Update app.py

app.py

@@ -505,55 +505,46 @@
 
 import os
 import gradio as gr
-from langchain_openai import ChatOpenAI
-from langchain.prompts import PromptTemplate
-from langchain.memory import ConversationBufferMemory
+from langchain.chat_models import ChatOpenAI
+from langchain.schema import AIMessage, HumanMessage
 
 # Set OpenAI API Key
-OPENAI_API_KEY = os.getenv('OPENAI_API_KEY')
-
-# Define the template for the chatbot's response
-template = """You are a helpful assistant to answer all user queries.
-{chat_history}
-User: {user_message}
-Chatbot:"""
-
-# Define the prompt template
-prompt = PromptTemplate(
-    input_variables=["chat_history", "user_message"], 
-    template=template
-)
+os.environ["OPENAI_API_KEY"] = "sk-3_mJiR5z9Q3XN-D33cgrAIYGffmMvHfu5Je1U0CW1ZT3BlbkFJA2vfSvDqZAVUyHo2JIcU91XPiAq424OSS8ci29tWMA"  # Replace with your key
 
-# Initialize conversation memory
-memory = ConversationBufferMemory(memory_key="chat_history")
+# Initialize the ChatOpenAI model
+llm = ChatOpenAI(temperature=1.0, model="gpt-3.5-turbo-0613")
 
-# Function to get chatbot response
-def get_text_response(user_message, history=None):
+# Function to predict response
+def get_text_response(message, history=None):
     # Ensure history is a list
     if history is None:
         history = []
-
-    # Prepare the conversation history
-    chat_history = history + [f"User: {user_message}"]
     
-    # Create the full prompt string
-    full_prompt = prompt.format(chat_history="\n".join(chat_history), user_message=user_message)
-
-    llm = ChatOpenAI(temperature=0.5, model="gpt-3.5-turbo")
+    # Convert the Gradio history format to LangChain message format
+    history_langchain_format = []
+    for human, ai in history:
+        history_langchain_format.append(HumanMessage(content=human))
+        history_langchain_format.append(AIMessage(content=ai))
     
-    # Use the invoke method instead of __call__
-    response = llm.invoke(full_prompt)
+    # Add the new user message to the history
+    history_langchain_format.append(HumanMessage(content=message))
+
+    # Get the model's response
+    gpt_response = llm(history_langchain_format)
+
+    # Append AI response to history
+    history.append((message, gpt_response.content))
 
     # Return the response and updated history
-    return response['choices'][0]['message']['content'], chat_history
-
-# Create a Gradio chat interface using ChatInterface
-with gr.Blocks() as demo:
-    chatbot = gr.ChatInterface(
-        get_text_response,  # Function to get responses
-        memory=ConversationBufferMemory()  # Gradio stateful chat memory
-    )
-    
+    return gpt_response.content, history
+
+# Create a Gradio chat interface
+demo = gr.Interface(
+    fn=get_text_response, 
+    inputs=["text", "state"], 
+    outputs=["text", "state"]
+)
+
 if __name__ == "__main__":
     demo.launch()
 
@@ -566,3 +557,4 @@ if __name__ == "__main__":
 
 
 
+