Update app.py

app.py

@@ -1,47 +1,27 @@
 import streamlit as st
-# Import from the correct package
 from langchain_google_genai import ChatGoogleGenerativeAI  
-from langchain.tools import Tool
-from langchain.agents import initialize_agent
-from langchain.agents import AgentType
-from langchain.tools import DuckDuckGoSearchRun
 from google.ai.generativelanguage_v1beta.types import Tool as GenAITool
-import os
+import json
 
-
-# Configure LangChain LLM with Gemini
-llm = ChatGoogleGenerativeAI(model="gemini-1.5-flash", google_api_key=os.getenv("gemini_api"))
-
-# Use DuckDuckGo Search (no API key needed)
-ddgs = DuckDuckGoSearchRun()
-search_tool = Tool(
-    name="Web Search",
-    func=ddgs.run,
-    description="Searches the web for relevant certification information."
-)
-
-# Create LangChain agent
-agent = initialize_agent(
-    tools=#[search_tool],
-    [GenAITool(google_search={})],
-    llm=llm,
-    agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
-    verbose=True
-)
+# Configure LangChain LLM with Gemini 2.0
+llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash", google_api_key=os.getenv("gemini_api"))
 
 def azure_cert_bot(cert_name):
     query = f"Microsoft Azure {cert_name} certification curriculum site:microsoft.com"
-    search_results = ddgs.run(query).split("\n")
-    prompt = f"Based on the following curriculum details, generate key questions and answers in markdown format for the {cert_name} certification exam. Do not include any metadata or unnecessary text, only return the formatted Q&A:\n{search_results}"
-    response = llm.invoke(prompt)
+    
+    # Use Gemini's built-in Google Search tool
+    resp = llm.invoke(
+        f"Find the Microsoft documentation for {cert_name} and generate key Q&A based on the curriculum.",
+        tools=[GenAITool(google_search={})],
+    )
     
     try:
-        response_text = response.get("content", "No response generated.") if isinstance(response, dict) else response
+        response_text = resp.get("content", "No response generated.") if isinstance(resp, dict) else resp
         response_text = "\n".join([line for line in response_text.split("\n") if not line.lower().startswith("content=") and "metadata" not in line.lower()])
     except Exception as e:
         response_text = f"Error processing response: {str(e)}"
     
-    return search_results, response_text
+    return query, response_text
 
 # Streamlit UI Enhancements
 st.set_page_config(page_title="Azure Certification Prep Assistant", layout="wide")
@@ -80,12 +60,10 @@ st.markdown("<div class='title'>Azure Certification Prep Assistant</div>", unsaf
 cert_name = st.text_input("Enter Azure Certification Name (e.g., AZ-900)", "")
 if st.button("Get Certification Details"):
     if cert_name:
-        links, qa_content = azure_cert_bot(cert_name)
+        search_query, qa_content = azure_cert_bot(cert_name)
         
-        st.markdown("<div class='subheader'>Certification Links & Curriculum</div>", unsafe_allow_html=True)
-        for link in links:
-            if link.strip():
-                st.markdown(f"<div class='markdown-text-container'>- <a href='{link}' target='_blank'>{link}</a></div>", unsafe_allow_html=True)
+        st.markdown("<div class='subheader'>Certification Search Query</div>", unsafe_allow_html=True)
+        st.markdown(f"<div class='markdown-text-container'>🔍 {search_query}</div>", unsafe_allow_html=True)
         
         st.markdown("<div class='subheader'>Exam Questions & Answers</div>", unsafe_allow_html=True)
         st.markdown(f"<div class='markdown-text-container'>{qa_content}</div>", unsafe_allow_html=True)