Update app.py

app.py

@@ -7,8 +7,12 @@ import nest_asyncio
 import os
 import subprocess
 import io
+from langchain_google_genai import ChatGoogleGenerativeAI, GoogleGenerativeAIEmbeddings
+from langchain.vectorstores import FAISS
+from langchain.text_splitter import CharacterTextSplitter
+from langchain.chains import ConversationalRetrievalChain
+from langchain.memory import ConversationBufferMemory
 import urllib.parse
-from langchain_google_genai import ChatGoogleGenerativeAI
 
 # Ensure Playwright installs required browsers and dependencies
 subprocess.run(["playwright", "install"])
@@ -58,6 +62,8 @@ def process_multiple_urls(urls):
                 f"""
     **Processing Grant Opportunities** 🚀
     Scanning URL {index+1} of {total_urls}:  `{url}`
+    <br>
+    <p style='font-size: 0.9em; color: #6699CC;'>Completed: {index}/{total_urls} | Remaining: {total_urls - index - 1}</p>
     """,
                 unsafe_allow_html=True,
             )
@@ -88,31 +94,22 @@ def convert_to_excel(data):
 
 
 def create_knowledge_base(data):
-    # Store JSON representation of data in session state
-    st.session_state.knowledge_base_json = json.dumps(data, indent=2)
+    documents = []
+    for grant in data["grants"]:
+        doc_parts = [f"{key.replace('_', ' ').title()}: {value}" for key, value in grant.items()]
+        documents.append("\n".join(doc_parts))
 
+    text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
+    texts = text_splitter.create_documents(documents)
 
-def chat_with_knowledge_base(query):
-    if "knowledge_base_json" not in st.session_state:
-        return "Knowledge base not initialized. Please load grant data first."
-
-    context = st.session_state.knowledge_base_json
-    prompt = f"""
-    You are an AI assistant that helps users analyze grant opportunities.
-    Here is the extracted grant data in JSON format:
-    
-    {context}
-
-    User's question: {query}
-    Answer the question based on the provided grant data.
-    """
+    embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001", google_api_key=GOOGLE_API_KEY)
+    vectorstore = FAISS.from_documents(texts, embeddings)
 
     llm = ChatGoogleGenerativeAI(
         model="gemini-2.0-flash-thinking-exp", google_api_key=GOOGLE_API_KEY, temperature=0
     )
-    
-    response = llm.invoke(prompt)
-    return response
+    memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
+    return ConversationalRetrievalChain.from_llm(llm, vectorstore.as_retriever(), memory=memory)
 
 
 def get_shareable_link(file_data, file_name, file_type):
@@ -124,9 +121,20 @@ def main():
     st.set_page_config(page_title="Quantilytix Grant Finder", page_icon="💰", layout="wide")
     st.title("💰 Quantilytix Grant Finder")
 
+    # --- Introduction and Motivation ---
+    st.markdown("""
+    <div style="text-align: justify;">
+        <p>
+            Welcome to <b>Quantilytix Grant Finder</b>, an AI-powered platform designed to streamline the grant discovery process, especially for academics and researchers in Zimbabwe.
+        </p>
+            
+    </div>
+    """, unsafe_allow_html=True)
+
     st.sidebar.image("logoqb.jpeg", use_container_width=True)
     st.sidebar.header("Scrape & Configure")
 
+    # Initialize session state
     if "scraped_data" not in st.session_state:
         st.session_state.scraped_data = None
     if "chat_history" not in st.session_state:
@@ -134,6 +142,7 @@ def main():
     if "chat_interface_active" not in st.session_state:
         st.session_state.chat_interface_active = False
 
+    # URL Input in Sidebar
     url_input = st.sidebar.text_area(
         "Enter Grant URLs (one per line)",
         height=150,
@@ -141,50 +150,104 @@ def main():
         placeholder="e.g.,\nhttps://www.example-grants.org/opportunities\nhttps://another-funding-source.com/grants-list"
     )
 
+    # Get Grants Button with Icon
     if st.sidebar.button("🔍 Get Grant Opportunities"):
         if url_input:
-            urls = [url.strip() for url in url_input.split("\n") if url.strip()]  # FIXED TYPO
+            urls = [url.strip() for url in url_input.split("\n") if url.strip()]
             if urls:
                 try:
                     with st.spinner("Scraping in progress... Please wait patiently."):
                         result = process_multiple_urls(urls)
                         st.session_state.scraped_data = result
-                        st.success(f"✅ Successfully scraped {len(result['grants'])} grant opportunities!")
+                        st.success(f"✅ Successfully scraped {len(result['grants'])} grant opportunities from {len(urls)} URLs!")
                 except Exception as e:
-                    st.error(f"🚨 Scraping error: {e}")
+                    st.error(f"🚨 Scraping process encountered an error: {e}")
             else:
                 st.warning("⚠️ Please enter valid URLs.")
         else:
-            st.warning("⚠️ Please enter at least one URL.")
+            st.warning("⚠️ Please enter at least one URL to begin scraping.")
 
+    # --- Main Panel for Data Display and Chat ---
     st.markdown("---")
 
-    if st.session_state.scraped_data and st.session_state.scraped_data["grants"]:
+    if st.session_state.scraped_data and st.session_state.scraped_data['grants']:
         st.header("📊 Scraped Grant Data")
 
-        with st.expander(f"📊 Preview Grant Data ({len(st.session_state.scraped_data['grants'])} grants)"):
+        # Data Preview and Download Options in Main Panel
+        with st.expander(f"📊 Preview Grant Data {len(st.session_state.scraped_data['grants'])} grants"):
             st.dataframe(st.session_state.scraped_data["grants"])
 
+        col1, col2, col3 = st.columns([1, 1, 2]) # Adjust column widths for better layout
+
+        with col1:
+            selected_format = st.selectbox("Download As:", ("CSV", "Excel"), key="download_format_selector")
+
+        with col2:
+            if selected_format == "CSV":
+                file_data = convert_to_csv(st.session_state.scraped_data)
+                file_name = "grants_data.csv"
+                file_type = "text/csv"
+            else:
+                file_data = convert_to_excel(st.session_state.scraped_data)
+                file_name = "grants_data.xlsx"
+                file_type = "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
+
+            download_link_html = f"<a href='data:{file_type};base64,{base64.b64encode(file_data).decode()}' download='{file_name}'><button style='background-color:#4CAF50;color:white;padding:10px 15px;border:none;border-radius:4px;'>⬇️ Download {selected_format}</button></a>"
+            st.markdown(download_link_html, unsafe_allow_html=True)
+
+        with col3:
+            shareable_link = get_shareable_link(file_data, file_name, file_type)
+            whatsapp_url = f"https://api.whatsapp.com/send?text={urllib.parse.quote(f'Check out these grant opportunities: {shareable_link}')}"
+            email_subject = urllib.parse.quote("Grant Opportunities File")
+            email_body = urllib.parse.quote(f"Download the grant opportunities file here: {shareable_link}")
+            email_url = f"mailto:?subject={email_subject}&body={email_body}"
+
+            st.markdown("<div style='margin-top:10px;'>Share via:</div>", unsafe_allow_html=True) # Add some margin for better spacing
+            st.markdown(f"📱 [WhatsApp]({whatsapp_url}) | 📧 [Email]({email_url})", unsafe_allow_html=True)
+
+
+        # Knowledge Base and Chat Interface
         if st.button("🧠 Load as Knowledge Base & Chat"):
             with st.spinner("Loading data into knowledge base..."):
-                create_knowledge_base(st.session_state.scraped_data)
+                st.session_state.qa_chain = create_knowledge_base(st.session_state.scraped_data)
                 st.session_state.chat_interface_active = True
-                st.session_state.chat_history = []
-            st.success("Knowledge base loaded! Start chatting.")
+                st.session_state.chat_history = [] # Clear chat history on reload
+            st.success("Knowledge base loaded! You can now chat with the Grants Bot.")
 
         if st.session_state.get("chat_interface_active"):
+            st.markdown("---")
             st.header("💬 Chat with Grants Bot")
-            query = st.text_input("Your question:")
+            st.markdown("Ask questions about the scraped grants to get quick insights!")
+
+            query = st.text_input("Your question:", key="chat_input")
             if query:
-                with st.spinner("Generating response..."):
-                    response = chat_with_knowledge_base(query)
-                    st.session_state.chat_history.append({"query": query, "response": response})
+                if st.session_state.qa_chain:
+                    with st.spinner("Generating response..."):
+                        response = st.session_state.qa_chain({"question": query})
+                        st.session_state.chat_history.append({"query": query, "response": response["answer"]})
+                else:
+                    st.error("Knowledge base not initialized. Please load data as knowledge base.")
 
             if st.session_state.chat_history:
                 st.subheader("Chat History")
                 for chat in st.session_state.chat_history:
-                    st.markdown(f"**You:** {chat['query']}")
-                    st.markdown(f"**Grants Bot:** {chat['response']}")
+                    st.markdown(f"<div style='padding: 10px; border-radius: 5px; margin-bottom: 5px; background-color: #f0f2f6;'><strong>You:</strong> {chat['query']}</div>", unsafe_allow_html=True)
+                    st.markdown(f"<div style='padding: 10px; border-radius: 5px; margin-bottom: 10px; background-color: #e0e2e6;'><strong>Grants Bot:</strong> {chat['response']}</div>", unsafe_allow_html=True)
+
+
+    else:
+        st.info("⬅️  Enter URLs in the sidebar and click 'Get Grant Opportunities' to start scraping.")
+
+    st.sidebar.markdown("---")
+    st.sidebar.markdown(
+        """
+        <div style='text-align: center; font-size: 0.8em; color: grey;'>
+            Powered by <a href="https://quantilytix.com" style='color: grey;'>Quantilytix</a> | &copy; 2025
+        </div>
+        """,
+        unsafe_allow_html=True,
+    )
+
 
 if __name__ == "__main__":
     main()