task.py

import pandas as pd
from sklearn.preprocessing import LabelEncoder
from sklearn.model_selection import train_test_split
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import LSTM, Dense, Dropout

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder

# Load Smart Home Dataset
def load_data(partition_id, num_partitions):
    # Load both datasets to ensure consistent label encoding
    data_1 = pd.read_csv("hh101.ann.features.csv")
    data_2 = pd.read_csv("hh102.ann.features.csv")

    # Combine data to fit the LabelEncoder with all possible classes
    combined_data = pd.concat([data_1, data_2], axis=0)

    # Fit the LabelEncoder on the combined data to ensure consistent encoding
    label_encoder = LabelEncoder()
    label_encoder.fit(combined_data['activity'])

    # Load the correct partition's dataset 
    if partition_id == 0:
        data = data_1
    else:
        data = data_2

    # Separate features and target variable
    X = data.drop(columns=['activity'])
    y = data['activity']

    # Normalize the features column-wise
    X = X.apply(lambda x: x / (x.max() + 1e-8), axis=0)  # Adding a small constant to avoid division by zero

    # Encode the labels using the previously fitted LabelEncoder
    y = label_encoder.transform(y)  # Consistent encoding across both clients

    # Debug: Check the range of the encoded labels
    print(f"Max label value for partition {partition_id}: {y.max()}")

    # Split into training and test sets (80% train, 20% test)
    X_train, X_test, y_train, y_test = train_test_split(X.values, y, test_size=0.2, random_state=42)

    # Reshape X for LSTM input
    X_train = X_train.reshape((X_train.shape[0], X_train.shape[1], 1))
    X_test = X_test.reshape((X_test.shape[0], X_test.shape[1], 1))

    return X_train, y_train, X_test, y_test


# Define LSTM Model for Multi-Class Classification
def load_model(input_shape=(45, 1), num_classes=37):  # Updated to 37 classes
    model = Sequential()
    model.add(LSTM(128, input_shape=input_shape, return_sequences=False))
    model.add(Dropout(0.2))
    model.add(Dense(num_classes, activation='softmax'))  # Multi-class classification with softmax

    model.compile(
        optimizer='adam',
        loss='sparse_categorical_crossentropy',  # Multi-class classification loss
        metrics=['accuracy']
    )
    return model