Fix the default value of side in torch.searchsorted

[lancerts]

Fixes #119999, currently the doc shows the default value of side = "left"

while the implementation gives the default value of side = c10::nullopt.

• fix the torch doc such that the default value of side is None.

• fix the comment in cpp such that the default value of side is None.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/120066

• 📄 Preview Python docs built from this PR
• 📄 Preview C++ docs built from this PR
• ❓ Need help or want to give feedback on the CI? Visit the bot commands wiki or our office hours

Note: Links to docs will display an error until the docs builds have been completed.

✅ You can merge normally! (5 Unrelated Failures)

As of commit 874f38a with merge base 62e5840 ():

FLAKY - The following jobs failed but were likely due to flakiness present on trunk:

• pull / linux-focal-cuda12.1-py3.10-gcc9-sm86 / test (default, 1, 5, linux.g5.4xlarge.nvidia.gpu) (gh)
  inductor/test_torchinductor_dynamic_shapes.py::TestInductorDynamicCUDA::test_full_cuda
• pull / linux-focal-cuda12.1-py3.10-gcc9-sm86 / test (default, 5, 5, linux.g5.4xlarge.nvidia.gpu) (gh)
  dynamo/test_trace_rules.py::TraceRuleTests::test_torch_name_rule_map_updated
• trunk / linux-focal-cuda12.1-py3.10-gcc9 / test (nogpu_AVX512, 1, 1, linux.2xlarge) (gh)
  dynamo/test_trace_rules.py::TraceRuleTests::test_torch_name_rule_map_updated
• trunk / linux-focal-cuda12.1-py3.10-gcc9 / test (nogpu_NO_AVX2, 1, 1, linux.2xlarge) (gh)
  dynamo/test_trace_rules.py::TraceRuleTests::test_torch_name_rule_map_updated
• trunk / macos-12-py3-arm64 / test (default, 1, 3, macos-m1-12) (gh)
  dynamo/test_trace_rules.py::TraceRuleTests::test_torch_name_rule_map_updated

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[lezcano]

This has BC-breaking implications so we won't be able to land it as-is. Can you instead change the docs?

[malfet]

This feels wrong, we should update documentation instead see #120086

But running the test to see if something will fail (as I really hope we test for it)

[lancerts]

This feels wrong, we should update documentation instead see #120086

But running the test to see if something will fail (as I really hope we test for it)

@malfet Thanks for the suggestion, totally agree. Modified the change to doc only.

However, changed to side=None instead of side: Optional[str] = None.

Or, we have to change out_int32, right, out and sorter in the docs as explicit optional to.
Seems other methods in torch docs don't mark the args explicitly optional, for example,

https://github.com/pytorch/pytorch/blob/main/torch/_torch_docs.py#L13865

Also need to change the comment in aten/src/ATen/native/Bucketization.cpp : P. Thanks
cc @lezcano

[malfet]

@lancerts if you are doing it anyway, do you mind just go over all functions in say Bucketization and update docs accordingly and document how you've verified the behavior in PR description (as right now it contradicts to what PR is actually doing)

[lancerts]

@lancerts if you are doing it anyway, do you mind just go over all functions in say Bucketization and update docs accordingly and document how you've verified the behavior in PR description (as right now it contradicts to what PR is actually doing)

@malfet sure.
1, Did a pass for bucketization cpp, h, torch doc, yaml, etc. There are no issues since side is not a function arg for the bucketize function, see code.

2, Updated the PR description to align with the changes.

3, Did a pass in torch doc. It seems the convention of the optional arg is to mark it as optional in the description (instead of in function API), for example in the pooling doc. Therefore, will follow this convention and make the change consistent with other functions in torchdoc. Thanks

[lancerts]

@malfet please take another look, comments addressed, thanks

[malfet]

Looks good, thank you for the updates

[lancerts]

@pytorchbot merge

[pytorchmergebot]

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here