iowarp/clio-data
The clio-data agent's plugin: its mission, assess-data-quality (what a variable states, checks, comparisons, visual checks) and convert-scientific-data (the HDF5/ADIOS2 to Parquet dtype map, skip-or-ask rules, and a bundled converter that reads back and verifies rows, nulls, NaNs, dtypes and per-column checksums).
Judge whether variables in HDF5, netCDF-4, ADIOS2 or Parquet files can be trusted and compared - what a variable states about itself, the checks to run with bounded reads, how to compare two variables, and which visual checks reveal problems.
Convert HDF5 (including netCDF-4) datasets or ADIOS2 BP4/BP5 variables to Parquet with an explicit dtype map, decide when a field is skipped and when to ask, and prove the result with row, null, NaN, dtype and per-column checksum checks. Bundles convert_to_parquet.py, which writes, reads back and verifies.