AIModel.py

import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Embedding, LSTM, Dense
from tensorflow.keras.preprocessing.text import Tokenizer
from tensorflow.keras.preprocessing.sequence import pad_sequences
from sklearn.model_selection import train_test_split
import numpy as np

# Preprocess Data
def preprocess_data(conversations, max_len=20):
    questions = [q for q, _ in conversations]
    answers = [a for _, a in conversations]

    # Tokenize the text data
    tokenizer = Tokenizer()
    tokenizer.fit_on_texts(questions + answers)

    X = tokenizer.texts_to_sequences(questions)
    y = tokenizer.texts_to_sequences(answers)

    # Pad sequences to ensure uniform input size
    X = pad_sequences(X, maxlen=max_len, padding='post')
    y = pad_sequences(y, maxlen=max_len, padding='post')

    return X, y, tokenizer

# Build RNN Model
def build_rnn_model(vocab_size, max_len):
    model = Sequential([
        Embedding(input_dim=vocab_size, output_dim=128, input_length=max_len),
        LSTM(128, return_sequences=False),
        Dense(vocab_size, activation='softmax')
    ])
    model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])
    return model

# Load and preprocess data
conversations = [
    ("How are you?", "I am fine."),
    ("What is your name?", "I am an AI assistant."),
    ("Tell me a joke.", "Why did the chicken cross the road? To get to the other side."),
    # Add more training data as needed
]

# Preprocessing
max_len = 20  # Set maximum length for padding
X, y, tokenizer = preprocess_data(conversations, max_len)
vocab_size = len(tokenizer.word_index) + 1

# Split data into training and testing sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

# Build and train the RNN model
model = build_rnn_model(vocab_size, max_len)
model.fit(X_train, y_train, epochs=10, batch_size=32, validation_data=(X_test, y_test))

# Save the model and tokenizer
model.save('ai_assistant_rnn_model.h5')
with open('tokenizer.pickle', 'wb') as f:
    import pickle
    pickle.dump(tokenizer, f)