SWIT/Datatset_Image_Encoding.py at main · pms5343/SWIT · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
import torch
from collections import defaultdict
import numpy as np
from sklearn.preprocessing import MinMaxScaler
from pyts.image import GramianAngularField
import pandas as pd
import os
from collections import Counter

def save_user_data(X_list, y_list, user_name):
    X_array = np.concatenate(X_list, axis=0)
    y_array = np.concatenate(y_list, axis=0)
    torch.save((torch.tensor(X_array, dtype=torch.float32), torch.tensor(y_array, dtype=torch.long)), f'{user_name}.pt')

# def segment_and_convert_to_gaf_mtf(raw_df, start_time, end_time, step=10, segment_length=200):
#     segment_df = raw_df[(raw_df['Time'] >= start_time) & (raw_df['Time'] <= end_time)]
#     gaf = GramianAngularField(image_size=segment_length, method='difference')
#     transformed_segments = []
#     scaler = MinMaxScaler(feature_range=(-1, 1))
#     for start_idx in range(0, len(segment_df) - segment_length + 1, step):
#         segment = segment_df.iloc[start_idx:start_idx+segment_length]
#         sensor_data = segment[['Ax', 'Ay', 'Az', 'Gx', 'Gy', 'Gz']].values
#         sensor_data_normalized = scaler.fit_transform(sensor_data)
#         temp_gaf_mtf = np.zeros((6, segment_length, segment_length))
#         for i in range(sensor_data.shape[1]):
#             gaf_transformed = gaf.fit_transform(sensor_data_normalized[:, i].reshape(1, -1))
#             temp_gaf_mtf[i] = gaf_transformed[0]
#         transformed_segments.append(temp_gaf_mtf)
#     return np.stack(transformed_segments, axis=0) if transformed_segments else np.array([])

# def load_and_process_task_data(task_number, user_type, task_mapping):
#     task_raw_data_file1 = f'/home/worker/SWIT_Dataset/Task_{task_number}_{user_type}.csv'

#     task_raw_df1 = pd.read_csv(task_raw_data_file1)

#     if task_number in [8, 6]:
#         start_time = 0
#         end_time = 20
#         gaf_data = segment_and_convert_to_gaf_mtf(task_raw_df1, start_time, end_time, step=10)
#         if gaf_data.size > 0:
#             X = np.concatenate([gaf_data], axis=0)
#             y = np.full(X.shape[0], task_mapping.get(task_number, task_number))
#             return X, y
#     else:
#         st_task_file = f'/home/worker/SWIT_Dataset/Segmentation_Status/ST_Task_{task_number}.csv'
#         if not os.path.exists(st_task_file):
#             print(f"File {st_task_file} does not exist.")
#             return None, None

#         st_task_df = pd.read_csv(st_task_file)
#         user_starts_ends = st_task_df.loc[st_task_df['Status'].str.contains('Start|End'), user_type].values
#         adjusted_times = [(time * 0.01) - 1 if idx % 2 == 0 else (time * 0.01) + 1 for idx, time in enumerate(user_starts_ends)]
#         adjusted_times = [max(0, time) for time in adjusted_times]

#         gaf_data_list = []
#         for i in range(0, len(adjusted_times), 2):
#             gaf_data = segment_and_convert_to_gaf_mtf(task_raw_df1, adjusted_times[i], adjusted_times[i + 1], n_bins=n_bins)
#             if gaf_data.size > 0:
#                 gaf_data_list.append(gaf_data)

#         if gaf_data_list:
#             X = np.concatenate(gaf_data_list, axis=0)
#             y = np.full(X.shape[0], task_mapping.get(task_number, task_number))
#         else:
#             X = np.array([])
#             y = np.array([])

#         return X, y


def segment_and_convert_to_gaf_mtf(raw_df, start_time, end_time, step=10, segment_length=200):
    segment_df = raw_df[(raw_df['Time'] >= start_time) & (raw_df['Time'] <= end_time)]
    gaf = GramianAngularField(image_size=segment_length, method='difference')
    transformed_segments = []
    scaler = MinMaxScaler(feature_range=(-1, 1))
    for start_idx in range(0, len(segment_df) - segment_length + 1, step):
        segment = segment_df.iloc[start_idx:start_idx+segment_length]
        sensor_data = segment[['Ax', 'Ay', 'Az', 'Gx', 'Gy', 'Gz']].values
        sensor_data_normalized = scaler.fit_transform(sensor_data)
        temp_gaf_mtf = np.zeros((6, segment_length, segment_length))
        for i in range(sensor_data.shape[1]):
            gaf_transformed = gaf.fit_transform(sensor_data_normalized[:, i].reshape(1, -1))
            temp_gaf_mtf[i] = gaf_transformed[0]
        transformed_segments.append(temp_gaf_mtf)
    return np.stack(transformed_segments, axis=0) if transformed_segments else np.array([])

def load_and_process_task_data(task_number, user_type, task_mapping):
    task_raw_data_file1 = f'/home/worker/SWIT_Dataset/Task_{task_number}_{user_type}.csv'
    task_raw_df1 = pd.read_csv(task_raw_data_file1)

    st_task_file = f'/home/worker/SWIT_Dataset/Segmentation_Status/ST_Task_{task_number}.csv'
    if not os.path.exists(st_task_file):
        print(f"File {st_task_file} does not exist.")
        return None, None

    st_task_df = pd.read_csv(st_task_file)
    user_starts_ends = st_task_df.loc[st_task_df['Status'].str.contains('Start|End'), user_type].values
    adjusted_times = [(time * 0.01) - 1 if idx % 2 == 0 else (time * 0.01) + 1 for idx, time in enumerate(user_starts_ends)]
    adjusted_times = [max(0, time) for time in adjusted_times]

    gaf_data_list = []
    for i in range(0, len(adjusted_times), 2):
        gaf_data = segment_and_convert_to_gaf_mtf(task_raw_df1, adjusted_times[i], adjusted_times[i + 1], n_bins=n_bins)
        if gaf_data.size > 0:
            gaf_data_list.append(gaf_data)

    if gaf_data_list:
        X = np.concatenate(gaf_data_list, axis=0)
        y = np.full(X.shape[0], task_mapping.get(task_number, task_number))
    else:
        X = np.array([])
        y = np.array([])

    return X, y

def process_task_data_for_user(task_number, user_id, task_mapping):
    X_list, y_list = [], []
    user_data_count = defaultdict(int)

    X, y = load_and_process_task_data(task_number, user_id, task_mapping)
    if X is not None and X.size > 0:
        X_list.append(X)
        y_list.append(y)
        user_data_count[task_number] += X.shape[0]

    return X_list, y_list, user_data_count

def balance_classes(X_list, y_list):
    combined_X = np.concatenate(X_list, axis=0)
    combined_y = np.concatenate(y_list, axis=0)

    counter = Counter(combined_y)
    min_count = min(counter.values())

    balanced_X, balanced_y = [], []
    for label in counter.keys():
        indices = np.where(combined_y == label)[0]
        selected_indices = np.random.choice(indices, min_count, replace=False)
        balanced_X.append(combined_X[selected_indices])
        balanced_y.append(combined_y[selected_indices])

    balanced_X = np.concatenate(balanced_X, axis=0)
    balanced_y = np.concatenate(balanced_y, axis=0)

    return balanced_X, balanced_y

tasks_to_process = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
task_mapping = {1: 0, 2: 1, 3: 2, 4: 3, 5: 4, 6: 5, 7: 6, 8: 7, 9: 8, 10: 9}

# List of users
user_ids = ['User_1', 'User_2', 'User_3', 'User_4', 'User_5', 'User_6', 'User_7', 'User_8', 'User_9', 'User_10',
            'User_11', 'User_12', 'User_13','User_14', 'User_15', 'User_16', 'User_17', 'User_18', 'User_19', 'User_20',
            'User_21', 'User_22', 'User_23', 'User_24', 'User_25', 'User_26', 'User_27']

user_data_counts = defaultdict(lambda: defaultdict(int))  # Use nested defaultdict

for user_id in user_ids:
    user_X_list, user_y_list = [], []

    for task_number in tasks_to_process:
        X_user, y_user, task_counts = process_task_data_for_user(task_number, user_id, task_mapping)
        if X_user:
            user_X_list.extend(X_user)
            user_y_list.extend(y_user)
            for task, count in task_counts.items():
                user_data_counts[user_id][task] += count

    if user_X_list and user_y_list:
        balanced_user_X, balanced_user_y = balance_classes(user_X_list, user_y_list)
        save_user_data([balanced_user_X], [balanced_user_y], f'user_{user_id}')

for user_id, tasks in user_data_counts.items():
    for task, count in tasks.items():
        print(f"{user_id} Task {task}: {count} samples")