Greenmask 0.2.26

Changes

• feat: add --diff-mode=summary to the validate command. It compares the same rows as --diff but prints only counts per table and column — changed, unchanged and NULLs kept — with no original or transformed value and no primary key. This makes validate usable where production data may be read but must not be shown, such as kept CI logs, a shared terminal or an AI agent's transcript. Text and --format=json output are supported #495 #503
• feat: the summary warns when a transformed column keeps more than --diff-unchanged-threshold percent of its non-NULL values (default 90, 100 turns it off), which usually means a no-op transformer, and when a column changed although no transformer affects it. These are validation warnings: they print with --warnings or --strict, have a hash that can go into resolved_warnings, and --strict exits non-zero on an unresolved one. The hash ignores row counts, so it stays the same when --rows-limit, the data or the threshold change #503
• fix: make the gender chosen by RandomPerson reproducible across runs. The list of genders was built from a map, so its order changed between runs and the deterministic engine could pick a different gender, and so a different name, for the same input #502. Closes #501
• docs: move the documentation site to Docusaurus and drop MkDocs #498 #500 #504
• docs: add the Save and reset test environments with arcave page: restore a masked dump into a Docker Compose project once, save it as an arcave savepoint and reset the whole environment in one command #499
• docs: add make docs-serve and make docs-build to preview the documentation locally with the stock Docusaurus theme, and a CONTRIBUTING.md guide for contributors #505
• docs: correct the documented defaults of Hash function, RandomNumeric decimal, RandomString symbols and RandomPerson gender_mapping to match the code #493

note

--diff-mode=values stays the default, so the existing validate --diff output is unchanged. --data with --diff-mode=summary and without --diff is an error, because plain --data prints every value. Without --data the mode is ignored, so diff_mode: summary can stay in a shared config.

Full Changelog: v0.2.25...v0.2.26

Links

Feel free to reach out to us if you have any questions or need assistance:

• Discord
• Email
• Twitter
• Telegram [RU]
• DockerHub