app.py

import streamlit as st
import torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification

# Load the model and tokenizer (make sure your model is correctly loaded here)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

# Specify the model path
model_name = "ipc_refined_approach_model"  # Replace with your actual model path or name
model = AutoModelForSequenceClassification.from_pretrained(model_name)
model = model.to(device)

tokenizer = AutoTokenizer.from_pretrained(model_name)

# Define your legal sections
sections = ['465', '467', '395', '332','353']  # Example sections, modify as per your actual list

# Streamlit UI setup
st.title("Legal Case Section Prediction")

# Get input text from user
st.subheader("Enter the legal text to predict the sections it belongs to:")
input_text = st.text_area("Input Text", height=250)

# Prediction function
def predict_text(text):
    # Tokenize and encode input text
    inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True, max_length=512)
    
    # Move inputs to the same device as the model
    inputs = {key: value.to(device) for key, value in inputs.items()}
    
    # Perform inference
    model.eval()
    with torch.no_grad():
        outputs = model(**inputs)
        logits = outputs.logits  # Ensure logits are accessed correctly
    
    # Apply sigmoid to get probabilities
    probs = torch.sigmoid(logits).detach().cpu().numpy()  # Move to CPU for processing
    
    # Convert probabilities to binary predictions (threshold 0.5)
    predictions = {section: int(prob > 0.5) for section, prob in zip(sections, probs[0])}
    
    # Return the sections the case belongs to
    sections_belongs_to = [section for section, pred in predictions.items() if pred == 1]
    return sections_belongs_to

# Show results if input text is provided
if input_text:
    st.subheader("Prediction Results")
    
    # Get predictions for the input text
    predicted_sections = predict_text(input_text)
    
    # Show predictions
    if predicted_sections:
        st.write(f"This case belongs to Section(s): {', '.join(predicted_sections)}")
    else:
        st.write("This case does not belong to any known section.")
else:
    st.write("Please enter some text to predict the sections.")