Update app.py

app.py

@@ -2,90 +2,127 @@ import os
 import streamlit as st
 from bs4 import BeautifulSoup
 import urllib3
-import faiss
-import pandas as pd
-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 langchain_community.vectorstores import FAISS
+from langchain_community.embeddings import HuggingFaceEmbeddings
 from groq import Groq
 
 # Initialize Groq client
 client = Groq(api_key=os.environ.get('GroqApi'))
 
-# Initialize FAISS index
-if "vectorstore" not in st.session_state:
-    st.session_state.vectorstore = None
+# Initialize session state for scraped data
+if "scraped_data" not in st.session_state:
+    st.session_state.scraped_data = []
 
 def scrape_web_data(url):
-    """Scrape tariff data from the given URL."""
+    """Scrape text 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")
+            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()]
+            return [{'Data': line.strip()} for line in all_text.split('\n') if line.strip()]
         else:
-            st.warning(f"Failed to fetch data: {response.status}")
-            return []
+            st.write(f"Error: {response.status}")
     except Exception as e:
-        st.error(f"An error occurred: {e}")
-        return []
+        st.write(f"An error occurred: {e}")
+    return []
 
 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!")
+    """Store the scraped data in FAISS vectorstore."""
+    if not data:
+        return
+    combined_text = "\n".join([item["Data"] for item in data])
+    embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
+    vectorstore = FAISS.from_texts([combined_text], embeddings)
+    vectorstore.save_local("vectorstore/")
+
+def query_with_groq(prompt):
+    """Query Groq LLM with the provided prompt."""
+    try:
+        chat_completion = client.chat.completions.create(
+            messages=[
+                {
+                    "role": "user",
+                    "content": prompt,
+                }
+            ],
+            model="llama3-8b-8192",
+        )
+        return chat_completion.choices[0].message.content
+    except Exception as e:
+        return f"Error querying Groq: {e}"
 
 # Streamlit UI
-st.title("Electricity Bill Calculator - Pakistan")
-st.subheader("Real-Time Tariff Data Retrieval")
+st.title("Pakistani Electricity Bill Calculator with Groq RAG")
 
-# Step 1: Scraping Tariff Data
-website_url = st.text_input("Enter NEPRA Tariff URL:")
-if st.button("Scrape Tariff Data"):
+# Step 1: Scraping
+st.subheader("Step 1: Scrape Tariff Data")
+website_url = st.text_input("Enter the tariff URL:")
+if st.button("Scrape Data"):
     scraped_data = scrape_web_data(website_url)
     if scraped_data:
+        st.session_state.scraped_data = scraped_data
+        st.success(f"Scraping completed. {len(scraped_data)} items found.")
         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"])
+    else:
+        st.warning("No data found. Please check the URL.")
 
-# Step 2: Appliance Data Input
+# Step 2: Appliance Inputs
 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
-]
+if "appliance_data" not in st.session_state:
+    st.session_state.appliance_data = []
 
-# 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)
+appliance_name = st.text_input("Appliance Name")
+appliance_load = st.number_input("Load (Watts per Appliance)", min_value=1, value=100)
+appliance_quantity = st.number_input("Quantity", min_value=1, value=1)
+appliance_usage_hours = st.number_input("Usage Hours per Day", min_value=0.0, value=6.0)
 
-        # 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}")
+if st.button("Add Appliance"):
+    st.session_state.appliance_data.append({
+        "name": appliance_name,
+        "load": appliance_load,
+        "quantity": appliance_quantity,
+        "usage_hours": appliance_usage_hours
+    })
+    st.success(f"Added {appliance_quantity} {appliance_name}(s) to the list!")
+
+# Display Appliance List
+if st.session_state.appliance_data:
+    st.subheader("Appliance List")
+    for idx, appliance in enumerate(st.session_state.appliance_data):
+        st.write(f"{idx+1}. {appliance['name']} - {appliance['quantity']} units, "
+                 f"{appliance['load']}W each, {appliance['usage_hours']} hours/day")
+
+# Step 3: Query the RAG system
+st.subheader("Step 3: Query Tariff Information")
+user_query = st.text_input("Enter your query about the tariff:")
+if st.button("Ask Query"):
+    if not st.session_state.scraped_data:
+        st.warning("Please scrape tariff data first.")
     else:
-        st.error("No tariff data available. Please scrape tariff data first.")
+        # Load vectorstore
+        embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
+        vectorstore = FAISS.load_local("vectorstore/", embeddings)
+        retriever = vectorstore.as_retriever()
+
+        # Combine scraped data for context
+        context = "\n".join([item["Data"] for item in st.session_state.scraped_data])
+        prompt = f"Context: {context}\n\nUser Query: {user_query}\nAnswer:"
+        response = query_with_groq(prompt)
+        st.write("**Answer:**")
+        st.write(response)
 
+# Step 4: Calculate Bill
+st.subheader("Step 4: Calculate Bill")
+tariff_rate = st.number_input("Enter Rate per Unit (PKR)", min_value=0.0, value=25.0)
+if st.button("Calculate Bill"):
+    if not st.session_state.appliance_data:
+        st.warning("Please add at least one appliance.")
+    else:
+        total_energy_kwh = sum(
+            (appliance["load"] * appliance["quantity"] * appliance["usage_hours"] * 30) / 1000
+            for appliance in st.session_state.appliance_data
+        )
+        monthly_bill = total_energy_kwh * tariff_rate
+        st.write(f"**Total Monthly Electricity Bill: PKR {monthly_bill:.2f}**")