A fast toolkit for “gleaning” structured facts from messy files. Turn folders of CSV/JSON/Excel/Parquet into tidy, typed dataframes with validated metadata and simple pipelines you can run in code or from the CLI.
- Smart column normalization – map vendor/legacy headers to your canonical schema.
- Lightweight pipelines – compose steps like
load → normalize → validate → enrich → write. - Schema & units checks – optional soft/hard validation of required fields and units.
- Fast I/O – pandas/pyarrow based; reads CSV, XLSX, JSON, Parquet.
- Declarative rules – YAML/JSON rules for synonyms, dtypes, units, computed fields.
- CLI & Python API – run one-off or batch jobs locally or in CI.
- Extensible – drop-in custom steps and validators.
Clone the repository. In the directory, install the package in editable mode.
pip install -e .