MAINT: A few updates to the array_api #20066

asmeurer · 2021-10-07T19:28:43Z

A few minor updates to the array_api namespace based on recent updates to the spec.

CC @seberg you may actually want to review this one before it is merged this time.

The spec has recently been updated to only require multiaxis (i.e., tuple) indices in the case where every axis is indexed, meaning there are either as many indices as axes or the index has an ellipsis.

seberg · 2021-10-07T21:44:17Z

The main comment I remember was that you had a lot of complexity to support arr[boolean_arr,] when maybe all you need to support is arr[boolean_arr]. But I guess there was no change/decision w.r.t. to that one made. And isn't the most important point the non-multi-dim index part? arr3d[0] not being strictly defined or did you decide against that after all?

mattip

linting

numpy/array_api/tests/test_array_object.py

numpy/array_api/_array_object.py

where does value-based promotion for 0-dimensional arrays, so we use the same trick as in the Array operators to avoid this.

asmeurer · 2021-10-19T19:53:01Z

I pushed a fix to the type promotion on where. As far as I can tell, the fix suggested at #18585 (comment) wouldn't work since where isn't a ufunc.

Printing behavior isn't required by the spec. This is just to make things easier to understand, especially with the array API test suite.

This test will not pass NumPy without the changes in numpy/numpy#20066 due to an update in indexing behavior in the spec.

from_dlpack() should be used to create arrays using DLPack.

unique() in the array API was replaced with three separate functions, unique_all(), unique_inverse(), and unique_values(), in order to avoid polymorphic return types. Additionally, it should be noted that these functions to not currently conform to the spec with respect to NaN behavior. The spec requires multiple NaNs to be returned, but np.unique() returns a single NaN. Since this is currently an open issue in NumPy to possibly revert, I have not yet worked around this. See numpy#20326.

This does nothing in NumPy, and is just present so that the signature is valid according to the spec.

The "multiaxis indexing must index every axis explicitly or use an ellipsis" was supposed to include any type of index, not just tuple indices.

asmeurer · 2021-11-09T01:03:16Z

This is ready for review. There have been several small changes to the spec since the last time anyone has looked at this. With this PR, everything should now be up-to-date, except for a handful of things which are being updated in separate pull requests.

rgommers · 2021-11-10T11:19:49Z

@asmeurer there is one test failure, test_device_property. Could you fix that?

rgommers · 2021-11-10T11:20:53Z

The boolean indexing question higher up (#20066 (comment)) also deserves a reply.

rgommers

This all LGTM modulo the test failure.

numpy/array_api/_set_functions.py

numpy/array_api/tests/test_creation_functions.py

asmeurer · 2021-11-10T23:05:35Z

@asmeurer there is one test failure, test_device_property. Could you fix that?

Sorry about that. #20335 has messed up my ability to run the tests locally.

numpy/array_api/_array_object.py

asmeurer · 2021-11-10T23:27:58Z

The main comment I remember was that you had a lot of complexity to support arr[boolean_arr,] when maybe all you need to support is arr[boolean_arr]. But I guess there was no change/decision w.r.t. to that one made. And isn't the most important point the non-multi-dim index part? arr3d[0] not being strictly defined or did you decide against that after all?

I did address the point of treating a[0] and a[0,] differently. That was just a misreading of the spec on my part. I think we could drop a[boolean_arr,] if that simplifies the code. What was the suggested way to do that? Any refactoring of the index validation should be fine so long as it passes the tests in test_array_object.py as well as the test_indexing test in the array API testsuite.

Co-authored-by: Bas van Beek <43369155+BvB93@users.noreply.github.com>

The array_api cannot use the NumPy testing functions because array_api arrays do not mix with NumPy arrays, and also NumPy testing functions may use APIs that aren't supported in the array API.

…rray-api-updates2

rgommers · 2021-11-12T13:37:30Z

This all looks good. CI failures are okay: linter ones are uninteresting, the typing error was fixed in main yesterday, and the segfault is unrelated (happening in other PRs). So in this goes.

I did address the point of treating a[0] and a[0,] differently. That was just a misreading of the spec on my part. I think we could drop a[boolean_arr,] if that simplifies the code. What was the suggested way to do that?

Seems fine either way to me, and can be done in a separate follow-up given that it's a discussion that predates this PR. Let's see if @seberg has a preference here.

rgommers · 2021-11-12T13:39:16Z

Ah wait, there's even more failures. They all look unrelated, but rerunning CI anyway to get this a bit more green.

rgommers · 2021-11-12T15:05:53Z

Looking much better now, in it goes. Thanks @asmeurer & reviewers!

seberg · 2021-11-12T15:10:57Z

A followup is fine and would be nice. I don't recall the logic exactly, what I do remember is that the indexing check was very complex (with some recursive calls), and a lot of that complexity was there to accomodate arr[boolean_arr,]. I believe in the original PR I had proposed a simplification.
That said, I think it is correct, just seemed more complex than necessary.

seberg · 2021-11-12T15:11:37Z

PS: And since this is experimental and a corner case, I don't think it would matter if you change arr[boolean_arr,] from being allowed to being rejected.

This test will not pass NumPy without the changes in numpy/numpy#20066 due to an update in indexing behavior in the spec.

asmeurer added 2 commits October 1, 2021 16:43

Allow casting in the array API asarray()

48005c1

Restrict multidimensional indexing in the array API namespace

9af4814

The spec has recently been updated to only require multiaxis (i.e., tuple) indices in the case where every axis is indexed, meaning there are either as many indices as axes or the index has an ellipsis.

github-actions bot added the 03 - Maintenance label Oct 7, 2021

mattip reviewed Oct 10, 2021

View reviewed changes

numpy/array_api/tests/test_array_object.py Show resolved Hide resolved

numpy/array_api/_array_object.py Outdated Show resolved Hide resolved

Fix type promotion for numpy.array_api.where

cb0b9c6

where does value-based promotion for 0-dimensional arrays, so we use the same trick as in the Array operators to avoid this.

asmeurer added 2 commits October 20, 2021 15:55

Print empty array_api arrays using empty()

7da7a99

Printing behavior isn't required by the spec. This is just to make things easier to understand, especially with the array API test suite.

Fix an incorrect slice bounds guard in the array API

0169af7

asmeurer added a commit to data-apis/array-api-tests that referenced this pull request Oct 27, 2021

Add test that svd values are sorted

139e83e

This test will not pass NumPy without the changes in numpy/numpy#20066 due to an update in indexing behavior in the spec.

asmeurer added 5 commits November 5, 2021 14:16

Disallow multiple different dtypes in the input to np.array_api.meshgrid

e0589fd

Remove DLPack support from numpy.array_api.asarray()

21faf34

from_dlpack() should be used to create arrays using DLPack.

Remove __len__ from the array API array object

d0f591d

Add astype() to numpy.array_api

17d7886

rgommers added the component: numpy.array_api label Nov 8, 2021

asmeurer added 6 commits November 8, 2021 16:17

Add the stream argument to the array API to_device method

cb335d2

This does nothing in NumPy, and is just present so that the signature is valid according to the spec.

Use the NamedTuple classes for the type signatures

5cae94d

Add unique_counts to the array API namespace

f6053aa

Remove some unused imports

680e0a4

Update the array_api indexing restrictions

49a3cc9

The "multiaxis indexing must index every axis explicitly or use an ellipsis" was supposed to include any type of index, not just tuple indices.

Merge branch 'main' into array-api-updates2

475b01d

asmeurer mentioned this pull request Nov 9, 2021

BUG: Error when importing NumPy main (No module named 'numpy.core._multiarray_umath') #20335

Closed

rgommers approved these changes Nov 10, 2021

View reviewed changes

numpy/array_api/_set_functions.py Outdated Show resolved Hide resolved

numpy/array_api/tests/test_creation_functions.py Show resolved Hide resolved

BvB93 reviewed Nov 10, 2021

View reviewed changes

numpy/array_api/_array_object.py Outdated Show resolved Hide resolved

asmeurer and others added 6 commits November 10, 2021 17:29

Use a simpler type annotation for the array API to_device method

6352d11

Co-authored-by: Bas van Beek <43369155+BvB93@users.noreply.github.com>

Fix a typo

b7bf06e

Fix a test failure in the array_api submodule

61bc679

The array_api cannot use the NumPy testing functions because array_api arrays do not mix with NumPy arrays, and also NumPy testing functions may use APIs that aren't supported in the array API.

Merge branch 'array-api-updates2' of github.com:asmeurer/numpy into a…

f72ed85

…rray-api-updates2

Merge branch 'main' into array-api-updates2

09aa2b5

Add dlpack support to the array_api submodule

580a616

rgommers added this to the 1.22.0 release milestone Nov 12, 2021

Merge branch 'main' into array-api-updates2

b2af8e9

rgommers merged commit ff2e2a1 into numpy:main Nov 12, 2021

leofang pushed a commit to leofang/cupy that referenced this pull request Nov 12, 2021

apply diff from numpy/numpy#20066

797b2af

leofang mentioned this pull request Nov 12, 2021

Apply upstream patch to cupy.array_api cupy/cupy#6086

Merged

cr313 added a commit to cr313/test-array-api that referenced this pull request Apr 19, 2024

Add test that svd values are sorted

43ff81d

This test will not pass NumPy without the changes in numpy/numpy#20066 due to an update in indexing behavior in the spec.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

MAINT: A few updates to the array_api #20066

MAINT: A few updates to the array_api #20066

asmeurer commented Oct 7, 2021

seberg commented Oct 7, 2021

mattip left a comment

asmeurer commented Oct 19, 2021

asmeurer commented Nov 9, 2021

rgommers commented Nov 10, 2021

rgommers commented Nov 10, 2021

rgommers left a comment

asmeurer commented Nov 10, 2021

asmeurer commented Nov 10, 2021

rgommers commented Nov 12, 2021

rgommers commented Nov 12, 2021

rgommers commented Nov 12, 2021

seberg commented Nov 12, 2021

seberg commented Nov 12, 2021

MAINT: A few updates to the array_api #20066

MAINT: A few updates to the array_api #20066

Conversation

asmeurer commented Oct 7, 2021

seberg commented Oct 7, 2021

mattip left a comment

Choose a reason for hiding this comment

asmeurer commented Oct 19, 2021

asmeurer commented Nov 9, 2021

rgommers commented Nov 10, 2021

rgommers commented Nov 10, 2021

rgommers left a comment

Choose a reason for hiding this comment

asmeurer commented Nov 10, 2021

asmeurer commented Nov 10, 2021

rgommers commented Nov 12, 2021

rgommers commented Nov 12, 2021

rgommers commented Nov 12, 2021

seberg commented Nov 12, 2021

seberg commented Nov 12, 2021