Every input path in faircode/cli.py (profile and compare) expects a real file on disk - .csv, .tsv, .xlsx, .json, .parquet, dispatched by extension. There's no way to pipe data in, e.g. cat data.csv | faircode profile - or curl ... | faircode profile -, which is a common Unix CLI convention and would let the tool slot into a shell pipeline without writing a temp file first.
Proposed scope (keep it small): support - as the path argument for CSV/TSV input specifically (reading sys.stdin via pd.read_csv(sys.stdin, ...)), since format-sniffing-by-extension has no extension to go on for a stdin stream anyway - .xlsx/.parquet/.json support over stdin can be a follow-up if there's demand, not a blocker for this one.
Every input path in
faircode/cli.py(profileandcompare) expects a real file on disk -.csv,.tsv,.xlsx,.json,.parquet, dispatched by extension. There's no way to pipe data in, e.g.cat data.csv | faircode profile -orcurl ... | faircode profile -, which is a common Unix CLI convention and would let the tool slot into a shell pipeline without writing a temp file first.Proposed scope (keep it small): support
-as the path argument for CSV/TSV input specifically (readingsys.stdinviapd.read_csv(sys.stdin, ...)), since format-sniffing-by-extension has no extension to go on for a stdin stream anyway -.xlsx/.parquet/.jsonsupport over stdin can be a follow-up if there's demand, not a blocker for this one.