Version 0.9.0

Version 0.9.0 (2026-07-03)

• Fixed a bug where empty string values, e.g. for the watermark, were not processed correctly resulting in error and shifts in columns in the generated TSV file cf. #29. Thanks @Athemis
• Fixed a bug where having different order of keys in the yaml definition of a metadata block caused the columns in the generated TSV file to be misaligned. #31 Thanks @Athemis
• Made the parent keyword optional. While no parent had to be specified, the entry still had to be present in the yaml file. Now empty parent entries can be omitted.

Full Changelog: 0.8.1...0.9.0

Version 0.8.1

Version 0.8.1 (2025-01-05)

• Fixed a bug where changing a lint level to WARNING using the --warn flag resulted in an error.
• Added tests for lint level flags.

Version 0.8.0

Lint Changes

• Added lints (e005 and e006) to detect nested compound metadata.
• Demote e004 to be a warning by default (Fixes #23). Since Dataverse 6.5, trailing whitespaces are trimmed.
• Tweak unique_title (b003) lint to allow identical titles as long as they are part of different compound field. (Fixed #22)

CLI Changes

• Add --error flag that allows warnings to be treated as errors instead.

Other Changes

• Fixed multi version tests not working properly
• Dropped support for Python versions below 3.12 for better string formatting (cf. PEP 701)
• Line and column numbers are now reported for "leaves" in yaml input files.

Version 0.7.0

• Fixed bug where converting files without lint violations would result in an error (#16). Thanks @Athemis
• Fixed bug where non-string watermarks caused an error (#19). Thanks @Athemis
• Added minimal test for conversion function.
• Add tests to PRs.
• Add lint b003 that checks if all titles within the DatasetField block are unique.

Version 0.6.0

• Added a CHANGELOG file.
• Switched yaml input to use the ruamel.yaml round trip parser. This parser is also "safe", i.e. does not allow code execution.
• Added line numbers and columns to error messages where possible. For yaml files, these are taken from the ruamel.yaml round trip parser. For tsv files, the column names within the specified block are reported.
• Added --strict flag to the convert sub-command that prevents converting files that produced warnings.
• The rules keys_valid and keys_unique now correctly report their name in output.
• Added integration tests checking tsv files.
• Improved user feedback for parameters dealing with rule names, e.g. --skip.

Version 0.5.0

• Improved documentation and user feedback
• Added option to check multiple files. You can now specify multiple input file paths for the check sub-command and those file paths can be glob patterns.
• Added suggestions for fixes to some lints.

Please note that the convert sub-command currently only works with one input file.

Version 0.4.0

• Added option to skip lints or reduce their severity to warning.
• Added first suggestions for fixes, e.g. minor typo detection with string matching.
• Added RULES.md to have a human readable documentation of employed lints.

Version 0.3.0

This release adds:

• Lints to detect trailing spaces
• Internal preparations to support multiple files
• Minor documentation improvements