diff --git a/.gitignore b/.gitignore index 0dbf2f2..173a377 100644 --- a/.gitignore +++ b/.gitignore @@ -168,3 +168,12 @@ cython_debug/ # option (not recommended) you can uncomment the following to ignore the entire idea folder. #.idea/ +HYS/Edf/* +HYS/event/* +HYS/Output/* +HYS/stage/* +ZDWY/Edf/* +ZDWY/event/* +ZDWY/Output/* +ZDWY/stage/* +.idea/* \ No newline at end of file diff --git a/HYS/ReadEDF.py b/HYS/ReadEDF.py new file mode 100644 index 0000000..2bc0efd --- /dev/null +++ b/HYS/ReadEDF.py @@ -0,0 +1,148 @@ +import os +import pyedflib +import numpy as np +import pandas as pd +import shutil +from scipy.signal import resample + +def writefile(filename,data,style='w'): + """ + 函数说明： + 在指定地点创建一个文件，写入data数据 + :param filename: 需写入的文件地址和名称 + :param data: 需写入的数据 + :return: 无返回 + """ + data = data.astype('str') #将data数组的数据转换成字符串 + with open(filename,style) as f : + for str in data : + f.write(str+'\n') + return 0 + +def event_handle(df): + if df.columns[0].strip() == 'Event type': + df['Event type'] = df['Event type'].replace('μ-arousal', 'u-arousal') + elif df.columns[0].strip() == '事件类型': + column_mapping = { + "事件类型\r\r": "Event type", + "睡眠期\r\r": "Stage", + "时间\r\r": "Time", + "屏\r": "Epoch", + "日期\r\r": "Date", + "持续时间\r\r": "Duration", + "体位记录.": "Body Position", + "判定\r\r": "Validation" + } # 列名映射 + df = df.rename(columns=column_mapping) # 重命名列 + target_columns = [ + "Event type", "Stage", "Time", "Epoch", "Date", "Duration", + "HR bef.", "HR extr.", "HR delta", "O2 bef.", "O2 min.", + "O2 delta", "Body Position", "Validation", "Unnamed: 14" + ] # 添加目标文件中要求的其他列，并赋值为None + df = df[target_columns] # 按目标列顺序排序 + + df['Event type'] = df['Event type'].replace('腿动', 'Leg Movement') + df['Event type'] = df['Event type'].replace('低通气', 'Hypopnea') + df['Event type'] = df['Event type'].replace('混合性暂停', 'Mixed apnea') + df['Event type'] = df['Event type'].replace('中枢性暂停', 'Central apnea') + df['Event type'] = df['Event type'].replace('阻塞性暂停', 'Obstructive apnea') + df['Event type'] = df['Event type'].replace('氧减', 'Desaturation') + df['Event type'] = df['Event type'].replace('长RR间距', 'Long RR') + df['Event type'] = df['Event type'].replace('心率上升', 'Heart Rate Rise') + df['Event type'] = df['Event type'].replace('心率下降', 'Heart Rate Drop') + df['Event type'] = df['Event type'].replace('PTT下降', 'PTT drop') + df['Event type'] = df['Event type'].replace('鼾声', 'Snore') + df['Event type'] = df['Event type'].replace('微觉醒', 'u-arousal') + else: + assert (df.columns[0].strip() == '事件类型') + + return df + +def upsample_resample(data,data_fs,target_fs): + up_ratio = target_fs / data_fs + target_num = int(len(data) * up_ratio) + + # 执行重采样 + if up_ratio > 1: + data_up = resample(data, target_num) + target_fs = int(target_fs) + else: + print("上采样倍数不大于1，不进行上采样操作") + data_up = data + target_fs = int(data_fs) + + return data_up,target_fs + +def read_edf(edfpath,stagepath,eventpath,outpath): + list = np.array([f for f in os.listdir(edfpath) if f.endswith('.edf')]) + for sample_num in list: + print("样本编号：", sample_num) + + output_sample = outpath + sample_num.split('.')[0][-4:] + '/' + if not os.path.exists(output_sample): + os.makedirs(output_sample) + + ##读取edf文件数据 + fpath = edfpath + sample_num + shutil.copy(fpath, output_sample) + with pyedflib.EdfReader(fpath) as file: + ##读取导出的睡眠分期标签csv文件 + df = pd.read_csv(stagepath + sample_num.split('.')[0] + '_Stage.csv') + stage_values = df.values[:,0].reshape(-1) + + ##读edf文件信息 + start_time = file.getStartdatetime() + sec = file.getFileDuration() + day_start = start_time.strftime("%Y") + "/" + str(start_time.month) + "/" + str(start_time.day) + sec_start = start_time.strftime("%H") + ":" + start_time.strftime("%M") + ":" + start_time.strftime("%S") + # sec_start = str(start_time.hour) + ":" + str(f"{start_time.minute:02d}") + ":" + str(f"{start_time.second:02d}") + assert (df.columns[3] == day_start) + assert (df.columns[4] == sec_start) + print("edf文件与睡眠分期标签开始时间一致！！！！！！！") + + ##读取Event列表 + df1 = pd.read_csv(eventpath + sample_num.split('.')[0] + '_Event.csv',header=0,encoding='GBK') + df1 = event_handle(df1) + df1.to_csv(output_sample + 'SA Label_Raw.csv', index=False, encoding='GBK') + + signal_num = file.signals_in_file + signal_label = file.getSignalLabels() + min_length = int(np.min([sec,df.shape[0]])) + print("裁剪为最小长度:",min_length) + flag = 0 + for i, index in enumerate(signal_label): + signal = file.readSignal(i) + sample_frequency = int(file.getSampleFrequency(i)) + data_org = signal[0:min_length * sample_frequency] + + if (index== 'Flow Patient') : + if flag == 0: + writefile(output_sample + 'Flow T_Raw_' + str(sample_frequency) + '.txt', data_org) + flag = 1 + else: + writefile(output_sample + 'Flow P_Raw_' + str(sample_frequency) + '.txt', data_org) + + else: + writefile(output_sample + index + '_Raw_' + str(sample_frequency) + '.txt', data_org) + with open(os.path.join(output_sample, "StartTime_Raw.txt"), "a") as text_file: + text_file.write(start_time.strftime("%Y-%m-%d %H:%M:%S\n")) + text_file.write("{:g}\n".format(int(sec))) + + ##裁剪对齐睡眠分期标签 + stage_values = stage_values.astype(int) + stage_values = np.where(stage_values == 11, 'W', stage_values) + stage_values = np.where(stage_values == '12', 'R', stage_values) + stage_values = np.where(stage_values == '13', 'N1', stage_values) + stage_values = np.where(stage_values == '14', 'N2', stage_values) + stage = np.where(stage_values == '15', 'N3', stage_values) + writefile(output_sample + '5_class_Raw_1.txt',stage[0:min_length]) + +if __name__ == '__main__': + edf_dir = 'E:\DB\\testdata\HYS\PSG_Origin/Edf/' + stage_dir = 'E:\DB\\testdata\HYS\PSG_Origin/Stage/' + event_dir = 'E:\DB\\testdata\HYS\PSG_Origin/Event/' + output_dir = 'E:\DB\\testdata\HYS\PSG_Text/' + + read_edf(edf_dir,stage_dir,event_dir,output_dir) + + print("finished") diff --git a/ZDWY/ZDWY_ReadEDF.py b/ZDWY/ZDWY_ReadEDF.py new file mode 100644 index 0000000..4ea652f --- /dev/null +++ b/ZDWY/ZDWY_ReadEDF.py @@ -0,0 +1,143 @@ +import os +import pyedflib +import numpy as np +import pandas as pd +import shutil +from scipy.signal import resample + +def writefile(filename,data,style='w'): + """ + 函数说明： + 在指定地点创建一个文件，写入data数据 + :param filename: 需写入的文件地址和名称 + :param data: 需写入的数据 + :return: 无返回 + """ + data = data.astype('str') #将data数组的数据转换成字符串 + with open(filename,style) as f : + for str in data : + f.write(str+'\n') + return 0 + +def upsample_resample(data,data_fs,target_fs): + up_ratio = target_fs / data_fs + target_num = int(len(data) * up_ratio) + + # 执行重采样 + if up_ratio > 1: + data_up = resample(data, target_num) + target_fs = int(target_fs) + else: + print("上采样倍数不大于1，不进行上采样操作") + data_up = data + target_fs = int(data_fs) + + return data_up,target_fs + +def event_handle(df): + df['类型'] = df['类型'].replace('腿部运动', 'Leg Movement') + df['类型'] = df['类型'].replace('低通气', 'Hypopnea') + df['类型'] = df['类型'].replace('混合型呼吸暂停', 'Mixed apnea') + df['类型'] = df['类型'].replace('中枢性呼吸暂停', 'Central apnea') + df['类型'] = df['类型'].replace('阻塞性呼吸暂停', 'Obstructive apnea') + df['类型'] = df['类型'].replace('相对氧减', 'Desaturation') + df['类型'] = df['类型'].replace('长 RR', 'Long RR') + df['类型'] = df['类型'].replace('心率上升', 'Heart Rate Rise') + df['类型'] = df['类型'].replace('心率下降', 'Heart Rate Drop') + df['类型'] = df['类型'].replace('PTT 下降', 'PTT drop') + df['类型'] = df['类型'].replace('打鼾', 'Snore') + df['类型'] = df['类型'].replace('觉醒', 'u-arousal') + df['睡眠期'] = df['睡眠期'].replace('WK', 'W') + df['睡眠期'] = df['睡眠期'].replace('REM', 'R') + + column_mapping = { + "类型": "Event type", + "睡眠期": "Stage", + "时间": "Time", + "时期": "Epoch", + "持续时间": "Duration", + "体位": "Body Position", + "验证": "Validation" + } # 列名映射 + df = df.rename(columns=column_mapping) # 重命名列 + target_columns = [ + "Event type", "Stage", "Time", "Epoch", "Date", "Duration", + "HR bef.", "HR extr.", "HR delta", "O2 bef.", "O2 min.", + "O2 delta", "Body Position", "Validation", "Unnamed: 14" + ] # 添加目标文件中要求的其他列，并赋值为None + for col in target_columns: # 添加缺失的列并用None填充 + if col not in df.columns: + df[col] = None + df = df[target_columns] # 按目标列顺序排序 + + return df + +def read_edf(edfpath,stagepath,eventpath,outpath): + list = np.array([f for f in os.listdir(edfpath) if f.endswith('.edf')]) + for sample_num in list: + print("样本编号：", sample_num) + + output_sample = outpath + sample_num.split('.')[0] + '/' + if not os.path.exists(output_sample): + os.makedirs(output_sample) + + ##读取edf文件数据 + fpath = edfpath + sample_num + shutil.copy(fpath, output_sample) + with pyedflib.EdfReader(fpath) as file: + ##读取导出的睡眠分期标签csv文件 + df = pd.read_csv(stagepath + sample_num.split('.')[0] + '_Stage.csv') + stage_values = df.values[:,2].reshape(-1) + + ##读edf文件信息 + start_time = file.getStartdatetime() + sec = file.getFileDuration() + sec_start = start_time.strftime("%H") + ":" + start_time.strftime("%M") + ":" + start_time.strftime("%S") + assert (df.values[0,1] == sec_start) + print("edf文件与睡眠分期标签开始时间一致！！！！！！！") + + ##读取Event列表 + df1 = pd.read_csv(eventpath + sample_num.split('.')[0] + '_Event.csv', header=0, encoding='utf-8') + df1 = event_handle(df1) + df1.to_csv(output_sample + 'SA Label_Raw.csv', index=False, encoding='GBK') + + signal_num = file.signals_in_file + signal_label = file.getSignalLabels() + min_length = int(np.min([sec,df.shape[0]*30])) + print("裁剪为最小长度:",min_length) + flag = 0 + for i, index in enumerate(signal_label): + signal = file.readSignal(i) + sample_frequency = int(file.getSampleFrequency(i)) + data_org = signal[0:min_length * sample_frequency] + + if (index== 'Flow Patient') : + if flag == 0: + writefile(output_sample + 'Flow T_Raw_' + str(sample_frequency) + '.txt', data_org) + flag = 1 + else: + writefile(output_sample + 'Flow P_Raw_' + str(sample_frequency) + '.txt', data_org) + else: + writefile(output_sample + index + '_Raw_' + str(sample_frequency) + '.txt', data_org) + with open(os.path.join(output_sample, "StartTime_Raw.txt"), "a") as text_file: + text_file.write(start_time.strftime("%Y-%m-%d %H:%M:%S\n")) + text_file.write("{:g}\n".format(int(sec))) + + ##裁剪对齐睡眠分期标签 + stage_values = np.where(stage_values == 'WK', 'W', stage_values) + stage_values = np.where(stage_values == 'REM', 'R', stage_values) + stage_values = np.where(stage_values == 'N1', 'N1', stage_values) + stage_values = np.where(stage_values == 'N2', 'N2', stage_values) + stage = np.where(stage_values == 'N3', 'N3', stage_values) + stage = stage.repeat(30) + writefile(output_sample + '5_class_Raw_1.txt',stage[0:min_length]) + +if __name__ == '__main__': + edf_dir = 'E:\DB\\testdata\ZD5Y\PSG_Origin/Edf/' + stage_dir = 'E:\DB\\testdata\ZD5Y\PSG_Origin/Stage/' + event_dir = 'E:\DB\\testdata\ZD5Y\PSG_Origin/Event/' + output_dir = 'E:\DB\\testdata\ZD5Y\PSG_Text/' + + read_edf(edf_dir,stage_dir,event_dir,output_dir) + + print("finished") diff --git a/requirements.txt b/requirements.txt new file mode 100644 index 0000000..1f7261d --- /dev/null +++ b/requirements.txt @@ -0,0 +1,5 @@ +mne==1.9.0 +numpy==2.2.5 +pandas==2.2.3 +pyEDFlib==0.1.38 +xlwt==1.3.0