I built a Python CLI for Parquet interoperability. It runs a given Parquet file through selected readers and shows where it fails with a command. Also uses Hypothesis for fuzzing. Generates tables, runs through a matrix of engines and minimizes the found failures. You can also create a schema beforehand and fuzz within it.I believe it can be useful for Parquet engine maintainers and people with multiple Parquet engines involved in their pipeline.Already found and filed two issues with it, one from scanning Apache s own corpus and the other from fuzzing:- https://github.com/pola-rs/polars/issues/28787 - https://github.com/duckdb/duckdb/issues/24731repo: https://github.com/sovsparrow/parquity