Update app.py

app.py

@@ -1,62 +1,18 @@
 import os
 from flask import Flask, render_template, request, jsonify
 from flask_cors import CORS
-from langchain_huggingface import HuggingFaceEndpoint
-# THE STABLE IMPORTS - NO MORE MISSING NAMES
-from langchain.agents import AgentExecutor, create_react_agent
-from langchain_core.tools import Tool
-from langchain_core.prompts import PromptTemplate
-from duckduckgo_search import DDGS
+from huggingface_hub import InferenceClient
 
 app = Flask(__name__)
 CORS(app)
 
-# 1. Setup LLM
-sec_token = os.getenv("HF_TOKEN")
-llm = HuggingFaceEndpoint(
-    repo_id="meta-llama/Llama-3.2-3B-Instruct",
-    huggingfacehub_api_token=sec_token,
+# 1. Initialize the Official Client
+# This uses your HF_TOKEN secret automatically
+client = InferenceClient(
+    model="meta-llama/Llama-3.2-3B-Instruct",
+    token=os.getenv("HF_TOKEN")
 )
 
-# 2. Manual Search
-def manual_search(query: str):
-    try:
-        with DDGS() as ddgs:
-            results = [r['body'] for r in ddgs.text(query, max_results=3)]
-            return "\n".join(results) if results else "No results found."
-    except Exception as e:
-        return f"Search error: {str(e)}"
-
-tools = [
-    Tool(
-        name="Search",
-        func=manual_search,
-        description="Useful for finding current info on the web."
-    )
-]
-
-# 3. Explicit Prompt Template (This is required for create_react_agent)
-template = """Answer the following questions:
-{tools}
-Format:
-Question: {input}
-Thought: {agent_scratchpad}
-Action: the action to take, should be one of [{tool_names}]
-Action Input: the input to the action
-Observation: the result of the action
-... (repeat)
-Final Answer: the final answer
-
-Begin!
-Question: {input}
-Thought: {agent_scratchpad}"""
-
-prompt = PromptTemplate.from_template(template)
-
-# 4. Initialize Agent
-agent = create_react_agent(llm, tools, prompt)
-agent_executor = AgentExecutor(agent=agent, tools=tools, handle_parsing_errors=True)
-
 @app.route('/')
 def index():
     return render_template('index.html')
@@ -65,12 +21,25 @@ def index():
 def ask():
     try:
         data = request.get_json()
-        query = data.get("query")
-        # Run agent
-        result = agent_executor.invoke({"input": query})
-        return jsonify({"answer": result["output"]})
+        user_query = data.get("query")
+
+        # 2. Simple, Direct Call to Llama
+        response = ""
+        for message in client.chat_completion(
+            messages=[{"role": "user", "content": user_query}],
+            max_tokens=500,
+            stream=True,
+        ):
+            token = message.choices[0].delta.content
+            if token:
+                response += token
+
+        return jsonify({"answer": response})
+
     except Exception as e:
-        return jsonify({"answer": f"System Error: {str(e)}"}), 500
+        print(f"Error: {str(e)}")
+        # This will tell us EXACTLY if the token is the problem
+        return jsonify({"answer": f"System Status: {str(e)}"}), 500
 
 if __name__ == "__main__":
     app.run(host="0.0.0.0", port=7860)