Update app.py

app.py

@@ -34,6 +34,9 @@ def get_data(url):
             "List me all grants or funds, short summary of grant description, "
             "the organisations funding them, the value of the grant as an integer, "
             "the due date, eligible countries, sector and eligibility criteria for applicants."
+
+            "if you cant find grants return any useful information you can find"
+            
         ),
         source=url,
         config=graph_config,
@@ -72,6 +75,8 @@ def process_multiple_urls(urls):
             if result and "grants" in result:
                 all_data["grants"].extend(result["grants"])
         except Exception as e:
+            st.write(result)
+            all_data = result
             st.error(f"⚠️ Error processing URL: {url} - {str(e)}")
             continue
 
@@ -81,35 +86,57 @@ def process_multiple_urls(urls):
 
 
 def convert_to_csv(data):
-    df = pd.DataFrame(data["grants"])
-    return df.to_csv(index=False).encode("utf-8")
+    try:
+        df = pd.DataFrame(data["grants"])
+        return df.to_csv(index=False).encode("utf-8")
+    except:
+        df = pd.DataFrame(data)
+        return df.to_csv(index=False).encode("utf-8")
 
 
 def convert_to_excel(data):
-    df = pd.DataFrame(data["grants"])
-    buffer = io.BytesIO()
-    with pd.ExcelWriter(buffer, engine="xlsxwriter") as writer:
-        df.to_excel(writer, sheet_name="Grants", index=False)
-    return buffer.getvalue()
+    try:
+        df = pd.DataFrame(data["grants"])
+        buffer = io.BytesIO()
+        with pd.ExcelWriter(buffer, engine="xlsxwriter") as writer:
+            df.to_excel(writer, sheet_name="Grants", index=False)
+        return buffer.getvalue()
+
+    except:
+        df = pd.DataFrame(data)
+        buffer = io.BytesIO()
+        with pd.ExcelWriter(buffer, engine="xlsxwriter") as writer:
+            df.to_excel(writer, sheet_name="Grants", index=False)
+        return buffer.getvalue()
+            
 
 
 def create_knowledge_base(data):
-    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))
+    # Store JSON representation of data in session state
+    st.session_state.knowledge_base_json = json.dumps(data, indent=2)
 
-    text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
-    texts = text_splitter.create_documents(documents)
 
-    embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001", google_api_key=GOOGLE_API_KEY)
-    vectorstore = FAISS.from_documents(texts, embeddings)
+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.
+    """
 
     llm = ChatGoogleGenerativeAI(
         model="gemini-2.0-flash-thinking-exp", google_api_key=GOOGLE_API_KEY, temperature=0
     )
-    memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
-    return ConversationalRetrievalChain.from_llm(llm, vectorstore.as_retriever(), memory=memory)
+    
+    response = llm.invoke(prompt)
+    return response
 
 
 def get_shareable_link(file_data, file_name, file_type):
@@ -175,7 +202,10 @@ def main():
 
         # 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"])
+            try:
+                st.dataframe(st.session_state.scraped_data["grants"])
+            except:
+                st.dataframe(st.session_state.scraped_data)
 
         col1, col2, col3 = st.columns([1, 1, 2]) # Adjust column widths for better layout
 
@@ -223,7 +253,7 @@ def main():
             if query:
                 if st.session_state.qa_chain:
                     with st.spinner("Generating response..."):
-                        response = st.session_state.qa_chain({"question": query})
+                        response = chat_with_knowledge_base(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.")