

  • #15139: Use constrained_layout in Lomb-Scargle example

  • #15142: ENH: stats.sampling: add SROU method

  • #15143: MAINT: Remove some unused imports.

  • #15144: BUG: Add missing import of ‘errno’ to runtests.py

  • #15157: ENH: rebased version of gh-14279

  • #15159: DOC: stats: fix a header in `stats.sampling` tutorial

  • #15161: DOC: 1.8.0 relnotes update

  • #15175: MAINT: 1.8.0 backports for relnotes and .gitignore

  • #15181: BUG: The pytest decorator for conditional skipping is ‘skipif’

  • #15191: MAINT: version bounds before 1.8.0rc1

  • #15192: MAINT: Replace use of `pytest.warns(None)` with `warnings.catch_warnings`

  • #15194: BUG: stats: Fix numerical issues of recipinvgauss

  • #15214: TST: sparse.linalg: store only PROPACK test matrices; generate…

  • #15220: BUG: sparse.linalg: Fix deprecation warnings.

  • #15225: Make 0th power of a sparse array/matrix return the identity with…

  • #15229: BUG: minimize should work with a scalar closes #15228

  • #15232: BUG: Add rmul for sparse arrays

  • #15236: BLD: update setup.py for Python 3.10

  • #15248: MAINT: 1.8.0rc2 backports

  • #15249: FIX: PROPACK MKL compatibility

  • #15253: BUG: special: fix `stdtr` and `stdtrit` for infinite df

  • #15256: MAINT: use PEP440 vs. distutils

  • #15268: CI: pin setuptools to 59.6.0 and Pythran to 0.10.0 for Windows…

  • #15270: MAINT: rename `moduleTNC` extension back to `_moduleTNC`

  • #15271: TST: slightly bump test tolerance for a new lobpcg test

  • #15275: MAINT: Fix imports in `signal._signaltools`

  • #15278: MAINT: remove non-default settings (except `shallow`) in `.gitmodules`

  • #15288: BLD Respect the –skip-build flag in setup.py

  • #15293: BUG: fix Hausdorff int overflow

  • #15301: TST: update `sparse.linalg` tests for failures due to tolerances

  • #15318: BLD: update pyproject.toml to not pin numpy for aarch64 + PyPy

  • #15322: BLD: update minimum Pythran version to 0.10.0 for SciPy 1.8.0

  • #15323: MAINT: filter RuntimeWarnings in stats functions

  • #15328: MAINT: interpolate: csr_matrix -> csr_array

  • #15331: BUG: stats._unuran: fix invalid attribute lookups

  • #15332: CI: pin numpy to 1.21.5 for the doc build on CircleCI

  • #15334: BUG: stats._unuran: fix remaining attribute lookup errors

  • #15335: CI: pin numpy to 1.21.5 in the Azure refguide check job

  • #15341: BUG: `highs-ds` returns memoryviews instead of np.arrays for…

  • #15397: BUG: ensured vendored pep440 is imported

  • #15416: BUG: Fix PyUFunc for wasm targets

  • #15418: MAINT: 1.8.0 rc3 backports round 1

  • #15421: BUG: stats: mode: fix negative axis issue with np.moveaxis instead…

  • #15432: MAINT: release branch PROPACK switch (default off)

  • #15515: MAINT: fix broken link and remove CI badges