app.py

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

# Load pre-trained Sentence Transformer model
model = SentenceTransformer('all-MiniLM-L6-v2')

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

# Generate embeddings using Sentence Transformers
def generate_embeddings(text):
    return model.encode(text)

# Build FAISS index for retrieval
def build_faiss_index(data):
    index = faiss.IndexFlatL2(384)  # 384-dimensional embeddings for MiniLM
    embeddings = [generate_embeddings(row.to_string()) for _, row in data.iterrows()]
    embeddings = np.array(embeddings).astype("float32")
    index.add(embeddings)
    return index, embeddings

# Query FAISS index
def query_index(query, data, index):
    query_embedding = generate_embeddings(query).astype("float32")
    distances, indices = index.search(np.array([query_embedding]), k=5)
    results = data.iloc[indices[0]]
    return results

# Generate a detailed report using Groq's generative model
def generate_report_with_groq(query, results):
    input_text = f"Based on the query '{query}', the following insights are generated:\n\n{results.to_string(index=False)}"
    response = client.chat.completions.create(
        messages=[{"role": "user", "content": input_text}],
        model="llama3-8b-8192",
        stream=False
    )
    return response.choices[0].message.content

# Main function to execute the workflow
if __name__ == "__main__":
    # Load dataset
    csv_path = "energy_usage_data.csv"  # Ensure this CSV is uploaded to your working directory
    data = pd.read_csv(csv_path)

    # Preprocess data (if needed)
    data.fillna("", inplace=True)

    # Build FAISS index
    print("Building FAISS index...")
    index, embeddings = build_faiss_index(data)

    # User query
    query = "Show households with high energy usage in the North region"
    print(f"User Query: {query}")

    # Query FAISS index
    print("Retrieving relevant data...")
    results = query_index(query, data, index)

    # Generate report
    print("Generating report using Groq API...")
    report = generate_report_with_groq(query, results)

    # Output the report
    print("Generated Report:\n")
    print(report)