Improve error handling for invalid date input

[loganthomas]

Improve error messages for invalid dates

Closes #94

Summary

When utt encounters invalid date input—either from command-line arguments or from entries in the log file—it now displays a user-friendly error message instead of a Python traceback.

Problem

Previously, invalid date usage produced unhelpful tracebacks:

$ utt report 324-4242-43
Traceback (most recent call last):
  File "/opt/homebrew/bin/utt", line 8, in <module>
    sys.exit(main())
  ...
  File "/opt/homebrew/lib/python3.9/site-packages/utt/components/report_args.py", line 74, in parse_absolute_date
    return datetime.datetime.strptime(datestring, "%Y-%m-%d").date()
  ...
ValueError: time data '324-4242-43' does not match format '%Y-%m-%d'

Solution

Added a simple UttError exception class for user-facing errors. Invalid dates now produce clean, actionable messages.

Examples

Invalid date argument

$ utt report invalid-date
error: Invalid date: invalid-date (expected YYYY-MM-DD)

Invalid month argument

$ utt report --month invalid
error: Invalid month: invalid (expected YYYY-MM)

Invalid date in log file

$ utt report
error: Invalid date at line 14: 2025-27-27 17:00 misc: testing

Entries not in chronological order

$ utt report
error: Line 14 not in chronological order: 2025-11-26 17:00 misc: testing

Changes

• utt/components/entries.py: Added UttError exception class; catch ValueError from date parsing and convert to user-friendly error
• utt/components/report_args.py: Wrap strptime calls to raise UttError on invalid input
• utt/__main__.py: Catch UttError at top level and print clean error message

Tests

Added tests for the new error handling:

• test_invalid_date_raises_utt_error
• test_invalid_month_raises_utt_error
• test_invalid_date_raises_value_error (for log file entries)

[larose]

Thanks for the PR! It looks good overall, but I've left a few comments.

Could you also add a single integration test to confirm that UttError is handled at the top level and formats a clean message without a stack trace? (We only need to do this for one case)

Also, please add yourself to the contributors list in the readme.

[larose]

Thanks, this looks good overall and should be the last round of feedback. I left a minor comment (plus one I apparently forgot to submit in my first review).

[larose]

@loganthomas some checks are failing, I'll merged once they are fixed.

[larose]

Thank you for your contribution @loganthomas! This is now available in version 1.32. Apologies for the delay.