Update app.py

app.py

@@ -1,119 +1,91 @@
 import os
 import streamlit as st
-import urllib3
 from bs4 import BeautifulSoup
+import urllib3
+import faiss
 import pandas as pd
-import tempfile
+from langchain.vectorstores import FAISS
+from langchain.embeddings import HuggingFaceEmbeddings
+from langchain.chains import RetrievalQA
+from langchain.prompts import PromptTemplate
+from langchain.llms import HuggingFaceHub
 from groq import Groq
 
 # Initialize Groq client
 client = Groq(api_key=os.environ.get('GroqApi'))
 
-# Tariff URL for scraping
-TARIFF_URL = "https://iesco.com.pk/index.php/customer-services/tariff-guide"
-
-# Predefined appliance list with typical loads in watts
-APPLIANCE_OPTIONS = {
-    "Fan": 75,
-    "Air Conditioner (1 Ton)": 1500,
-    "Air Conditioner (1.5 Ton)": 2200,
-    "Refrigerator": 150,
-    "LED Bulb (20W)": 20,
-    "Tube Light": 40,
-    "Iron": 1000,
-    "Microwave Oven": 1200,
-    "Washing Machine": 500,
-    "Electric Heater": 1500,
-    "Laptop": 50,
-    "Desktop Computer": 200,
-    "Television (LCD/LED)": 120,
-    "Water Pump": 1000,
-    "Geyser (Electric)": 3000
-}
+# Initialize FAISS index
+if "vectorstore" not in st.session_state:
+    st.session_state.vectorstore = None
 
-# Initialize session state
-if "appliance_data" not in st.session_state:
-    st.session_state.appliance_data = []
-if "tariff_data" not in st.session_state:
-    st.session_state.tariff_data = {"rate_per_unit": 25}  # Default fallback rate
-
-def scrape_tariff_data(url):
-    """Scrape tariff data from the specified URL."""
+def scrape_web_data(url):
+    """Scrape tariff data from the given URL."""
     try:
         http = urllib3.PoolManager()
-        response = http.request('GET', url)
+        response = http.request("GET", url)
         if response.status == 200:
-            soup = BeautifulSoup(response.data, 'html.parser')
-            text = soup.get_text()
-            # Simulating tariff rate extraction (replace with accurate extraction logic)
-            rate = 25  # Placeholder value, replace with extraction from `text`
-            return {"rate_per_unit": rate}
+            soup = BeautifulSoup(response.data, "html.parser")
+            all_text = soup.get_text()
+            return [{"Data": line.strip()} for line in all_text.split("\n") if line.strip()]
         else:
-            st.warning(f"Failed to fetch tariff data. Status code: {response.status}")
+            st.warning(f"Failed to fetch data: {response.status}")
+            return []
     except Exception as e:
-        st.error(f"Error scraping tariff data: {e}")
-    return {"rate_per_unit": 25}  # Default fallback rate
-
-# Streamlit App Configuration
-st.title("Pakistani Electricity Bill Calculator")
+        st.error(f"An error occurred: {e}")
+        return []
 
-# Step 1: Fetch Tariff Data
-st.header("Step 1: Fetch Tariff Data")
-if st.button("Fetch Tariff Data"):
-    st.session_state.tariff_data = scrape_tariff_data(TARIFF_URL)
-    st.success("Tariff data updated successfully!")
+def store_tariff_data(data):
+    """Store tariff data into a FAISS vector database."""
+    if data:
+        df = pd.DataFrame(data)
+        embeddings = HuggingFaceEmbeddings()
+        vectorstore = FAISS.from_dataframe(df, embeddings)
+        st.session_state.vectorstore = vectorstore
+        st.success("Tariff data successfully stored in the vector database!")
 
-# Display current tariff data
-st.write(f"**Current Tariff Rate:** PKR {st.session_state.tariff_data['rate_per_unit']} per kWh")
+# Streamlit UI
+st.title("Electricity Bill Calculator - Pakistan")
+st.subheader("Real-Time Tariff Data Retrieval")
 
-# Step 2: Add Appliances
-st.sidebar.header("Add Appliances")
-appliance = st.sidebar.selectbox("Select Appliance", options=list(APPLIANCE_OPTIONS.keys()))
-default_load = APPLIANCE_OPTIONS[appliance]
-quantity = st.sidebar.number_input("Number of Appliances", min_value=1, value=1)
-load = st.sidebar.number_input("Load (Watts per Appliance)", min_value=1, value=default_load)
-usage_hours = st.sidebar.number_input("Usage Time (Hours per Day)", min_value=0.0, value=6.0)
+# Step 1: Scraping Tariff Data
+website_url = st.text_input("Enter NEPRA Tariff URL:")
+if st.button("Scrape Tariff Data"):
+    scraped_data = scrape_web_data(website_url)
+    if scraped_data:
+        store_tariff_data(scraped_data)
+        st.write("Scraped Data:")
+        for item in scraped_data[:5]:  # Show first 5 data items for reference
+            st.write(item["Data"])
 
-if st.sidebar.button("Add Appliance"):
-    st.session_state.appliance_data.append({
-        "appliance": appliance,
-        "load": load,
-        "quantity": quantity,
-        "usage_hours": usage_hours
-    })
-    st.sidebar.success(f"Added {quantity} {appliance}(s) to the list!")
+# Step 2: Appliance Data Input
+st.subheader("Step 2: Enter Appliance Details")
+appliances = st.text_area(
+    "Enter appliance details in the format: Appliance, Load(Watts), Usage(Hours per Day)",
+    "Fan, 75, 6\nRefrigerator, 150, 24\nLED Bulb, 20, 5",
+)
+appliance_list = [line.split(",") for line in appliances.split("\n") if line.strip()]
+appliance_data = [
+    {"appliance": item[0].strip(), "load": float(item[1].strip()), "usage": float(item[2].strip())}
+    for item in appliance_list
+]
 
-# Display Appliance List
-st.header("Step 3: Appliance List")
-if st.session_state.appliance_data:
-    total_load = 0
-    total_energy_wh = 0
-    for idx, data in enumerate(st.session_state.appliance_data):
-        st.write(f"{idx + 1}. {data['appliance']} - {data['quantity']} units, {data['load']}W each, {data['usage_hours']} hours/day")
-        total_load += data["quantity"] * data["load"]
-        total_energy_wh += data["quantity"] * data["load"] * data["usage_hours"]
-    st.write(f"**Total Load: {total_load} Watts**")
-else:
-    st.write("No appliances added yet.")
+# Step 3: Calculate Bill
+if st.button("Calculate Bill"):
+    if st.session_state.vectorstore:
+        # Query FAISS for tariff information
+        retriever = st.session_state.vectorstore.as_retriever()
+        chain = RetrievalQA.from_chain_type(llm=HuggingFaceHub(model="google/flan-t5-base"), retriever=retriever)
+        tariff_query = "What is the rate per unit for electricity in PKR?"
+        result = chain.run(tariff_query)
 
-# Step 4: Calculate Monthly Bill
-if st.button("Calculate Monthly Bill"):
-    if not st.session_state.appliance_data:
-        st.warning("Please add appliances before calculating the bill.")
+        # Parse the result and calculate the bill
+        try:
+            rate_per_unit = float(result.split()[0])  # Extract numerical value from result
+            total_units = sum((item["load"] * item["usage"] * 30) / 1000 for item in appliance_data)
+            total_bill = total_units * rate_per_unit
+            st.success(f"Your estimated monthly electricity bill is: PKR {total_bill:.2f}")
+        except Exception as e:
+            st.error(f"Error calculating bill: {e}")
     else:
-        total_energy_kwh = total_energy_wh / 1000  # Convert energy to kWh
-        monthly_energy_kwh = total_energy_kwh * 30
-        rate_per_unit = st.session_state.tariff_data.get("rate_per_unit", 25)
-        monthly_bill = monthly_energy_kwh * rate_per_unit
-
-        # Query AI Model (optional, for additional insights)
-        query = f"Calculate the monthly electricity bill for appliances with a total energy consumption of {monthly_energy_kwh:.2f} kWh per month."
-        response = client.chat.completions.create(
-            messages=[{"role": "user", "content": query}],
-            model="llama3-8b-8192",
-        )
+        st.error("No tariff data available. Please scrape tariff data first.")
 
-        # Display Results
-        st.write(f"**Total Monthly Electricity Bill: PKR {monthly_bill:.2f}**")
-        st.write("**AI Model Insight:**")
-        st.write(response.choices[0].message.content)