Update app.py

app.py

@@ -25,28 +25,38 @@ async def search_api(request: ChatRequest, req: Request):
         tavily_api_key = os.environ.get("TAVILY_API_KEY", "")
         nvidia_api_key = os.environ.get("NVIDIA_API_KEY", "")
 
-        tavily_client = AsyncTavilyClient(api_key=tavily_api_key)
-        search_res = await tavily_client.search(query=request.query, search_depth="basic", max_results=5)
+        search_query = request.query.strip().replace("\n", " ")
+        if len(search_query) > 300:
+            search_query = search_query[:300]
 
-        results = search_res.get("results", [])
         sources = []
         context_parts = []
 
-        for idx, r in enumerate(results, 1):
-            url = r.get("url", "")
-            favicon = get_favicon_url(url)
-            sources.append({
-                "title": r.get("title", ""),
-                "url": url,
-                "favicon": favicon
-            })
-            context_parts.append(f"[{idx}] Title: {r.get('title')}\nURL: {url}\nContent: {r.get('content')}\n")
+        if tavily_api_key and search_query:
+            try:
+                tavily_client = AsyncTavilyClient(api_key=tavily_api_key)
+                search_res = await tavily_client.search(query=search_query, search_depth="basic", max_results=5)
+                results = search_res.get("results", [])
+                for idx, r in enumerate(results, 1):
+                    url = r.get("url", "")
+                    favicon = get_favicon_url(url)
+                    sources.append({
+                        "title": r.get("title", ""),
+                        "url": url,
+                        "favicon": favicon
+                    })
+                    context_parts.append(f"[{idx}] Title: {r.get('title')}\nURL: {url}\nContent: {r.get('content')}\n")
+            except Exception:
+                pass
 
         context_str = "\n".join(context_parts)
 
         yield f"data: {json.dumps({'type': 'sources', 'content': sources})}\n\n"
 
-        system_prompt = f"Answer the user's question accurately based on the search context below. Use inline citations like [1], [2] corresponding to sources.\nContext:\n{context_str}"
+        if context_str:
+            system_prompt = f"Answer the user's question accurately based on the search context below. Use inline citations like [1], [2] corresponding to sources.\nContext:\n{context_str}"
+        else:
+            system_prompt = "You are a helpful AI assistant. Answer the user's request accurately."
 
         payload = {
             "model": "stepfun-ai/step-3.7-flash",
@@ -66,28 +76,31 @@ async def search_api(request: ChatRequest, req: Request):
             "Accept": "text/event-stream"
         }
 
-        async with httpx.AsyncClient(timeout=120.0) as client:
-            async with client.stream("POST", "https://integrate.api.nvidia.com/v1/chat/completions", headers=headers, json=payload) as response:
-                async for line in response.aiter_lines():
-                    if await req.is_disconnected():
-                        break
-                    if line.startswith("data: ") and line != "data: [DONE]":
-                        data_str = line[6:].strip()
-                        if not data_str:
-                            continue
-                        try:
-                            data = json.loads(data_str)
-                            choices = data.get("choices", [])
-                            if choices:
-                                delta = choices[0].get("delta", {})
-                                reasoning = delta.get("reasoning_content") or delta.get("reasoning")
-                                if reasoning:
-                                    yield f"data: {json.dumps({'type': 'reasoning', 'content': reasoning})}\n\n"
-                                content = delta.get("content")
-                                if content:
-                                    yield f"data: {json.dumps({'type': 'answer', 'content': content})}\n\n"
-                        except Exception:
-                            continue
+        try:
+            async with httpx.AsyncClient(timeout=120.0) as client:
+                async with client.stream("POST", "https://integrate.api.nvidia.com/v1/chat/completions", headers=headers, json=payload) as response:
+                    async for line in response.aiter_lines():
+                        if await req.is_disconnected():
+                            break
+                        if line.startswith("data: ") and line != "data: [DONE]":
+                            data_str = line[6:].strip()
+                            if not data_str:
+                                continue
+                            try:
+                                data = json.loads(data_str)
+                                choices = data.get("choices", [])
+                                if choices:
+                                    delta = choices[0].get("delta", {})
+                                    reasoning = delta.get("reasoning_content") or delta.get("reasoning")
+                                    if reasoning:
+                                        yield f"data: {json.dumps({'type': 'reasoning', 'content': reasoning})}\n\n"
+                                    content = delta.get("content")
+                                    if content:
+                                        yield f"data: {json.dumps({'type': 'answer', 'content': content})}\n\n"
+                            except Exception:
+                                continue
+        except Exception:
+            pass
 
         yield "data: [DONE]\n\n"