Skip to content

Latest commit

 

History

History
78 lines (51 loc) · 2.72 KB

CHANGELOG.md

File metadata and controls

78 lines (51 loc) · 2.72 KB

Version 3.5.0

  • Added KRADFILE/RADKFILE
  • Minor fixes in documentation
  • Bugfix/workaround for invalid format of cross-references in JMdict

Version 3.4.0

  • Added Kanjidic
  • Publishing NPM packages: @scriptin/jmdict-simplified-types, @scriptin/jmdict-simplified-loader

Version 3.3.1

  • Bugfix #24: sense.appliesToKanji and sense.appliesToKana are set to ["*"] (meaning "applies to all/any") by default. This was the logic of the original XQuery converter

Version 3.3.0

  • Added missing field descriptions in readme
  • Added one more Xref type variant
  • Added Node.js modules for JSON schema validation of dictionary JSON files
  • Updated the documentation

Version 3.2.1

  • Bugfix #22: add missing "misc" tags on sense elements in JMdict

Version 3.2.0

  • Converter is fully rewritten from XQuery to Kotlin - XQuery parser was limited by memory and couldn't process large files
  • Converter now supports language filtering, i.e. can produce language-specific versions (not only English)
  • Converter is now capable of generating multiple output files in parallel
  • Fixed an error with non-case-sensitive XML entities
  • Documentation is updated to include TypeScript types, cleaner explanations, missing fields, etc.

Version 3.1.0

  • Added conversion for g_type attribute on gloss elements (words->sense->gloss->type)

Version 3.0.1

  • BaseX is no longer required as an external binary, now it's used internally as a build script dependency
  • Latest source files (new data in JMnedict)

Version 3.0.0

  • ⚠️ Change in the format: entries' id fields in all dictionaries are now strings (were numbers), this is to prevent BaseX converting long numbers to exponential notation (like 1.001e+8). ID values themselves haven't changed
  • Migrated from Zorba to BaseX, JSON output formatting changed as a result
  • Migrated from bash script to Gradle build
  • JMnedict is now being processed in batches and then concatenated. This way the conversion process doesn't consume all system memory
  • Distribution files now have version is their names

Version 2.0.0

  • #7: Rename fields: jm(ne)dict-date -> dictDate, jm(ne)dict-revisions -> dictRevisions

Version 1.2.0

  • #6: JMnedict included
  • Latest versions of both dictionaries
  • Improved build script
  • Updated docs
  • Many minor fixes

Version 1.1.1

  • #3: Fixed duplicated XML entities text bug
  • #4: Removed binary files from the repository

Version 1.1

  • #2: Add 'jmdict-date' and 'jmdict-revisions' fields
  • Add 'version' field
  • Fix a typo

Version 1.0

Initial release, with full English version and shorter English version with only common kanji/kana entries.