Add empty, empty_like, arange, and linspace

[mtsokol]

Hi @hameerabbasi,

This is the last PR with the creation functions. empty and empty_like are implemented the same way as in Numba backend. arange and linspase are strictly dense functions therefore I decided to use NumPy functions and pass results to Tensor (no-copy operation).

[hameerabbasi]

Do the tests now pass in the Array API for these functions?

[mtsokol]

Do the tests now pass in the Array API for these functions?

Yes! Here's the diff:

--- a/ci/Finch-array-api-xfails.txt
+++ b/ci/Finch-array-api-xfails.txt
@@ -72,10 +72,6 @@ array_api_tests/test_signatures.py::test_array_method_signature[__setitem__]
 # test_creation_functions

 # not implemented
-array_api_tests/test_creation_functions.py::test_arange
-array_api_tests/test_creation_functions.py::test_empty
-array_api_tests/test_creation_functions.py::test_empty_like
-array_api_tests/test_creation_functions.py::test_linspace
 array_api_tests/test_creation_functions.py::test_meshgrid

 # test_array_object

[hameerabbasi]

From the CI failures, looks like we need a NumPy 2.0 compat PR. 🤔

[mtsokol]

It was a typo in the test change. Rebased.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 89.86%. Comparing base (aa2f8fb) to head (d61c579).

❗ Current head d61c579 differs from pull request most recent head 0bfed32

Please upload reports for the commit 0bfed32 to get more accurate results.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #72      +/-   ##
==========================================
+ Coverage   89.73%   89.86%   +0.13%     
==========================================
  Files           9        9              
  Lines         828      839      +11     
==========================================
+ Hits          743      754      +11     
  Misses         85       85              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.