app.py

import streamlit as st
import numpy as np
import tensorflow as tf
from tensorflow.keras.models import load_model
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
import tensorflow as tf
from tensorflow.keras.models import Model
from tensorflow.keras.layers import (Input, Conv1D, MaxPooling1D, Bidirectional, LSTM, 
                                     Dense, GlobalAveragePooling1D, Attention)

# Assuming your one_hot_encode and pad_sequence functions are defined as before

def one_hot_encode(seq):
    mapping = {'A': [1, 0, 0, 0], 
               'C': [0, 1, 0, 0], 
               'G': [0, 0, 1, 0], 
               'T': [0, 0, 0, 1]}
    return np.array([mapping[i] for i in seq])

def pad_sequence(encoded_seq, max_len=1000):  # Adjust max_len based on your model
    padding_needed = max_len - len(encoded_seq)
    return np.pad(encoded_seq, ((0, padding_needed), (0, 0)), 'constant')

# Load the model
model_path = "Model/ssr_model_v2.h5"  # Update the path to where your model is saved
model = load_model(model_path)

# Define the label encoder classes based on your dataset labels
class_labels = ['di', 'hexa', 'penta', 'tetra', 'tri']

st.title("DNA Sequence Classification")

user_sequence = st.text_area("Enter DNA sequence:", "Type or paste sequence here...")

if st.button("Classify"):
    if user_sequence:
        # Preprocess the input sequence
        encoded_sequence = one_hot_encode(user_sequence.upper())
        padded_sequence = pad_sequence(encoded_sequence)
        sequence_to_classify = np.expand_dims(padded_sequence, axis=0)
        
        # Predict
        prediction = model.predict(sequence_to_classify)
        predicted_index = np.argmax(prediction, axis=1)[0]
        predicted_label = class_labels[predicted_index]
        
        st.write(f"Predicted Class: {predicted_label}")
    else:
        st.write("Please enter a valid sequence.")