feat: add multilingual glossary support (backend + frontend types)

Backend:
- Add source_language column to glossaries table
- Add translations JSON column to glossary_terms table
- Alembic migration for schema changes
- format_glossary_for_prompt now language-aware: extracts correct
  translation per target language, falls back to EN reference for
  templates with only FR→EN data
- CRUD routes accept/return source_language and translations
- Pydantic schemas updated

Frontend:
- Types updated: GlossaryTerm now has translations: Record<string, string>
- Glossary/GlossaryListItem now have source_language
- Added SUPPORTED_LANGUAGES constant (13 languages)

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

routes/glossary_routes.py

@@ -46,6 +46,7 @@ def _format_term(term: GlossaryTerm) -> dict:
         "id": term.id,
         "source": term.source,
         "target": term.target,
+        "translations": term.translations or {},
         "created_at": term.created_at.isoformat() if term.created_at else None,
     }
 
@@ -55,6 +56,7 @@ def _format_glossary(glossary: Glossary) -> dict:
     return {
         "id": glossary.id,
         "name": glossary.name,
+        "source_language": glossary.source_language,
         "terms": [_format_term(t) for t in glossary.terms] if glossary.terms else [],
         "created_at": glossary.created_at.isoformat() if glossary.created_at else None,
         "updated_at": glossary.updated_at.isoformat() if glossary.updated_at else None,
@@ -103,6 +105,7 @@ async def create_glossary(
             glossary = Glossary(
                 user_id=user.id,
                 name=body.name,
+                source_language=body.source_language,
                 created_at=datetime.now(timezone.utc),
                 updated_at=datetime.now(timezone.utc),
             )
@@ -112,6 +115,7 @@ async def create_glossary(
                     glossary=glossary,
                     source=term_data.source,
                     target=term_data.target,
+                    translations=term_data.translations or {},
                     created_at=datetime.now(timezone.utc),
                 )
                 session.add(term)
@@ -180,6 +184,7 @@ async def list_glossaries(
                 GlossaryListItem(
                     id=g.id,
                     name=g.name,
+                    source_language=g.source_language or "fr",
                     terms_count=len(g.terms) if g.terms else 0,
                     created_at=g.created_at,
                 )
@@ -331,6 +336,9 @@ async def update_glossary(
             if body.name is not None:
                 glossary.name = body.name
 
+            if body.source_language is not None:
+                glossary.source_language = body.source_language
+
             if body.terms is not None:
                 # Delete existing terms
                 session.query(GlossaryTerm).filter(
@@ -343,6 +351,7 @@ async def update_glossary(
                         glossary_id=glossary.id,
                         source=term_data.source,
                         target=term_data.target,
+                        translations=term_data.translations or {},
                         created_at=datetime.now(timezone.utc),
                     )
                     session.add(term)

routes/translate_routes.py

@@ -915,10 +915,13 @@ async def _run_translation_job(
 
         # Story 3.10: Retrieve and format glossary terms for LLM prompt
         glossary_terms = None
+        glossary_source_lang = "fr"
         if glossary_id and user_id:
             try:
-                glossary_terms = get_glossary_terms(glossary_id, user_id)
-                logger.info(f"Job {job_id}: Loaded {len(glossary_terms)} glossary terms")
+                glossary_data = get_glossary_terms(glossary_id, user_id)
+                glossary_terms = glossary_data["terms"]
+                glossary_source_lang = glossary_data.get("source_language", "fr")
+                logger.info(f"Job {job_id}: Loaded {len(glossary_terms)} glossary terms (source: {glossary_source_lang})")
             except GlossaryNotFoundError as e:
                 tracker.set_error(str(e))
                 logger.error(f"Job {job_id}: Glossary error - {e}")
@@ -940,7 +943,10 @@ async def _run_translation_job(
             effective_prompt = custom_prompt
 
         # Build the full prompt combining effective prompt and glossary
-        full_prompt = build_full_prompt(effective_prompt, glossary_terms)
+        full_prompt = build_full_prompt(
+            effective_prompt, glossary_terms,
+            source_lang=glossary_source_lang, target_lang=target_lang,
+        )
 
         translation_provider = None
         _p = provider.lower()