update

backend/main.py

@@ -18,8 +18,16 @@ def _env(name: str, default: str) -> str:
     return value if value else default
 
 
-# Example for local Ollama: http://localhost:11434/api/generate
+# Example: http://host:11434/api/generate — для чата берётся тот же хост + /api/chat
 OLLAMA_API_URL = _env("OLLAMA_API_URL", "http://host.docker.internal:11434/api/generate")
+
+
+def _ollama_chat_url() -> str:
+    base = OLLAMA_API_URL.replace("/api/generate", "").rstrip("/")
+    return f"{base}/api/chat"
+
+
+OLLAMA_CHAT_URL = _env("OLLAMA_CHAT_URL", _ollama_chat_url())
 REQUIRE_AUTH_FOR_GENERATE = _env("REQUIRE_AUTH_FOR_GENERATE", "true").lower() in ("1", "true", "yes", "on")
 
 from backend.auth import (  # noqa: E402
@@ -40,6 +48,16 @@ class GenerationRequest(BaseModel):
     model: str = "codellama"
 
 
+class ChatMessageIn(BaseModel):
+    role: str
+    content: str
+
+
+class ChatRequest(BaseModel):
+    model: str
+    messages: list[ChatMessageIn]
+
+
 def validate_config(config_text: str):
     try:
         yaml.safe_load(config_text)
@@ -151,6 +169,38 @@ def me(user: User = Depends(get_current_user)):
     return {"id": str(user.id), "email": user.email, "full_name": user.full_name, "is_admin": user.is_admin}
 
 
+@app.post("/chat")
+def chat_with_ollama(
+    req: ChatRequest,
+    user: User | None = Depends(get_current_user) if REQUIRE_AUTH_FOR_GENERATE else None,
+):
+    if not req.messages:
+        raise HTTPException(status_code=400, detail="messages must not be empty")
+    for m in req.messages:
+        if m.role not in ("user", "assistant", "system"):
+            raise HTTPException(status_code=400, detail=f"invalid role: {m.role}")
+
+    payload = {
+        "model": req.model,
+        "messages": [{"role": m.role, "content": m.content} for m in req.messages],
+        "stream": False,
+    }
+    try:
+        response = requests.post(OLLAMA_CHAT_URL, json=payload, timeout=120)
+        response.raise_for_status()
+        data = response.json()
+        msg = data.get("message") or {}
+        content = msg.get("content", "")
+        if not content and isinstance(data.get("response"), str):
+            content = data["response"]
+        return {"role": "assistant", "content": content}
+    except requests.HTTPError as e:
+        detail = e.response.text if e.response is not None else str(e)
+        raise HTTPException(status_code=502, detail=f"Ollama HTTP error: {detail}")
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Ollama error: {str(e)}")
+
+
 @app.post("/generate")
 async def generate_pipeline(
     req: GenerationRequest,