aws / aws/sagemaker-python-sdk

AsyncPredictor fails if name is None, despite it being the default

Open
#4,774 1 comment 0 reactions 0 assignees View on GitHub
component: utility apis type: bug
Dominant language
Python
Stars
2.3k
Forks
1.3k
Avg merge
1d 22h
Merged PRs (30d)
35

Description

**Describe the bug**

The `name` argument of the `AsyncPredictor.__init__.py` is optional with default=None

https://github.com/aws/sagemaker-python-sdk/blob/b535ed87ae55fd1020ea1d7c7c9cd7e3a068a683/src/sagemaker/predictor_async.py#L34

But calling `async_predictor.predict(data=...)` requires name to be non-null, i.e.` _upload_data_to_s3 ` calls `name_from_base` on `self.name` which fails if name is None (i.e. the default)

https://github.com/aws/sagemaker-python-sdk/blob/b535ed87ae55fd1020ea1d7c7c9cd7e3a068a683/src/sagemaker/predictor_async.py#L171

**To reproduce**

```
from sagemaker.predictor import Predictor
from sagemaker.predictor_async import AsyncPredictor

predictor = Predictor(endpoint_name=endpoint_name, sagemaker_session=sagemaker_session)
async_predictor = AsyncPredictor(predictor)

result = async_predictor.predict(data=request_body)
```
TypeError: 'NoneType' object is not subscriptable

**Expected behavior**
The name field of `AsyncPredictor` should have a "non-None" default (a guid, or the endpoint name?)

**System information**
A description of your system. Please provide:
- **SageMaker Python SDK version**: '2.224.2'
- **Framework name (eg. PyTorch) or algorithm (eg. KMeans)**: PyTorch / Triton (custom inference container)
- **Framework version**:
- **Python version**:
- **CPU or GPU**:
- **Custom Docker image (Y/N)**: Y

**Additional context**
Add any other context about the problem here.

Contributor guide

Open the contributing guide

Research direction

Start in src/sagemaker/predictor_async.py, especially AsyncPredictor.__init__ and _upload_data_to_s3 near the referenced lines. Reproduce the failure with AsyncPredictor(Predictor(...)) and predict(data=...), then determine an appropriate non-None default for name. Done means the documented default path no longer raises the reported TypeError.

Written by the indexing model from the issue text.

Assessment

Tech stack
aws, python
Domain
machine-learning
Issue type
Bug
Difficulty
2/5
Estimated time
1-3 hours
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.