fix(loaders): say "unsupported format" for .xls, .ods and other binary files (#1018)

[minutechreview]

Summary

faircode audit on an .xls or .ods file used to fail with a confusing message about --encoding and invalid UTF-8, because the file was read as text. It now says which format is not supported and what to do:

error: could not read dataset data.xls: unsupported format .xls: save it as .xlsx or .csv

The exit code is still 2.

In faircode/loaders.py, read_table now raises a ValueError for .xls, .ods, .feather, .pkl, .zip and .7z, before falling back to the delimited-text reader. The check only applies when the file exists, so a missing file still reports "file not found". The suffix check ignores case (.XLS too). Unknown text extensions such as .dat are still sniffed as before.

Tests are in tests/test_loaders.py (each unsupported suffix, uppercase, missing file, unknown text extension) and tests/test_cli.py (exit code 2, the message names the format, no --encoding hint).

The web UI already rejects anything that is not csv, tsv, json or xlsx, so I left the JS alone.

Type

• Audit
• Explainer
• Bug fix
• Other

---

Linked issue

Closes #1018

Checks: pytest tests --ignore=tests/browser gives 707 passed, 41 skipped (scipy installed). ruff check and scripts/check_em_dash.py pass.

[vercel]

@minutechreview is attempting to deploy a commit to the yashkewlani2020-gmailcom's projects Team on Vercel.

A member of the Team first needs to authorize it.

[github-actions]

@yakew7 @ahmdkaml - new PR touching a path you own, please review.