Sagemaker-python-sdk: Processing job: Cannot specify s3 uri as code location

[elvinx]

According to the docs, one can specify either S3 URI or a local path as code location while running processing jobs with SageMaker:

This seems not to be the case.
When you specify the s3 path, SageMaker Python SDK still tries to create a default S3 bucket and upload the code to it.

I dug into the code base and can see that there is no check if the given path is an S3 path or not and default behavior is always to expect a local path.
https://github.com/aws/sagemaker-python-sdk/blob/master/src/sagemaker/processing.py#L373

Expected behavior:
Allow S3 paths as a code location as stated in the docs.

[ajaykarpur]

Hi @elvinx, thanks for bringing this to our attention. We'll address this and release a fix as soon as we're able.

[andremoeller]

Thanks for reporting this @elvinx .

I sent a pull request with the fix and tests: https://github.com/aws/sagemaker-python-sdk/pull/1167

[elvinx]

@andremoeller Thank you!

[andremoeller]

Hey @elvinx ,

This fix has been released: https://github.com/aws/sagemaker-python-sdk/releases/tag/v1.47.0

You can install this newest release into a notebook instance by:

1. Running !pip install -U --user sagemaker
2. Restarting the kernel.

I'm closing this issue. Please feel free to reach out or reopen this if you run into issues. Thanks!