DataRobot Storage

2.2.0 verified Fri May 01 auth: no python

A library providing reusable storage access (S3, GCS, Azure Blob, local filesystem) for DataRobot. Current version 2.2.0, compatible with Python >=3.9, <4.0. Released under the MIT license with quarterly updates.

pip install datarobot-storage

Common errors

error ImportError: No module named datarobot.storage ↓

cause Using the old dotted import path (pre-v2.0.0).

fix

Use from datarobot_storage import Storage instead of from datarobot.storage import Storage.

error TypeError: Storage() takes at least 1 argument (0 given) ↓

cause Calling Storage() without a backend argument (v2+).

fix

Provide a backend object: Storage(backend).

error datarobot_storage.exceptions.InvalidBackendError: The backend must be an instance of BaseBackend ↓

cause Passing a dict or string instead of a backend instance.

fix

Use a proper backend class: Storage(S3Backend(...)).

Warnings

breaking v2.0.0 renamed the main package from datarobot.storage to datarobot_storage. All imports must use underscores. ↓

fix Update imports: replace datarobot.storage with datarobot_storage.

breaking Storage constructor now requires a backend object instead of a dict configuration. ↓

fix Instantiate a backend explicitly (e.g., S3Backend(...)) and pass to Storage.

gotcha Environment variables for credentials are not automatically picked up; you must pass them explicitly unless using an IAM role or service account. ↓

fix Pass aws_access_key_id and aws_secret_access_key explicitly, or configure boto3 default session.

Imports

Storage
wrong
```
from datarobot.storage import Storage
```
correct
```
from datarobot_storage import Storage
```
Legacy package used old dotted path; renamed to underscore in v2

S3Backend

wrong

from datarobot_storage import S3Backend

correct

from datarobot_storage.backends import S3Backend

Backends are submodules, not top-level

Quickstart

Initialize an S3-backed storage and upload a file.

from datarobot_storage import Storage
from datarobot_storage.backends import S3Backend

backend = S3Backend(
    bucket='my-bucket',
    aws_access_key_id=os.environ.get('AWS_ACCESS_KEY_ID', ''),
    aws_secret_access_key=os.environ.get('AWS_SECRET_ACCESS_KEY', '')
)
storage = Storage(backend)
storage.upload('local/path.txt', 'remote/path.txt')