[OpenVINO] Implement nanmedian and nanpercentile ops

[goyaladitya05]

Description

ops nanmedian and nanpercentile were added to keras.ops in #22585 and #22679. This PR implements both for the OpenVINO backend.

Closes: openvinotoolkit/openvino/issues/35502

Contributor Agreement

Please check all boxes below before submitting your PR for review:

• I am a human, and not a bot.
• I will be responsible for responding to review comments in a timely manner.
• I will work with the maintainers to push this PR forward until submission.

Note: Failing to adhere to this agreement may result in your future PRs no longer being reviewed.

[gemini-code-assist]

Code Review

This pull request implements nanmedian and nanpercentile for the OpenVINO backend and enables the corresponding tests. The current implementation of nanmedian is overly complex and contains a bug where keepdims=True does not correctly restore the original rank for multi-axis reductions. Additionally, it fails to cast integer or boolean scalars to float, which is inconsistent with NumPy's behavior. It is recommended to simplify the logic by using the existing nanquantile function.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.03%. Comparing base (0dd8a15) to head (9dad42a).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff             @@
##           master   #22762       +/-   ##
===========================================
+ Coverage   59.77%   83.03%   +23.25%     
===========================================
  Files         596      596               
  Lines       69820    69822        +2     
  Branches    10883    10883               
===========================================
+ Hits        41737    57976    +16239     
+ Misses      25575     8998    -16577     
- Partials     2508     2848      +340     

Flag	Coverage Δ
keras	82.85% <100.00%> (+23.07%)	⬆️
keras-jax	58.36% <0.00%> (?)
keras-numpy	54.23% <0.00%> (?)
keras-openvino	59.79% <100.00%> (+0.01%)	⬆️
keras-tensorflow	59.92% <0.00%> (?)
keras-torch	58.70% <0.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[hertschuh]

Thanks for implementing those.