Update app.py

app.py

@@ -1,56 +1,46 @@
 import os
 import streamlit as st
-#from dotenv import load_dotenv
+import logging
 from langchain_groq import ChatGroq
 from langchain_community.tools import ArxivQueryRun, WikipediaQueryRun, DuckDuckGoSearchRun
 from langchain_community.utilities import ArxivAPIWrapper, WikipediaAPIWrapper
 from langchain import hub
-from langchain.agents import create_openai_tools_agent
-from langchain.agents import AgentExecutor
+from langchain.agents import create_openai_tools_agent, AgentExecutor
 from langchain.prompts import PromptTemplate
 from langchain_community.vectorstores import FAISS
 from langchain_huggingface import HuggingFaceEmbeddings
 from langchain.tools import Tool
 from pydantic import BaseModel, Field
 from typing import List, Dict
-groq_api_key = os.getenv("GROQ_API_KEY")
-
-import logging
 
-# Configure the logging
+# Configure logging
 logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
-
-# Example log messages
 logging.info("Streamlit app started")
-logging.error("This is an error log")
 
-if not os.getenv("GROQ_API_KEY"):
-    logging.info("GROQ API key is missing!")
-#git remote set-url origin https://<user_name>:<token>@huggingface.co/<repo_path>
+# Load the API key
+groq_api_key = "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" #os.getenv("GROQ_API_KEY")
+if not groq_api_key:
+    logging.error("GROQ API key is missing!")
+
 # Define a query condensing template
 condense_prompt_template = PromptTemplate(
     input_variables=["original_question", "conversation_history"],
     template="""
         Given the conversation history below, condense the user's query into a clear and specific question.
-
         Conversation History:
         {conversation_history}
-
         Original Question:
         {original_question}
-
         Condensed Question:"""
 )
 
 def condense_query(llm_model, original_question, conversation_history):
-    # Format the prompt with conversation history and original question
     prompt = condense_prompt_template.format(
         original_question=original_question,
         conversation_history=conversation_history
     )
-    # Generate the condensed query
     resp = llm_model.predict(prompt)
-    return resp.strip()  # Clean up whitespace
+    return resp.strip()
 
 # Initialize HuggingFace embeddings and FAISS vectorstore
 embeddings = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")
@@ -69,14 +59,14 @@ def retrieve_documents(query: str) -> List[Dict[str, str]]:
     results = retriever_tool.get_relevant_documents(query)
     return [{"content": doc.page_content} for doc in results]
 
-# Define the custom FAISS tool using Tool
+# Define the FAISS tool
 faiss_tool = Tool(
     name="retrieve_documents",
     description="Retrieve documents from FAISS vectorstore.",
     func=retrieve_documents,
 )
 
-# Arxiv, Wikipedia, and Search tool setup
+# Initialize tools
 arxiv_wrapper = ArxivAPIWrapper(top_k_results=1, doc_content_chars_max=250)
 arxiv_tool = ArxivQueryRun(api_wrapper=arxiv_wrapper)
 
@@ -84,16 +74,9 @@ wiki_wrapper = WikipediaAPIWrapper(top_k_results=1, doc_content_chars_max=250)
 wiki_tool = WikipediaQueryRun(api_wrapper=wiki_wrapper)
 
 search_tool = DuckDuckGoSearchRun(name="Search")
+tools = [arxiv_tool, wiki_tool, search_tool, faiss_tool]
 
-# Define tools, including the wrapped FAISS tool
-tools = [
-    arxiv_tool,
-    wiki_tool,
-    search_tool,
-    faiss_tool
-]
-
-# Set up the agent with prompt and LLM model
+# Set up the agent
 prompt = hub.pull("hwchase17/openai-functions-agent")
 llm = ChatGroq(model="Gemma2-9B-It", api_key=groq_api_key, streaming=True)
 agent = create_openai_tools_agent(llm, tools, prompt)
@@ -112,7 +95,7 @@ conversation_history = "\n".join(
     [f"{msg['role']}: {msg['content']}" for msg in st.session_state['messages']]
 )
 
-# CSS to fix the input box at the bottom of the screen
+# CSS for input box at bottom of screen
 st.markdown("""
     <style>
     .fixed-bottom-input-container {
@@ -131,21 +114,20 @@ st.markdown("""
     </style>
 """, unsafe_allow_html=True)
 
-# Create the input field
+# Chat input and processing
 user_input = st.text_input("Type your message here...", key="user_input", label_visibility="collapsed")
 
-# If there's input, process it
 if user_input:
     condensed_question = condense_query(llm, user_input, conversation_history)
     response = agent_executor.invoke({"input": condensed_question})
-
-    # Append user input and assistant response to messages
+    
+    # Update chat history
     st.session_state.messages.append({"role": "user", "content": user_input})
     st.session_state.messages.append({"role": "assistant", "content": response.get("output", "")})
 
-    # Display the user input and response
+    # Display the response
     st.chat_message("user").write(user_input)
     st.chat_message("assistant").write(response.get("output", ""))
 
-    # Clear the input box by creating a new one
-    st.text_input("Type your message here...", key="user_input_", value="", label_visibility="collapsed")  # Recreate input to clear it
+    # Clear the input field
+    st.session_state["user_input"] = ""