This repository is work in progress.
Roadmap:
- Read parquet file (v)
- Have a sample UDF that accesses the file data with custom logic (v)
- Add onnxruntime ()
- Predict using UDF + onnxruntime ()
- Package code as Python extension (bindings with PyO3) ()
- Write alternatives, such as pure pandas Python, pure polars Python etc ()
- Benchmark ()