mpara's picture
Create app.py
f64929b verified
Raw History Blame Contribute Delete
3.15 kB
# Import necessary libraries
import os
import pandas as pd
from sentence_transformers import SentenceTransformer
import faiss
import numpy as np
import gradio as gr
from groq import Groq
# Set up Groq API
GROQ_API_KEY = "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" # Replace with your Groq API key
client = Groq(api_key=GROQ_API_KEY)
# Load environmental data (replace 'dataset.csv' with your actual file path)
df = pd.read_csv('./environmental_impact_assessment_dataset.csv') # Path to your uploaded dataset
# Combine text columns relevant to the project for embedding (adjust as per your dataset)
text_column = df['Project Type'] + ' ' + df['Mitigation Measures'] # Example of combining fields
# Use SentenceTransformer for embeddings (lighter and no API key required)
embedding_model = SentenceTransformer('all-MiniLM-L6-v2') # A lighter sentence transformer model
# Generate embeddings for the texts
embeddings = embedding_model.encode(text_column.tolist())
# Convert embeddings to a numpy array for FAISS
embeddings_np = np.array(embeddings).astype(np.float32)
# Build the FAISS index
index = faiss.IndexFlatL2(embeddings_np.shape[1]) # Create the index (L2 distance)
index.add(embeddings_np) # Add embeddings to the index
# Function to perform search and retrieval from FAISS
def retrieve_relevant_document(query):
query_embedding = embedding_model.encode([query]) # Generate embedding for the query
query_embedding_np = np.array(query_embedding).astype(np.float32) # Convert to numpy array
# Perform similarity search in the FAISS index
_, indices = index.search(query_embedding_np, k=1) # Get the top 1 most similar document
retrieved_text = text_column.iloc[indices[0][0]] # Get the corresponding text
return retrieved_text
# Function to generate a report using the Groq API
def generate_report(user_input):
# Handle cases where the user input is empty
if not user_input.strip():
return "Please provide project details to generate an Environmental Impact Assessment (EIA) report."
# Retrieve the relevant document from FAISS
relevant_document = retrieve_relevant_document(user_input)
# Use Groq API to generate the response based on the retrieved document
chat_completion = client.chat.completions.create(
messages=[
{"role": "user", "content": f"Generate an environmental impact assessment report based on the following project details:\n\n{relevant_document}\n\nUser query: {user_input}"}
],
model="llama3-8b-8192", # Use a suitable model from Groq
)
return chat_completion.choices[0].message.content
# Define the Gradio interface for user input and output
def gradio_interface(project_details):
# Call the report generation function and return the output
return generate_report(project_details)
# Set up Gradio UI
iface = gr.Interface(
fn=gradio_interface,
inputs="text", # Textbox input for project details
outputs="text", # Text output for the generated report
live=False # Set live to False for better user experience
)
# Launch the Gradio interface
iface.launch()