azureml-fsspec

What it is and what it does

azureml-fsspec registers a custom fsspec filesystem handler for azureml:// URIs, bridging Azure Machine Learning datastores and Python's data-processing ecosystem. It lets you use fsspec's abstraction layer to read and write files directly from Azure ML datastores without manual credential management or intermediate downloads.

The package wraps fsspec around azureml-dataprep and pytz, handling URI parsing, authentication (including service principal and managed identity), and registry lookups. It is maintained by Microsoft, actively updated for recent Python versions (3.12, 3.13 support added in 1.4.0), and designed for data scientists and ML engineers working within Azure ML environments.

Use it for:

• Access training data from an Azure ML datastore via fsspec without downloading files locally.
• Read parquet or CSV files from Azure ML datastores for distributed processing.
• Integrate Azure ML datastores into existing fsspec-based workflows without rewriting data-loading logic.
• Automate data pipeline steps that reference Azure ML datastore paths using fsspec APIs.
• Support hybrid ML workflows where Azure ML is the central data repository.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides a filesystem interface to Azure Machine Learning datastores via fsspec, allowing you to access data files using azureml:// URIs with standard data-processing libraries.

Yes, if you work within Azure Machine Learning and need fsspec access to datastores. Install friction is low and maintenance is active. The proprietary license requires review before use in commercial or redistributed products. No known security vulnerabilities as of the latest release.

Install

azureml-fsspec on PyPI

pip install azureml-fsspec

uv add azureml-fsspec

poetry add azureml-fsspec

Installing azureml-fsspec

Before you install

Low install friction with a pure-Python wheel. Maintenance is active; the latest release (1.4.0) is recent and addressed setuptools compatibility issues in azureml-dataprep, a key runtime dependency.

License in practice

Licensed under a proprietary Microsoft license (Proprietary https://aka.ms/azureml-preview-sdk-license). License treatment is unclear; review the linked license terms before use in commercial or redistributed contexts.

Quickstart

pip install azureml-fsspec

import fsspec
fs = fsspec.filesystem('azureml')
files = fs.ls('azureml://subscriptions/{sub_id}/resourcegroups/{rs_group}/workspaces/{ws}/datastores/workspaceblobstore/paths/myfolder')

Requires Python 3.10–3.13 and valid Azure Machine Learning subscription, resource group, workspace, and datastore credentials.

Verify before relying

• Whether the proprietary license permits commercial use or redistribution without explicit Microsoft agreement.
• Authentication flow details beyond mlclient support (e.g., service principal, managed identity setup steps).
• Performance characteristics or limits when working with large datasets via fsspec.

Package facts

Evidence: azureml_fsspec-1.4.0-py3-none-any.whl

Keywords: file-system, dask, azure

Tags

More Scientific/Engineering packages