Array API support tracker

**upstream tracker: scipy/scipy#18867**

## Typing the array API

... that'll have to wait until *someone* manages to find the time to work on [array-api-typing](https://github.com/data-apis/array-api-typing) 😅 

## Relevant `scipy` bits

The Array API is partially supported `scipy`. The following packages support it fully:

- `cluster`
  - `vq`
    - `whiten`
    - `[py_]vq`
    - `kmeans[2]`
  - `hierarchy`
    - `int_floor`
    - `linkage` (used by `single`, `complete`, `average`, `weighted`, `centroid`, `median`, `ward`)
    - `{cut,to}_tree`
    - `optimal_leaf_ordering`
    - `cophenet`
    - `inconsistent`
    - `{from,to}_mlab_linkage`
    - `is_monotonic`
    - `is_valid_{im,linkage}`
    - `num_obs_linkage`
    - `correspond`
    - `fcluster[data]`
    - `leaves_list`
    - `dendrogram`
    - `max{dists,inconsts,Rstat}`
    - `leaders`
- `constants`
  - `convert_temperature`
  - `lambda2nu` [1.16+]
  - `nu2lambda` [1.16+]
- `datasets`
- `differentiate` [1.15+]
  - `derivative`
  - `jacobian`
  - `hessian`
- `fft`
  - `[i][r|h]fft[2|n]`
  - `[i]dct[n]`
  - `[i]dst[n]`
  - `[r]fftfreq`
  - `[i]fftshift`
- `integrate` [1.15+]
  - `cubature`
  - `tanhsinh`
  - `trapezoid`
  - `nsum`
- `io`
- ~`linalg`~ (not yet, see [scipy/scipy#19068](https://github.com/scipy/scipy/issues/19068))
- `ndimage`
  - `vectorized_filter` [1.16+]
- `optimize`
  - `_differentiable_functions`
  - `elementwise.*` [1.15+]
  - `rosen[_{der,hess,hess_prod}]` [1.15+]
- `signal` (see [scipy/scipy#20678](https://github.com/scipy/scipy/issues/20678))
  - `findfreqs` [1.16+]
  - `freqs[_{zpk,sos}]` [1.16+]
  - `freqz[_{zpk,sos}]` [1.16+]
  - `envelope` [1.16+]
  - `resample[_poly]` [1.16+]
  - `remez` [1.16+]
  - `fir{win[2],ls}` [1.16+]
  - `minimum_phase` [1.16+]
  - `zpk2{tf,sos}` [1.16+]
  - `tf2{zpk,sos}` [1.16+]
  - `sos2{tf,zpk}` [1.16+]
  - `lp2{lp,hp,bp,bs}` [1.16+]
  - `vectorstrength` [1.16+]
  - `upfirdn` [1.16+]
- `signal.windows` [1.16+] (https://github.com/scipy/scipy/pull/21783)
- ~`sparse`~ (not yet, see [scipy/scipy#18867](https://github.com/scipy/scipy/pull/20190))
- `special`
  - `log_ndtr`
  - `ndtr[i]`
  - `chdtr` [1.15+]
  - `chdtrc`
  - `erf[c]`
  - `i0[e]`
  - `i1[e]`
  - `betainc[c]` [1.15+]
  - `gammaln`
  - `gammainc[c]`
  - `{log,exp}it`
  - `[rel_]entr`
  - `xlogy`
  - `stdtrit` [1.16+]
  - `[log_]softmax` [1.16+]
  - `logsumexp` [1.16+]
- `stats` (see [scipy/scipy#20544](https://github.com/scipy/scipy/issues/20544))
  - `describe`
  - `moment`
  - `variation`
  - `sem`
  - `skew`
  - `kurtosis`
  - `kstat[var]`
  - `z{map,score}` [1.15+]
  - `gzscore` [1.15+]
  - `gstd` [1.16+]
  - `t{mean,var,str,sem,min,max}` [1.15+]
  - `{g,h,p,t}mean` [1.15+]
  - `quantile` [1.16+]
  - `rankdata` [1.16+]
  - `entropy`
  - `differential_entropy` [1.15+]
  - `power_divergence`
  - `circ{mean,var,std}`
  - `directional_stats` [1.15+]
  - `ttest_{1samp,ind,ind_from_stats,rel}` [1.15+]
  - `{skew,kurtosis,normal,monte_carlo}test`
  - `pearsonr`
  - `chisquare`
  - `jarque_bera`
  - `bartlett`
  - `boxcox_llf` [1.15+]
  - `combine_pvalues` [1.15+]

## How Scipy implements it

Internally, `scipy` implements the support in [`scipy._lib._array_api`](https://github.com/scipy/scipy/blob/59fd061b353881ad6cf84e3fd9aa398de8ecb136/scipy/_lib/_array_api.py). From the looks of it, this module will come with a bunch of extra utility functions in `scipy 1.15` ([`37174ad`](https://github.com/scipy/scipy/commit/37174adf7f98b6a39c7aaba3467533add51290b8)), as well as an additional [`scipy._lib._array_api_no_0d`](https://github.com/scipy/scipy/blob/59fd061b353881ad6cf84e3fd9aa398de8ecb136/scipy/_lib/_array_api_no_0d.py) module in [`281a8c3`](https://github.com/scipy/scipy/commit/281a8c3c408681f5683754dfc9ce22e0d31b547d)

## Libraries that support it

Some libraries require [`array-api-compat`](https://github.com/data-apis/array-api-compat) for them to support the array-api (2022):

- [`numpy < 2`](https://github.com/numpy/numpy)
- [`cupy`](https://github.com/cupy/cupy/)
- [`pytorch`](https://github.com/pytorch/pytorch)
- [`dask > 2023.12.0`](https://github.com/dask/dask)
- [`ndonnx`](https://github.com/Quantco/ndonnx)

the following libraries have native support:

- [`numpy >= 2`](https://github.com/numpy/numpy)
- [`cupy`](https://github.com/cupy/cupy/) ([but not really](https://docs.cupy.dev/en/stable/reference/array_api.html))
- [`jax`](https://github.com/jax-ml/jax/)
- [`sparse`](https://github.com/pydata/sparse)
- [`ndonnx`](https://github.com/Quantco/ndonnx)

## Other relevant links:

- https://docs.scipy.org/doc/scipy/dev/api-dev/array_api.html
- https://data-apis.org/array-api/latest/index.html
- https://github.com/data-apis/array-api

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

Array API support tracker #140

Typing the array API

Relevant `scipy` bits

How Scipy implements it

Libraries that support it

Other relevant links:

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Uh oh!

Array API support tracker #140

Description

Typing the array API

Relevant scipy bits

How Scipy implements it

Libraries that support it

Other relevant links:

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Relevant `scipy` bits