feat(B-LLM): Zentraler LLM Client — llm_complete() + llm_embed() + Migration + Tests + Doku

B-LLM: llm_client.py um generische llm_complete() und llm_embed() erweitert
- Provider-Auswahl, API-Key-Auflösung, Error-Handling, Cost-Tracking
- Retry mit Exponential-Backoff für transient errors
- Timeout konfigurierbar
- Helper: get_api_credentials(), build_model(), _classify_error()

B-LLM-MIG: Alle 8 direkten litellm.acompletion() Calls auf llm_complete() umgestellt
- agent_runner.py, query_understanding.py (2x), ai_proactive (3x), ai_assistant (2x)
- 0 verbleibende direkte litellm.acompletion() Calls außerhalb llm_client.py

B-LLM-TEST: 39 Tests in test_llm_client.py — alle grün
- Mock mode, error handling, embed, helpers, backward compat

B-LLM-DOC: Plugin-Dev-Guide Kapitel 7 (LLM Integration) hinzugefügt

app/plugins/builtins/unified_search/query_understanding.py

@@ -8,7 +8,7 @@ import logging
 import uuid
 from typing import Any
 
-import litellm
+from app.ai.llm_client import llm_complete
 from sqlalchemy.ext.asyncio import AsyncSession
 
 logger = logging.getLogger(__name__)
@@ -87,7 +87,7 @@ async def llm_analyze_query(
         api_key, api_base, provider_type = await _get_api_credentials(db, tenant_id)
         model = _build_model(DEFAULT_LLM_MODEL, provider_type)
 
-        litellm_kwargs: dict[str, Any] = dict(
+        result = await llm_complete(
             model=model,
             messages=[
                 {"role": "system", "content": QUERY_ANALYZE_SYSTEM},
@@ -96,14 +96,10 @@ async def llm_analyze_query(
             temperature=0.1,
             max_tokens=500,
             response_format={"type": "json_object"},
+            api_key=api_key,
+            api_base=api_base,
         )
-        if api_key:
-            litellm_kwargs["api_key"] = api_key
-        if api_base:
-            litellm_kwargs["api_base"] = api_base
-
-        response = await litellm.acompletion(**litellm_kwargs)
-        content = response.choices[0].message.content
+        content = result["content"]
         # Strip markdown code fences if present
         content = content.strip()
         if content.startswith("```"):
@@ -139,7 +135,7 @@ async def llm_aggregate_results(
         ]
         user_msg = json.dumps({"query": query, "results": compact})
 
-        litellm_kwargs: dict[str, Any] = dict(
+        result = await llm_complete(
             model=model,
             messages=[
                 {"role": "system", "content": RESULT_AGGREGATE_SYSTEM},
@@ -148,14 +144,10 @@ async def llm_aggregate_results(
             temperature=0.1,
             max_tokens=1000,
             response_format={"type": "json_object"},
+            api_key=api_key,
+            api_base=api_base,
         )
-        if api_key:
-            litellm_kwargs["api_key"] = api_key
-        if api_base:
-            litellm_kwargs["api_base"] = api_base
-
-        response = await litellm.acompletion(**litellm_kwargs)
-        content = response.choices[0].message.content
+        content = result["content"]
         # Strip markdown code fences if present
         content = content.strip()
         if content.startswith("```"):