import numpy as np
import tensorflow as tf
from tensorflow import keras
from keras.layers import Conv2D, Dense, Embedding, Flatten, LSTM, MaxPooling2D
from keras.models import Sequential
from xgboost import XGBClassifier

# (a) Prediksi Stok Barang (LSTM)
# Input shape: (samples, time_steps=12, features=1)
model_a = Sequential([
    LSTM(32, input_shape=(12, 1)),
    Dense(1)
])
model_a.compile(optimizer='adam', loss='mse')

# (b) Klasifikasi Foto Produk Rusak vs Normal (CNN)
# Input shape contoh citra RGB 128x128
model_b = Sequential([
    Conv2D(32, (3, 3), activation='relu', input_shape=(128, 128, 3)),
    MaxPooling2D((2, 2)),
    Flatten(),
    Dense(64, activation='relu'),
    Dense(1, activation='sigmoid')  # binary: rusak vs normal
])
model_b.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])

# (c) Analisis Sentimen Ulasan Play Store (LSTM + Embedding)
vocab_size = 5000
max_length = 100

model_c = Sequential([
    Embedding(input_dim=vocab_size, output_dim=64, input_length=max_length),
    LSTM(64),
    Dense(32, activation='relu'),
    Dense(1, activation='sigmoid')  # binary: positif vs negatif
])
model_c.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])

# (d) Prediksi Churn Pelanggan Tabular (ML Klasik - XGBoost)
# X_train memiliki dimensi (n_samples, 15)
model_d = XGBClassifier(
    n_estimators=100,
    max_depth=4,
    learning_rate=0.1,
    random_state=42
)
# model_d.fit(X_train, y_train)