Update app.py

fix event loop

app.py

@@ -5,9 +5,19 @@ import warnings
 import time
 import sqlite3
 import shutil
+import asyncio  # <--- NEW IMPORT
 
 # ==========================================
-# 1. PAGE CONFIG (MUST BE FIRST)
+# 0. ASYNC FIX (CRITICAL FOR STREAMLIT)
+# ==========================================
+# This fixes "There is no current event loop" errors
+try:
+    asyncio.get_running_loop()
+except RuntimeError:
+    asyncio.set_event_loop(asyncio.new_event_loop())
+
+# ==========================================
+# 1. PAGE CONFIG (MUST BE FIRST STREAMLIT CMD)
 # ==========================================
 st.set_page_config(page_title="Bank Loan Agent (Multi-Model)", layout="wide")
 
@@ -22,10 +32,9 @@ INDEX_PATH = "faiss_index"
 REQUIRED_PDFS = ["Bank Loan Overall Risk Policy.pdf", "Bank Loan Interest Rate Policy.pdf"]
 
 try:
-    # GROQ IMPORTS
+    # PROVIDER IMPORTS
     from langchain_groq import ChatGroq
-    # GOOGLE IMPORTS
-    from langchain_google_genai import ChatGoogleGenerativeAI
+    from langchain_google_genai import ChatGoogleGenerativeAI, GoogleGenerativeAIEmbeddings
     
     # SHARED IMPORTS
     from langchain_huggingface import HuggingFaceEmbeddings
@@ -41,7 +50,7 @@ try:
 
 except ImportError as e:
     st.error(f"❌ Critical Import Error: {e}")
-    st.info("💡 Suggestion: Add 'langchain-google-genai' to requirements.txt")
+    st.info("💡 Suggestion: Check requirements.txt contains 'langchain-google-genai'")
     st.stop()
 
 # ==========================================
@@ -49,8 +58,7 @@ except ImportError as e:
 # ==========================================
 def init_db():
     """Converts CSV files to SQLite DB."""
-    if os.path.exists(DB_FILE):
-        return
+    if os.path.exists(DB_FILE): return
 
     conn = sqlite3.connect(DB_FILE)
     csv_files = {
@@ -64,14 +72,9 @@ def init_db():
             if os.path.exists(file):
                 df = pd.read_csv(file)
                 df.columns = [c.strip() for c in df.columns] 
-                if 'ID' in df.columns:
-                    df['ID'] = df['ID'].astype(str)
-                try:
-                    df.to_sql(table, conn, if_exists='replace', index=False)
-                except Exception:
-                    pass 
-    except Exception as e:
-        st.error(f"DB Init Error: {e}")
+                if 'ID' in df.columns: df['ID'] = df['ID'].astype(str)
+                try: df.to_sql(table, conn, if_exists='replace', index=False)
+                except: pass 
     finally:
         conn.close()
 
@@ -83,30 +86,23 @@ def run_query(query, params=()):
             cursor = conn.cursor()
             cursor.execute(query, params)
             return cursor.fetchone()
-    except Exception as e:
-        return f"DB Error: {e}"
+    except Exception as e: return f"DB Error: {e}"
 
 # ==========================================
 # 4. DEFINE TOOLS
 # ==========================================
-
 @tool
 def get_credit_score(user_id: str) -> str:
     """Queries SQL DB for Credit Score."""
     clean_id = ''.join(filter(str.isdigit, str(user_id)))
     row = run_query("SELECT Credit_Score FROM credit_score WHERE ID = ?", (clean_id,))
-    if row and not isinstance(row, str):
-        return f"Credit Score: {row[0]}"
-    return "User ID not found."
+    return f"Credit Score: {row[0]}" if (row and not isinstance(row, str)) else "User ID not found."
 
 @tool
 def get_account_status(user_id: str) -> str:
     """Queries SQL DB for Name, Nationality, Status, and Email."""
     clean_id = ''.join(filter(str.isdigit, str(user_id)))
-    row = run_query(
-        "SELECT Name, Nationality, Account_Status, Email FROM account_status WHERE ID = ?", 
-        (clean_id,)
-    )
+    row = run_query("SELECT Name, Nationality, Account_Status, Email FROM account_status WHERE ID = ?", (clean_id,))
     if row and not isinstance(row, str):
         return f"Customer Name: {row[0]}, Nationality: {row[1]}, Status: {row[2]}, Email: {row[3]}"
     return "User ID not found."
@@ -118,20 +114,15 @@ def check_pr_status(user_id: str) -> str:
     row = run_query("SELECT PR_Status FROM pr_status WHERE ID = ?", (clean_id,))
     if not row or (isinstance(row, str) and "no such column" in row.lower()):
          row = run_query("SELECT Is_PR FROM pr_status WHERE ID = ?", (clean_id,))
-    if row and not isinstance(row, str):
-        return f"PR Status: {row[0]}"
-    return "PR Status: False."
+    return f"PR Status: {row[0]}" if (row and not isinstance(row, str)) else "PR Status: False."
 
 # ==========================================
 # 5. STREAMLIT APP UI
 # ==========================================
 st.title("🤖 Multi-Model Loan Assessor")
 st.markdown("Agent connects to **SQLite Database** and **Persistent Vector Store**")
-
-# Calculate missing PDFs
 pdfs_missing = [f for f in REQUIRED_PDFS if not os.path.exists(f)]
 
-# --- METRICS FUNCTION ---
 def update_metrics(placeholder):
     manual_time = 15 * 60  
     if 'execution_time' in st.session_state:
@@ -143,29 +134,27 @@ def update_metrics(placeholder):
             col_kpi1.metric("AI Processing", f"{ai_time:.1f}s")
             col_kpi2.metric("Time Saved", f"{time_saved/60:.1f} min", delta=f"{saved_pct:.1f}% faster")
 
-# --- SIDEBAR (MULTI-PROVIDER) ---
+# --- SIDEBAR ---
 with st.sidebar:
     st.header("🔐 Authentication")
     
-    # 1. Provider Selector
+    # 1. Select Provider
     provider_option = st.radio("Select AI Model:", ["Groq (Llama-3)", "Google (Gemini)"])
     
-    # Initialize State
+    # State Management
     if 'auth_status' not in st.session_state:
         st.session_state['auth_status'] = False
         st.session_state['api_key'] = None
         st.session_state['provider'] = None
 
-    # Reset if user switches provider
     if st.session_state.get('provider') != provider_option:
         st.session_state['auth_status'] = False
         st.session_state['api_key'] = None
         st.session_state['provider'] = provider_option
 
-    # 2. Authentication Logic
+    # 2. Auth Logic
     if not st.session_state['auth_status']:
         api_key_input = st.text_input(f"Enter {provider_option} API Key", type="password")
-        
         if st.button("Validate Key"):
             if not api_key_input:
                 st.error("⚠️ Please enter a key.")
@@ -178,9 +167,7 @@ with st.sidebar:
                         else:
                             test_llm = ChatGoogleGenerativeAI(google_api_key=api_key_input, model="gemini-1.5-flash")
                         
-                        test_llm.invoke("Test Connection")
-                        
-                        # Store Success
+                        test_llm.invoke("Test")
                         st.session_state['auth_status'] = True
                         st.session_state['api_key'] = api_key_input
                         st.success("✅ Valid Key!")
@@ -190,32 +177,22 @@ with st.sidebar:
                     st.error(f"❌ Connection Failed: {e}")
     else:
         st.success(f"✅ {st.session_state['provider']} Active")
-        if st.button("🔴 Logout / Change Model"):
+        if st.button("🔴 Logout"):
             st.session_state['auth_status'] = False
             st.session_state['api_key'] = None
             st.rerun()
 
     st.divider()
-    st.subheader("🛠️ Maintenance")
-    
-    if st.button("♻️ Rebuild Knowledge Base"):
+    if st.button("♻️ Rebuild DB/RAG"):
         if os.path.exists(INDEX_PATH): shutil.rmtree(INDEX_PATH)
         st.cache_resource.clear()
-        st.success("Cache cleared.")
-        time.sleep(1)
         st.rerun()
 
-    if st.button("💾 Reload CSVs"):
-        if os.path.exists(DB_FILE): os.remove(DB_FILE)
-        init_db()
-        st.success("Database refreshed.")
-        
-    st.divider()
     if os.path.exists(DB_FILE) and not pdfs_missing:
         st.success("✅ System Ready")
     else:
         st.warning(f"⚠️ Missing: {pdfs_missing}")
-
+    
     st.header("📊 Metrics")
     metrics_placeholder = st.empty()
     update_metrics(metrics_placeholder)
@@ -223,133 +200,106 @@ with st.sidebar:
 # --- MAIN LOGIC ---
 if st.session_state.get('auth_status', False):
     
-    # --- RAG SETUP (Shared Embeddings) ---
+    current_key = st.session_state['api_key']
+    current_provider = st.session_state['provider']
+
+    # --- DYNAMIC RAG SETUP ---
     @st.cache_resource
-    def setup_rag():
-        if pdfs_missing:
-            st.error(f"Missing PDFs: {pdfs_missing}")
-            st.stop()
-            
-        # We use HuggingFace embeddings for BOTH providers to keep the Vector Store compatible
-        # This prevents having to rebuild the index every time you switch models.
-        embeddings = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")
+    def setup_rag(_provider, _key): # Arguments included to force refresh on switch
+        if pdfs_missing: st.stop()
         
-        if os.path.exists(INDEX_PATH):
-            return FAISS.load_local(INDEX_PATH, embeddings, allow_dangerous_deserialization=True).as_retriever()
+        # Use Google Embeddings if Gemini, else HuggingFace
+        if "Google" in _provider:
+            embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001", google_api_key=_key)
         else:
-            documents = []
-            for pdf_file in REQUIRED_PDFS:
-                loader = PyPDFLoader(pdf_file)
-                documents.extend(loader.load())
-            
-            text_splitter = CharacterTextSplitter(chunk_size=600, chunk_overlap=50)
-            final_docs = text_splitter.split_documents(documents)
-            
-            vectorstore = FAISS.from_documents(final_docs, embeddings)
-            vectorstore.save_local(INDEX_PATH)
-            return vectorstore.as_retriever()
+            embeddings = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")
 
-    with st.spinner("Initializing Knowledge Base..."):
-        retriever = setup_rag()
+        if os.path.exists(INDEX_PATH):
+            # Attempt load. If schema mismatch (different embeddings), rebuild.
+            try:
+                return FAISS.load_local(INDEX_PATH, embeddings, allow_dangerous_deserialization=True).as_retriever()
+            except:
+                pass # Fall through to rebuild
+
+        # Build Index
+        documents = []
+        for pdf_file in REQUIRED_PDFS:
+            documents.extend(PyPDFLoader(pdf_file).load())
+        splits = CharacterTextSplitter(chunk_size=600, chunk_overlap=50).split_documents(documents)
+        vectorstore = FAISS.from_documents(splits, embeddings)
+        vectorstore.save_local(INDEX_PATH)
+        return vectorstore.as_retriever()
 
-    # --- DYNAMIC LLM INSTANTIATION ---
-    current_key = st.session_state['api_key']
-    current_provider = st.session_state['provider']
+    with st.spinner("Initializing Knowledge Base..."):
+        # We pass args so Streamlit sees them as dependencies
+        retriever = setup_rag(current_provider, current_key)
 
+    # --- LLM INSTANTIATION ---
     if "Groq" in current_provider:
         llm = ChatGroq(api_key=current_key, temperature=0, model_name="llama-3.3-70b-versatile")
     else:
         llm = ChatGoogleGenerativeAI(google_api_key=current_key, temperature=0, model="gemini-1.5-flash")
 
-    # --- AGENT CHAIN ---
-    rag_prompt = ChatPromptTemplate.from_template("Answer based on context:\n{context}\nQuestion: {question}")
+    # --- AGENT SETUP ---
     rag_chain = (
         {"context": retriever | (lambda d: "\n".join([x.page_content for x in d])), "question": RunnablePassthrough()}
-        | rag_prompt | llm | StrOutputParser()
+        | ChatPromptTemplate.from_template("{context}\nQ:{question}") | llm | StrOutputParser()
     )
     
     @tool
     def consult_policy_doc(query: str) -> str:
-        """Consults Policy Documents for Risk Rules."""
+        """Consults Policy Documents."""
         return rag_chain.invoke(query)
 
     tools = [get_credit_score, get_account_status, check_pr_status, consult_policy_doc]
-
     prompt = ChatPromptTemplate.from_messages([
-        ("system", "You are a Loan Risk Officer. Query SQL DB for customer info. Consult Policy Docs for rules."),
+        ("system", "Act as a Loan Officer. Query SQL DB for info. Check Policies via tool. Output Markdown report."),
         ("human", "{input}"),
         MessagesPlaceholder(variable_name="agent_scratchpad"),
     ])
     
-    agent = create_tool_calling_agent(llm, tools, prompt)
-    agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True, return_intermediate_steps=True)
+    agent_executor = AgentExecutor(agent=create_tool_calling_agent(llm, tools, prompt), tools=tools, verbose=True, return_intermediate_steps=True)
 
-    # --- UI INPUT ---
+    # --- UI ---
     col1, col2 = st.columns([1, 2])
     with col1:
-        st.subheader("1. Customer Details")
+        st.subheader("1. Details")
         uid = st.text_input("Customer ID", "1111")
-        use_simulation = st.checkbox("Simulation Mode")
-        
-        sim_score = 650
-        sim_status = "good-standing"
-        if use_simulation:
-            sim_score = st.slider("Sim Credit Score", 300, 900, 450)
-            sim_status = st.selectbox("Sim Status", ["good-standing", "closed", "delinquent"])
-
-        st.divider()
-        btn = st.button("Assess Loan Risk", type="primary")
-    
+        use_sim = st.checkbox("Simulation Mode")
+        sim_score = st.slider("Score", 300, 900, 450) if use_sim else 0
+        sim_status = st.selectbox("Status", ["good-standing", "closed", "delinquent"]) if use_sim else ""
+        btn = st.button("Assess Risk", type="primary")
+
     with col2:
         if btn:
-            # Build Prompt
-            if use_simulation:
-                 query = f"""
-                Process Loan for Customer ID: {uid}.
-                *** SIMULATION MODE ACTIVE ***
-                1. DO NOT query 'get_credit_score' or 'account_status' for Score/Status.
-                2. USE: Score: {sim_score}, Status: {sim_status}
-                3. Query 'get_account_status' ONLY for Name/Nationality.
-                4. Consult Policy Docs for risk/rates.
-                5. Output Final Report table + Justification.
-                """
-            else:
-                query = f"""
-                Process Loan for Customer ID: {uid}.
-                1. Query SQL tools for Name, Email, Nationality, Status, Score.
-                2. IF Nationality is 'Singaporean', SKIP 'check_pr_status'.
-                3. Consult Policy Docs for risk/rates.
-                4. Output Final Report table + Justification.
-                """
+            query = f"Process Loan {uid}. "
+            if use_sim: query += f"SIMULATION: Use Score {sim_score}, Status {sim_status}. Only query Name/Nationality from DB."
+            else: query += "Query SQL for all data."
+            query += " Check Policies. Output Final Report."
             
-            with st.status(f"🤖 Agent ({current_provider}) is processing...", expanded=True) as status:
+            with st.status(f"🤖 Agent ({current_provider}) Working...", expanded=True) as status:
                 st_callback = StreamlitCallbackHandler(st.container())
                 try:
                     start_time = time.time()
                     res = agent_executor.invoke({"input": query}, {"callbacks": [st_callback]})
-                    end_time = time.time()
-                    st.session_state.execution_time = end_time - start_time
+                    st.session_state.execution_time = time.time() - start_time
                     update_metrics(metrics_placeholder)
-                    status.update(label="✅ Complete!", state="complete", expanded=False)
+                    status.update(label="✅ Done", state="complete", expanded=False)
                 except Exception as e:
                     st.error(f"Error: {e}")
                     st.stop()
 
-            st.success("### 📋 Final Recommendation")
+            st.success("### 📋 Final Report")
             st.markdown(res['output'])
-            
-            with st.expander("🔍 Detailed Trace"):
-                steps = res.get("intermediate_steps", [])
-                for i, (action, observation) in enumerate(steps):
-                    st.markdown(f"**Step {i+1}:** Tool `{action.tool}` | Output: `{observation}`")
+            with st.expander("Trace"):
+                for action, obs in res.get("intermediate_steps", []):
+                    st.markdown(f"**Tool:** `{action.tool}` -> `{obs}`")
 
-            if not use_simulation:
+            if not use_sim:
                 st.divider()
-                with st.expander("✉️ Draft Email"):
-                    email_prompt = f"Write a formal email based on this decision: {res['output']}"
-                    with st.spinner("Drafting..."):
-                        email_draft = llm.invoke(email_prompt).content
-                    st.text_area("Email Draft", value=email_draft, height=200)
+                with st.expander("✉️ Email Draft"):
+                    email = llm.invoke(f"Draft formal email for: {res['output']}").content
+                    st.text_area("Draft", value=email, height=200)
 
 elif not st.session_state.get('auth_status', False):
-    st.info("👈 Please select a provider (Groq or Gemini) and validate your key.")
+    st.info("👈 Select AI Provider & Validate Key in Sidebar")