Use custom user agent string at all times #1
Conversation
…ssion or SageMaker client
src/sagemaker/session.py
Outdated
|
|
||
| SDK_VERSION = pkg_resources.require('sagemaker')[0].version | ||
| OS_NAME = platform.system() |
There was a problem hiding this comment.
If no OS name can be resolved, platform.system() will return an empty string. Can this be something like:
OS_NAME = platform.system() or 'UnresolvedOS'
Same for OS_VERSION
There was a problem hiding this comment.
done. Also moved everything to a new module.
src/sagemaker/session.py
Outdated
| botocore_session.user_agent_extra = 'Boto3/{} Botocore/{}'.format(boto3.__version__, | ||
| botocore_session.user_agent_version) | ||
| self.sagemaker_client = sagemaker_client or self.boto_session.client('sagemaker') | ||
| self.sagemaker_client._client_config.user_agent = user_agent_string |
There was a problem hiding this comment.
Is there a reason we don't preserve the original user agent string as well. Why not just append the SageMaker string to the end?
There was a problem hiding this comment.
The default user agent string is 'Boto3/1.4.8 Python/3.6.3 Darwin/15.6.0 Botocore/1.8.6', which is why I just added the Boto and Botocore versions to the end. It needs to start with 'AWS-SageMaker' and be followed by 'language/version OS/version' for parsing. Appending the original user agent string here is a good call, though.
Add Pytorch estimator and model
…ws#1) * Added client libraries notebooks + kmeans notebooks * Fixed delete endpoint cell in mxnet notebook * Removed leftover debugging cells from kmeans high level * Reset IOBytes buffer after writing in KMeans low-level * kmeans lowlevel notebook fixes * Move examples into im-python-sdk directory
Sync with AWS repo
merge from master branch
documentation: pytorchddp doc update
The previous implementation allowed for our user agent string to be overridden if the user supplied a boto3 session, SageMaker client, or SageMaker runtime client.