library · inferred from evidence
whylogs
A library that could be run under WSL as a workaround for the Windows dependency issue.
- Integer values in whylogs frequent_items show as floats (1.000000)
Logging a pandas DataFrame with an int64 column via why.log, then reading `frequent_items/frequent_strings`, returns values like '1.000000' instead of '1'. The column dtype is int64; the stringified frequent items show floats.
- pip install whylogs fails on Windows: No matching distribution for whylabs-datasketches
On Windows 10 with a conda Python 3.8 env, `pip install whylogs==0.4.8` fails with `ERROR: No matching distribution found for whylabs-datasketches==2.2.0b1`. The same install works on Linux. The protobuf import error was a red herring from