v0.10.2

Important Changes

• The Thai (th), Simplified Chinese (zh-hans), and Traditional Chinese (zh-hant) models have been updated.
• An optional wrap argument has been added to Python's budoux.html_processor.resolve() and Parser.translate_html_string(), as well as Java's HTMLProcessor.resolve() and Parser.translateHTMLString(), allowing you to control whether the returned string includes a wrapping SPAN element.
• Expose Parser.PARENT_CSS_STYLE ("word-break: keep-all; overflow-wrap: anywhere;").

What's Changed

• ci: consolidate language workflows and unify runtime versions by @tushuhei in #1468
• refactor(tests): replace sys.path.insert and unittest.main boilerplate with pytest pythonpath by @tushuhei in #1461
• feat(model): retrain zh-hans and zh-hant models on Unicode corpora by @tushuhei in #1467
• 🧪 test(java): add unit tests for loadDefault*Parser functions by @tushuhei in #1463
• ci: avoid canceling main runs and harden-runner errors on non-Linux by @tushuhei in #1469
• Update expected results in ParserTest for consistency by @tushuhei in #1470
• 🧪 test: add test for load_default_thai_parser by @tushuhei in #1462
• ci(java): replace googlejavaformat-action with spotless-maven-plugin by @tushuhei in #1471
• feat(python, java): support translating HTML strings without wrapping span by @tushuhei in #1474
• feat(demo): replace BR checkbox with custom separator input and add Maven link by @tushuhei in #1475
• feat(model): retrain th model on Wisesight corpus with LogitBoost by @tushuhei in #1473
• bump version to 0.10.2 by @tushuhei in #1476

Full Changelog: v0.10.1...v0.10.2