Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

dpnp.unique returns reshaped unique_inverse when axis is None #1999

Merged
merged 23 commits into from
Aug 19, 2024

Conversation

antonwolfy
Copy link
Contributor

@antonwolfy antonwolfy commented Aug 18, 2024

The shape of returned unique_inverse from numpy.unique(..., return_inverse=True, axis=None) has been changed since 2.0.1.
Now multi-dimensional inputs, numpy returns unique_inverse which is reshaped such that the input can be reconstructed using dpnp.take(unique, unique_inverse, axis=axis).

The PR proposes to update dpnp.unique and to align with that behavior.
The tests are updated to properly handle the result from different numpy versions.

  • Have you provided a meaningful PR description?
  • Have you added a test, reproducer or referred to 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?
  • If this PR is a work in progress, are you filing the PR as a draft?

Copy link
Contributor

github-actions bot commented Aug 18, 2024

View rendered docs @ https://intelpython.github.io/dpnp/index.html

Base automatically changed from skip-tests-for-new-umaths to master August 19, 2024 16:17
@antonwolfy antonwolfy merged commit 5c9752b into master Aug 19, 2024
36 of 46 checks passed
@antonwolfy antonwolfy deleted the align-unique-with-numpy-2.0 branch August 19, 2024 20:02
github-actions bot added a commit that referenced this pull request Aug 19, 2024
)

* Remove alias on numpy.complex_

* Remove alias on numpy.cfloat

* Remove alias on numpy.float and numpy.float_

* Remove alias on numpy.float and numpy.singlecomplex

* Remove aliases on numpy.Inf, numpy.Infinity and numpy.infty constants

* Get rid of dpnp.float in the code

* Remove aliases on numpy.NAN and numpy.NaN constants

* Remove aliases on numpy.NINF and numpy.PINF constants

* Remove aliases on numpy.NZERO and numpy.PZERO constants

* Remove use of removed numpy.longcomplex dtype

* Use proper import of numpy exceptions

* Applied pre-commit hooks

* Remove dpnp.issubsctype per numpy 2.0 migration guide

* Skip tests for asfarray with numpy 2.0

* Mute CFD relating tests

* Applied pre-commit hooks

* Add skip fixture in test_arraymanipulation.py::test_asfarray2

* Mute tests for new umaths from numpy 2.0

* dpnp.unique returns reshaped unique_inverse

* Corrected note sentense

* Update test_sycl_queue.py::test_unique 5c9752b
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants