[katerina20]

Part of updating the SDK to the current Crowdin API docs. This PR covers the Glossaries and Translation Memory APIs.

Glossaries

• Concordance search: SearchConcordanceRequest gets expressions, which the API now requires. expression is deprecated.
• Terms and concepts:
  • Status.DRAFT added.
  • Enterprise custom fields added to Term, AddTermRequest and Concept.
• Glossaries: isShared added to Glossary and AddGlossaryRequest.
• Update concept:
  • New method updateConceptWithRequest(glossaryId, conceptId, UpdateConceptRequest) sends only the fields the PUT endpoint accepts. It has a separate name, not an overload, so existing calls stay unambiguous.
  • The original updateConcept(glossaryId, conceptId, Concept), which sends the Concept response model, is deprecated. Its signature and behaviour are unchanged.
• Export:
  • The filters are now sent with the documented values: upper-case (NOT_RECOMMENDED, FULL_FORM, …), and UD tags for parts of speech (NOUN, ADJ, X, …).
  • New fields carry them: termStatuses, termTypes, termGenders, termPartsOfSpeech.
  • New filters: text, caseSensitive, searchStrict, searchFullMatch, dateFrom, dateTo.
• Export status: added the missing attributes (text, caseSensitive, searchFullMatch, searchStrict, status, partOfSpeech, type, gender, authorId, dateFrom, dateTo).
• listGlossaries: new filter query parameter.
• Deprecated: ListTermsParams.translationOfTermId, which the API deprecates.

Translation Memory

• Concordance search: expressions added, expression deprecated.
• Translation memories: languageId, isShared and webUrl added to TranslationMemory; isShared added to AddTranslationMemoryRequest.
• New overloads:
  • listTms(ListTmsParams) adds filter and takes a Long userId. The existing method takes Integer, so the change is an overload.
  • listTmSegments(tmId, ListTmSegmentsParams) adds the croql filter.
• Not added: the TM segment record endpoints (POST/PATCH/DELETE …/segments/{segmentId}/records), because the API deprecates them in favour of Edit TM Segment.

Backward compatibility

No public method, field or class was removed, renamed or retyped.

• Glossary export filters (behaviour correction):
  • Before, the export request sent term-style values ("preferred", "full form", "noun"), which the export endpoint doesn't document.
  • The deprecated setStatuses/setTypes/setGenders/setPartsOfSpeech now convert term enums to the documented values.
  • Part of speech other maps to X. The published API uses other for glossary terms and X for glossary export, and Universal Dependencies defines X as "other". This hasn't been confirmed by a live Crowdin API test. A dedicated unit test (exportGlossaryRequestDeprecatedSetterMapsOtherPartOfSpeechToXTest) checks that PartOfSpeech.OTHER is sent as "X" and maps back.
  • The deprecated getters convert back. They can't represent the export-only values SYM and COMMON, so they leave them out.
• Source compatibility: updateConcept(id, id, null) still compiles, because the new method has a different name. A compatibility test (updateConceptWithNullRequestCompatibilityTest) covers it.
• equals/hashCode/toString include the new fields.

Testing

• ./gradlew build: 647 tests, 0 failures, on both JDK 17 and JDK 8.
• ./gradlew javadoc on JDK 8: no errors.
• New and updated mock tests and fixtures cover:
  • concordance expressions (both APIs);
  • updateConceptWithRequest with the new request, and updateConcept(id, id, null) compiling and calling the original endpoint;
  • the export request JSON (new fields, and conversion from the deprecated setters);
  • the export status attributes;
  • fields, DRAFT, isShared, webUrl, languageId;
  • listTms(ListTmsParams) and listTmSegments with croql.

🤖 Generated with Claude Code