Update app.py

app.py

@@ -1,7 +1,7 @@
 import gradio as gr
 from langchain.prompts import ChatPromptTemplate
-from langchain_huggingface import HuggingFaceEndpoint
 from langchain.schema import HumanMessage, SystemMessage, AIMessage
+from huggingface_hub import InferenceClient
 import os
 import time
 import logging
@@ -11,18 +11,15 @@ import re
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
 
-# Make sure to set your Hugging Face API token in your environment variables
-# For example: export HUGGINGFACEHUB_API_TOKEN='your_token_here'
-if "HUGGINGFACEHUB_API_TOKEN" not in os.environ:
-    logger.warning("HUGGINGFACEHUB_API_TOKEN not set, the application may not work.")
+# Support both token names for flexibility
+hf_token = os.environ.get("HF_TOKEN") or os.environ.get("HUGGINGFACEHUB_API_TOKEN")
+if not hf_token:
+    logger.warning("Neither HF_TOKEN nor HUGGINGFACEHUB_API_TOKEN is set, the application may not work.")
 
-# --- LLM and Template Configuration ---
-llm = HuggingFaceEndpoint(
-    repo_id="google/gemma-2-2b-it",
-    temperature=0.7,
-    top_p=0.9,
-    max_new_tokens=1024,
-    huggingfacehub_api_token=os.getenv("HUGGINGFACEHUB_API_TOKEN")
+# --- LLM Configuration ---
+client = InferenceClient(
+    provider="together",
+    api_key=hf_token,
 )
 math_template = ChatPromptTemplate.from_messages([
     ("system", """{system_message}
@@ -131,12 +128,20 @@ def respond_with_enhanced_streaming(message, history):
         
         logger.info(f"Processing {mode} query: {message[:50]}...")
         
-        # Use the chain with proper message history (YOUR ORIGINAL APPROACH)
-        chain = template | llm
-        response = chain.invoke({
-            "question": message,
-            "system_message": "You are EduBot, an expert AI learning assistant. Provide comprehensive, educational responses that help students truly understand concepts."
-        })
+        # Use LangChain template to format the prompt
+        formatted_prompt = template.format(
+            question=message,
+            system_message="You are EduBot, an expert AI learning assistant. Provide comprehensive, educational responses that help students truly understand concepts."
+        )
+        
+        # Use the Together provider with text_generation
+        response = client.text_generation(
+            formatted_prompt,
+            model="moonshotai/Kimi-K2-Instruct",
+            max_new_tokens=1024,
+            temperature=0.7,
+            top_p=0.9,
+        )
         
         response = smart_truncate(response, max_length=3000)
         
@@ -157,7 +162,7 @@ def respond_with_enhanced_streaming(message, history):
         yield final_response
         
     except Exception as e:
-        logger.exception("Error in LangChain response generation")
+        logger.exception("Error in response generation")
         yield f"Sorry, I encountered an error: {str(e)}"
 
 # --- Fixed Gradio UI and CSS ---