feat: Add external expert sources (Wikidata SPARQL + ConceptNet API)

- Add expert_cache.py: TTL-based in-memory cache (1 hour default)
- Add expert_source_service.py: WikidataProvider and ConceptNetProvider
  - Wikidata SPARQL queries for occupations with Chinese labels
  - ConceptNet API queries for occupation-related concepts
  - Random selection from cached pool
- Update schemas.py: Add ExpertSource enum (llm/wikidata/conceptnet)
- Update ExpertTransformationRequest with expert_source and expert_language
- Update router: Conditionally use external sources with LLM fallback
  - New SSE events: expert_source, expert_fallback
- Update frontend types with ExpertSource

🤖 Generated with [Claude Code](https://claude.com/claude-code)

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

backend/app/routers/expert_transformation.py

@@ -13,6 +13,7 @@ from ..models.schemas import (
     ExpertKeyword,
     ExpertTransformationCategoryResult,
     ExpertTransformationDescription,
+    ExpertSource,
 )
 from ..prompts.expert_transformation_prompt import (
     get_expert_generation_prompt,
@@ -20,6 +21,7 @@ from ..prompts.expert_transformation_prompt import (
     get_single_description_prompt,
 )
 from ..services.llm_service import ollama_provider, extract_json_from_response
+from ..services.expert_source_service import expert_source_service
 
 logger = logging.getLogger(__name__)
 router = APIRouter(prefix="/api/expert-transformation", tags=["expert-transformation"])
@@ -35,36 +37,98 @@ async def generate_expert_transformation_events(
         model = request.model
 
         # ========== Step 0: Generate expert team ==========
-        yield f"event: expert_start\ndata: {json.dumps({'message': '正在組建專家團隊...'}, ensure_ascii=False)}\n\n"
+        yield f"event: expert_start\ndata: {json.dumps({'message': '正在組建專家團隊...', 'source': request.expert_source.value}, ensure_ascii=False)}\n\n"
 
         experts: List[ExpertProfile] = []
+        actual_source = request.expert_source.value
 
-        try:
-            expert_prompt = get_expert_generation_prompt(
-                query=request.query,
-                categories=all_categories,
-                expert_count=request.expert_count,
-                custom_experts=request.custom_experts
-            )
-            logger.info(f"Expert prompt: {expert_prompt[:200]}")
+        # 決定使用哪種來源生成專家
+        use_llm = (
+            request.expert_source == ExpertSource.LLM or
+            request.custom_experts  # 有自訂專家時，使用 LLM 補充
+        )
 
-            expert_response = await ollama_provider.generate(
-                expert_prompt, model=model, temperature=temperature
-            )
-            logger.info(f"Expert response: {expert_response[:500]}")
+        if use_llm:
+            # LLM 生成專家
+            try:
+                expert_prompt = get_expert_generation_prompt(
+                    query=request.query,
+                    categories=all_categories,
+                    expert_count=request.expert_count,
+                    custom_experts=request.custom_experts
+                )
+                logger.info(f"Expert prompt: {expert_prompt[:200]}")
 
-            expert_data = extract_json_from_response(expert_response)
-            experts_raw = expert_data.get("experts", [])
+                expert_response = await ollama_provider.generate(
+                    expert_prompt, model=model, temperature=temperature
+                )
+                logger.info(f"Expert response: {expert_response[:500]}")
 
-            for exp in experts_raw:
-                if isinstance(exp, dict) and all(k in exp for k in ["id", "name", "domain"]):
-                    experts.append(ExpertProfile(**exp))
+                expert_data = extract_json_from_response(expert_response)
+                experts_raw = expert_data.get("experts", [])
 
-        except Exception as e:
-            logger.error(f"Failed to generate experts: {e}")
-            yield f"event: error\ndata: {json.dumps({'error': f'專家團隊生成失敗: {str(e)}'}, ensure_ascii=False)}\n\n"
-            return
+                for exp in experts_raw:
+                    if isinstance(exp, dict) and all(k in exp for k in ["id", "name", "domain"]):
+                        experts.append(ExpertProfile(**exp))
 
+                actual_source = "llm"
+
+            except Exception as e:
+                logger.error(f"Failed to generate experts via LLM: {e}")
+                yield f"event: error\ndata: {json.dumps({'error': f'專家團隊生成失敗: {str(e)}'}, ensure_ascii=False)}\n\n"
+                return
+        else:
+            # 外部來源生成專家
+            try:
+                experts_data, actual_source = await expert_source_service.get_experts(
+                    source=request.expert_source.value,
+                    count=request.expert_count,
+                    language=request.expert_language
+                )
+
+                for i, exp_data in enumerate(experts_data):
+                    experts.append(ExpertProfile(
+                        id=f"expert-{i}",
+                        name=exp_data["name"],
+                        domain=exp_data["domain"],
+                        perspective=f"從{exp_data['domain']}角度思考"
+                    ))
+
+                logger.info(f"Generated {len(experts)} experts from {actual_source}")
+
+            except Exception as e:
+                # 外部來源失敗，fallback 到 LLM
+                logger.warning(f"External source failed: {e}, falling back to LLM")
+                yield f"event: expert_fallback\ndata: {json.dumps({'original': request.expert_source.value, 'fallback': 'llm', 'reason': str(e)}, ensure_ascii=False)}\n\n"
+
+                try:
+                    expert_prompt = get_expert_generation_prompt(
+                        query=request.query,
+                        categories=all_categories,
+                        expert_count=request.expert_count,
+                        custom_experts=request.custom_experts
+                    )
+
+                    expert_response = await ollama_provider.generate(
+                        expert_prompt, model=model, temperature=temperature
+                    )
+
+                    expert_data = extract_json_from_response(expert_response)
+                    experts_raw = expert_data.get("experts", [])
+
+                    for exp in experts_raw:
+                        if isinstance(exp, dict) and all(k in exp for k in ["id", "name", "domain"]):
+                            experts.append(ExpertProfile(**exp))
+
+                    actual_source = "llm"
+
+                except Exception as llm_error:
+                    logger.error(f"LLM fallback also failed: {llm_error}")
+                    yield f"event: error\ndata: {json.dumps({'error': f'專家團隊生成失敗: {str(llm_error)}'}, ensure_ascii=False)}\n\n"
+                    return
+
+        # 回報來源資訊
+        yield f"event: expert_source\ndata: {json.dumps({'source': actual_source}, ensure_ascii=False)}\n\n"
         yield f"event: expert_complete\ndata: {json.dumps({'experts': [e.model_dump() for e in experts]}, ensure_ascii=False)}\n\n"
 
         if not experts: