Implement `tensor.isin` #2098

ndgrigorian · 2025-06-06T08:54:47Z

This PR proposes an implementation for isin, a function likely coming to a future array API specification, which leverages a similar kernel to the implementation of searchsorted

This implementation uses the searchsorted kernel to check if the value has a position in the array. If that position is the number of elements in the array, it is not a member. Otherwise, if arr[pos] == val for some array arr being searched for value val, then val is a member.

Have you provided a meaningful PR description?
Have you added a test, reproducer or referred to an issue with a reproducer?
Have you tested your changes locally for CPU and GPU devices?
Have you made sure that new changes do not introduce compiler warnings?
Have you checked performance impact of proposed changes?
Have you added documentation for your changes, if necessary?
Have you added your changes to the changelog?
If this PR is a work in progress, are you opening the PR as a draft?

github-actions · 2025-06-06T09:31:48Z

View rendered docs @ https://intelpython.github.io/dpctl/pulls/2098/index.html

github-actions · 2025-06-06T09:40:22Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_8 ran successfully.
Passed: 1115
Failed: 6
Skipped: 119

coveralls · 2025-06-06T09:40:36Z

coverage: 84.892% (-0.1%) from 84.989%
when pulling 23c61a8 on feature/searchsorted-based-isin
into 35a8c26 on master.

github-actions · 2025-06-06T23:26:19Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_10 ran successfully.
Passed: 1114
Failed: 7
Skipped: 119

dpctl/tensor/_set_functions.py

antonwolfy · 2025-06-10T22:16:23Z

dpctl/tensor/_set_functions.py

+
+    dep_evs = _manager.submitted_events
+    ht_ev, s_ev = _isin(
+        needles=x1,


It seems the only case when the strided implementation (which assumes slower) will be used when x1 array is not contiguous (we sort test_elements array and no out keyword in isin function).
Would it make sense to flatten input array x and to pass order keyword there?

But, it makes sense also to keep strided implementation of _isin in case when it might be helpful in implementation of other set functions).

I can experiment and see values of flattening vs. not flattening

but in general, this implementation is going to be changed quite a bit soon, I have some local changes waiting

ndgrigorian · 2025-06-11T22:14:55Z

@antonwolfy
implementation is updated and now permits Python scalars in second argument

tests still need to be added, but isin itself is ready for review

github-actions · 2025-06-11T22:55:11Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_17 ran successfully.
Passed: 1115
Failed: 6
Skipped: 119

antonwolfy · 2025-06-11T22:40:01Z

dpctl/tensor/CMakeLists.txt

@@ -112,6 +112,7 @@ set(_reduction_sources
    ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/reductions/sum.cpp
 )
 set(_sorting_sources
+    ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/sorting/isin.cpp


It doesn't seem relating to sorting routine

it uses common utilities with searchsorted (i.e., from rich_comparisons.hpp) which is why it lives there

if the code from rich_comparisons gets factored out, I can go ahead and move it elsewhere, I guess to _tensor_impl for now

dpctl/tensor/_set_functions.py

antonwolfy · 2025-06-11T23:11:10Z

dpctl/tensor/libtensor/source/sorting/isin.cpp

+    fnT get() const
+    {
+        using dpctl::tensor::kernels::isin_contig_impl;
+        using Compare = typename AscendingSorter<argTy>::type;


Do we need to have Compare templated here? Is there any use case possible when another one will be required to be used by isin kernel?

It's not strictly necessary, but was done to reduce code duplication—these sorters are defined in tensor/source.

I can look at the normal sort implementation to refresh myself on what was done there, but if that isn't sufficient, it may be preferable to template and pass the sorter here as opposed to duplicating in isin.hpp

dpctl/tensor/libtensor/include/kernels/sorting/isin.hpp

dpctl/tensor/_set_functions.py

github-actions · 2025-06-16T08:56:49Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_18 ran successfully.
Passed: 1115
Failed: 6
Skipped: 119

github-actions · 2025-06-16T21:51:25Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_22 ran successfully.
Passed: 1115
Failed: 6
Skipped: 119

github-actions · 2025-06-16T22:17:02Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_23 ran successfully.
Passed: 1115
Failed: 6
Skipped: 119

github-actions · 2025-06-17T07:08:12Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_24 ran successfully.
Passed: 1115
Failed: 6
Skipped: 119

github-actions · 2025-06-17T08:30:21Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_23 ran successfully.
Passed: 1116
Failed: 5
Skipped: 119

dpctl/tensor/_set_functions.py

antonwolfy · 2025-06-17T08:21:36Z

dpctl/tensor/_set_functions.py

+
+    if not isinstance(x, dpt.usm_ndarray):
+        x_arr = dpt.asarray(
+            x, dtype=dt1, usm_type=res_usm_type, sycl_queue=exec_q


Do we need to cast here to result dtype to avoid unnecessary copy below?

Suggested change

x, dtype=dt1, usm_type=res_usm_type, sycl_queue=exec_q

x, dtype=dt, usm_type=res_usm_type, sycl_queue=exec_q

The same with test_arr casting here

And probably then it would sense to combine the checks, like:

if not isinstance(x, dpt.usm_ndarray): x_buf = dpt.asarray( x, dtype=dt, usm_type=res_usm_type, sycl_queue=exec_q ) elif x_dt != dt: x_buf = _empty_like_orderK(x, dt, res_usm_type, sycl_dev) ht_ev, ev = _copy_usm_ndarray_into_usm_ndarray( src=x, dst=x_buf, sycl_queue=exec_q, depends=dep_evs ) _manager.add_event_pair(ht_ev, ev) else: x_buf = x

I've left it this way because this is how element-wise functions handle it, as well: scalars are put first into the appropriate array type, then the array is cast into another type for computation

I'm not sure if it's strictly necessary, but may avoid some edge cases producing incorrect results

dpctl/tests/test_tensor_isin.py

antonwolfy · 2025-06-17T08:59:43Z

dpctl/tests/test_tensor_isin.py

+
+
+@pytest.mark.parametrize(
+    "dtype",


What is about case when inputs have different dtypes and casting is required?

github-actions · 2025-06-17T21:17:04Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_24 ran successfully.
Passed: 1112
Failed: 9
Skipped: 119

github-actions · 2025-06-17T22:16:52Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_26 ran successfully.
Passed: 1115
Failed: 6
Skipped: 119

github-actions · 2025-06-17T23:04:20Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_27 ran successfully.
Passed: 1115
Failed: 6
Skipped: 119

github-actions · 2025-06-18T06:36:39Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_30 ran successfully.
Passed: 1115
Failed: 6
Skipped: 119

isin leverages kernel very similar to searchsorted, but after the search, the position is checked, and if the position is equal to the number of elements in the searched array, existence is considered false

permit scalar input for second argument, address some review comments, add docstring

radix sort implementation asserts that array must be of size > 1

github-actions · 2025-06-18T20:50:10Z

Array API standard conformance tests for dpctl=0.21.0dev0=py310h93fe807_35 ran successfully.
Passed: 1115
Failed: 6
Skipped: 119

ndgrigorian force-pushed the feature/searchsorted-based-isin branch from 1805102 to 5355fb8 Compare June 6, 2025 22:41

antonwolfy reviewed Jun 10, 2025

View reviewed changes

ndgrigorian marked this pull request as ready for review June 11, 2025 22:15

ndgrigorian requested review from vlad-perevezentsev and vtavana as code owners June 11, 2025 22:15

ndgrigorian requested a review from antonwolfy June 11, 2025 22:20

antonwolfy reviewed Jun 11, 2025

View reviewed changes

ndgrigorian requested a review from antonwolfy June 17, 2025 06:24

ndgrigorian force-pushed the feature/searchsorted-based-isin branch from b3822f3 to f7e0967 Compare June 17, 2025 07:40

antonwolfy reviewed Jun 17, 2025

View reviewed changes

ndgrigorian force-pushed the feature/searchsorted-based-isin branch from 8f19cb5 to 3cf7445 Compare June 17, 2025 21:32

ndgrigorian force-pushed the feature/searchsorted-based-isin branch from 3cf7445 to 7c6a4be Compare June 17, 2025 22:20

ndgrigorian added 3 commits June 18, 2025 13:06

Implement tensor.isin

0fd0db1

isin leverages kernel very similar to searchsorted, but after the search, the position is checked, and if the position is equal to the number of elements in the searched array, existence is considered false

factor common utilities for scalar arguments to a new file

50e2051

Make constexpr variables in isin static

1ec28a6

ndgrigorian added 12 commits June 18, 2025 13:06

Update implementation of isin

27c03c7

permit scalar input for second argument, address some review comments, add docstring

Update per review comments

1cb8a13

Allow x to be a scalar in isin and remove assume_unique

ebe745b

Make comparator static constexpr

d223c3e

add basic tests for isin functionality

89ccd97

add a fast-path for size == 1 arrays in sort

910602b

radix sort implementation asserts that array must be of size > 1

Remove unused import of dpctl in _set_functions.py

dfe34ce

Add type hints to isin

d17837a

Add fast-path for size == 1 arrays to argsort

c09b9ff

Add usm_type to test_buf in isin

75a79e2

Address review comments for isin tests

aa957b6

Add test covering nans and +/- 0 in isin

23c61a8

ndgrigorian force-pushed the feature/searchsorted-based-isin branch from 16c63e4 to 23c61a8 Compare June 18, 2025 20:06

	x, dtype=dt1, usm_type=res_usm_type, sycl_queue=exec_q
	x, dtype=dt, usm_type=res_usm_type, sycl_queue=exec_q



		@pytest.mark.parametrize(
		"dtype",

Implement tensor.isin #2098

Are you sure you want to change the base?

Implement tensor.isin #2098

Uh oh!

Conversation

ndgrigorian commented Jun 6, 2025

Uh oh!

github-actions bot commented Jun 6, 2025

Uh oh!

github-actions bot commented Jun 6, 2025

Uh oh!

coveralls commented Jun 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Jun 6, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ndgrigorian commented Jun 11, 2025

Uh oh!

github-actions bot commented Jun 11, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

github-actions bot commented Jun 16, 2025

Uh oh!

github-actions bot commented Jun 16, 2025

Uh oh!

github-actions bot commented Jun 16, 2025

Uh oh!

github-actions bot commented Jun 17, 2025

Uh oh!

github-actions bot commented Jun 17, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Jun 17, 2025

Uh oh!

github-actions bot commented Jun 17, 2025

Uh oh!

github-actions bot commented Jun 17, 2025

Uh oh!

github-actions bot commented Jun 18, 2025

Uh oh!

github-actions bot commented Jun 18, 2025

Uh oh!

Implement `tensor.isin` #2098

Implement `tensor.isin` #2098

coveralls commented Jun 6, 2025 •

edited

Loading