Metadata-Version: 2.4
Name: repark
Version: 0.6.0
Requires-Dist: pyarrow>=25.0.0
Requires-Dist: pydantic>=2.10,<3
Requires-Dist: numpy>=1.26 ; extra == 'ml-ext'
Requires-Dist: pandas>=2.1 ; extra == 'ml-ext'
Requires-Dist: xgboost>=2.0 ; extra == 'ml-ext'
Requires-Dist: lightgbm>=4.0 ; extra == 'ml-ext'
Requires-Dist: scikit-learn>=1.3 ; extra == 'ml-ext'
Requires-Dist: numpy>=1.26 ; extra == 'numpy'
Requires-Dist: pandas>=2.1 ; extra == 'pandas'
Requires-Dist: polars>=1.0 ; extra == 'polars'
Provides-Extra: ml-ext
Provides-Extra: numpy
Provides-Extra: pandas
Provides-Extra: polars
Summary: Near-drop-in PySpark API on a pure-Rust, no-JVM Apache Iceberg engine.
License: Apache-2.0
Requires-Python: >=3.12
Description-Content-Type: text/markdown; charset=UTF-8; variant=GFM

# repark

A near-drop-in **PySpark** API on a pure-Rust, **no-JVM** Apache Iceberg engine.

```python
from repark import ReparkSession   # was: from pyspark.sql import SparkSession

# Drop-in alias for existing scripts:
# from repark import SparkSession
```

Compute runs in Rust (Apache DataFusion + iceberg-rust + Arrow); data crosses the Python boundary
as Apache Arrow, zero-copy. See the [repository root](../../README.md) for architecture and status.

