Installation¶

Requirements¶

Python 3.12 or later

Install¶

pip install datasluice

Or with uv:

uv add datasluice

Optional dependencies¶

DataSluice keeps its core dependency footprint small. Some features require optional packages:

Feature	Install command
pandas support	`pip install datasluice[pandas]`
Polars support	`pip install datasluice[polars]`
dlt integration	`pip install datasluice[dlt]`
DuckDB support	`pip install datasluice[duckdb]`
Apache Airflow provider	`pip install apache-airflow-providers-datasluice`
Parquet reading	`pip install datasluice[parquet]`
XLSX reading	`pip install datasluice[xlsx]`
All extras	`pip install datasluice[all]`

Airflow integration ships as a separate distribution called apache-airflow-providers-datasluice and imports from the airflow.providers.datasluice namespace. It is released and versioned independently of the core datasluice package.

Connector extras¶

The base installation already carries the full connector contract surface: typed models, sync and async client Protocols, capability profiles, reference fakes, and the public compliance runner.

Installable named connector extras (ckan, udata, socrata, and the all-connectors convenience extra) belong to Phase 2 packaging work (PACK-01/PACK-02) and their dependency tables. This release does not add or advertise them, and the library never downloads dependencies at runtime — missing capabilities explain the required install instead.

Verify installation¶

datasluice --version

Development install¶

See the README for local development setup instructions.