app.py

import os
import streamlit as st
import pandas as pd
import numpy as np
from groq import Groq
from sentence_transformers import SentenceTransformer
import faiss

GROQ_API_KEY = "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"
# Initialize Groq API Client
client = Groq(api_key=GROQ_API_KEY)

# Load Pretrained Embedding Model
embedding_model = SentenceTransformer("all-MiniLM-L6-v2")

# Initialize global variables
uploaded_file = None
faiss_index = None
dataframe = None

# Function to load and preprocess dataset
def load_dataset(file):
    global dataframe, faiss_index
    dataframe = pd.read_csv(file)
    st.success("Dataset loaded successfully!")
    # Create FAISS index for the dataset
    embeddings = embedding_model.encode(dataframe["Energy Usage (kWh)"].astype(str).tolist())
    faiss_index = faiss.IndexFlatL2(embeddings.shape[1])
    faiss_index.add(np.array(embeddings))

# Function to retrieve relevant rows
def retrieve_relevant_data(query, top_k=5):
    if dataframe is None or faiss_index is None:
        st.error("Please upload a dataset first.")
        return []
    query_embedding = embedding_model.encode([query])
    distances, indices = faiss_index.search(np.array(query_embedding), top_k)
    relevant_rows = dataframe.iloc[indices[0]].to_dict(orient="records")
    return relevant_rows

# Function to analyze data and generate cost-saving recommendations
def generate_cost_saving_recommendations(data):
    if not data:
        return "No relevant data found for recommendations."
    
    total_energy_usage = sum(row["Energy Usage (kWh)"] for row in data)
    avg_energy_usage = total_energy_usage / len(data)
    total_cost = sum(row["Cost"] for row in data)
    
    recommendations = [
        f"Total energy usage: {total_energy_usage:.2f} kWh",
        f"Average energy usage per household: {avg_energy_usage:.2f} kWh",
        f"Total cost: ${total_cost:.2f}",
        "Recommendations:",
        "- Implement energy-efficient appliances.",
        "- Use renewable energy sources like solar or wind.",
        "- Schedule high-energy tasks during off-peak hours.",
        "- Conduct regular maintenance to reduce energy wastage."
    ]
    return "\n".join(recommendations)

# Function to generate a detailed analysis report
def generate_report(query):
    relevant_data = retrieve_relevant_data(query)
    context = "\n".join([str(row) for row in relevant_data])
    chat_completion = client.chat.completions.create(
        messages=[
            {
                "role": "user",
                "content": f"Based on the following query: '{query}' and context:\n{context}\nProvide an energy usage analysis report."
            }
        ],
        model="llama3-8b-8192",
        stream=False,
    )
    detailed_report = chat_completion.choices[0].message.content
    cost_saving_recommendations = generate_cost_saving_recommendations(relevant_data)
    return detailed_report, cost_saving_recommendations

# Streamlit app configuration
st.title("Energy Usage Analysis & Cost-Saving Report Generator")
st.sidebar.header("Upload Dataset")

# File upload
uploaded_file = st.sidebar.file_uploader("Upload your energy usage dataset (CSV)", type=["csv"])
if uploaded_file is not None:
    load_dataset(uploaded_file)

# Query input
st.header("Generate Energy Usage Report")
query = st.text_input("Enter your query (e.g., 'Analyze peak usage times in urban areas')")

if st.button("Generate Report"):
    if uploaded_file is None:
        st.error("Please upload a dataset first.")
    elif query.strip() == "":
        st.error("Please enter a query.")
    else:
        with st.spinner("Generating report..."):
            try:
                detailed_report, cost_saving_recommendations = generate_report(query)
                st.subheader("Energy Usage Analysis Report")
                st.write(detailed_report)
                st.subheader("Cost-Saving Recommendations")
                st.write(cost_saving_recommendations)
            except Exception as e:
                st.error(f"An error occurred: {e}")