Datasets:
License:
| import os | |
| import re | |
| import pandas as pd | |
| from pathlib import Path | |
| from collections import defaultdict | |
| def natural_sort_key(s): | |
| """ | |
| 用于自然排序的键函数，确保10排在2后面 | |
| """ | |
| return [int(text) if text.isdigit() else text.lower() for text in re.split(r'(\d+)', s)] | |
| def determine_dataset_name(file_path): | |
| """ | |
| 从文件路径确定数据集名称 | |
| Args: | |
| file_path (str): 文件路径 | |
| Returns: | |
| str: 数据集名称 | |
| """ | |
| # 从路径中提取目录结构 | |
| path_parts = Path(file_path).parts | |
| # 尝试查找符合命名模式的部分 (如RM_001_CWRU) | |
| for part in path_parts: | |
| if re.match(r'RM_\d+_\w+', part): | |
| return part | |
| # 如果没有找到匹配的模式，则使用父目录名作为数据集名称 | |
| parent_dir = os.path.basename(os.path.dirname(file_path)) | |
| if parent_dir and parent_dir != ".": | |
| return parent_dir | |
| # 最后的备选方案是使用文件名的一部分 | |
| base_name = os.path.basename(file_path) | |
| name_parts = base_name.split('_') | |
| if len(name_parts) > 1: | |
| return name_parts[0] | |
| # 如果都不符合，返回无扩展名的文件名 | |
| return os.path.splitext(base_name)[0] | |
| def explore_datasets(root_dir): | |
| """ | |
| 遍历目录收集数据集信息，并按数据集分组 | |
| Args: | |
| root_dir (str): 数据集的根目录 | |
| Returns: | |
| list: 包含所有数据集信息的列表，已按数据集顺序排序 | |
| """ | |
| # 用于存储各数据集文件的字典 | |
| dataset_files = defaultdict(list) | |
| # 首先收集所有文件并按数据集分组 | |
| print("正在收集文件信息...") | |
| for root, dirs, files in os.walk(root_dir): | |
| # 对目录和文件进行自然排序 | |
| dirs.sort(key=natural_sort_key) | |
| files.sort(key=natural_sort_key) | |
| for file in files: | |
| # 跳过一些非数据文件 | |
| if file.startswith('.') or file == 'dataset_info.csv' or file == 'datainfo.py': | |
| continue | |
| # 检查是否在RM_004_IMS目录下，如果是，即使没有后缀也接受该文件 | |
| is_ims_file = 'RM_004_IMS' in root | |
| has_valid_extension = file.endswith(('.csv', '.mat', '.txt', '.dat', '.npy', '.npz', '.h5', '.hdf5', '.xlsx', '.xls')) | |
| # 如果文件有有效后缀或者在RM_004_IMS目录下，则处理它 | |
| if has_valid_extension or is_ims_file: | |
| file_path = os.path.join(root, file) | |
| # 确定此文件属于哪个数据集 | |
| dataset_name = determine_dataset_name(file_path) | |
| # 获取文件扩展名，如果没有则标记为'BIN'(二进制)或'DAT' | |
| ext = os.path.splitext(file)[1][1:].upper() | |
| if not ext and is_ims_file: | |
| ext = 'BIN' # 为IMS文件指定一个默认类型 | |
| dataset_files[dataset_name].append({ | |
| 'path': file_path, | |
| 'rel_path': os.path.relpath(file_path, root_dir + '/' + dataset_name), | |
| 'name': os.path.splitext(file)[0] if os.path.splitext(file)[0] else file, # 如果没有扩展名，使用完整文件名 | |
| 'file': file, | |
| 'type': ext | |
| }) | |
| # 对数据集名称进行排序，确保如RM_001在RM_002之前 | |
| datasets = [] | |
| sorted_dataset_names = sorted(dataset_files.keys(), key=natural_sort_key) | |
| # 为每个数据集分配ID并处理其文件 | |
| for dataset_id, dataset_name in enumerate(sorted_dataset_names, 1): | |
| print(f"处理数据集 {dataset_id}: {dataset_name}") | |
| file_id_in_dataset = 0 | |
| # 对每个数据集中的文件进行处理 | |
| for file_info in dataset_files[dataset_name]: | |
| file_id_in_dataset += 1 | |
| file_path = file_info['path'] | |
| rel_path = file_info['rel_path'] | |
| file = os.path.basename(file_path) | |
| file_part = file_info['file'] | |
| # 获取文件基本信息 | |
| file_info = { | |
| 'Id': len(datasets) + 1, | |
| 'Dataset_id': dataset_id, | |
| 'Name': dataset_name, | |
| 'Description': "", | |
| 'TYPE': '', | |
| 'File': rel_path, | |
| 'Visiable': 1, | |
| 'Label': '', | |
| 'Label_Description': '', | |
| 'RUL_label': '', | |
| 'RUL_label_description': '', | |
| 'Fault level': '', | |
| 'Domain_id': '', | |
| 'Domain_description': '', | |
| 'Sample_rate': '', | |
| 'Sample_lenth': '', | |
| 'Channel': '', | |
| 'Fault_Diagnosis': '', | |
| 'Anomaly_Detection': '', | |
| 'Remaining_Life': '', | |
| # 'Digital_Twin_Prediction': '', | |
| } | |
| # 从文件名和路径推断信息 | |
| path_lower = rel_path.lower() | |
| filename_lower = file.lower() | |
| # # 尝试识别数据集类型和领域 | |
| # domain_keywords = { | |
| # 'bearing': {'id': 1, 'description': '轴承故障诊断'}, | |
| # 'gear': {'id': 2, 'description': '齿轮故障诊断'}, | |
| # 'motor': {'id': 3, 'description': '电机故障诊断'}, | |
| # 'pump': {'id': 4, 'description': '泵故障诊断'}, | |
| # 'fan': {'id': 5, 'description': '风机故障诊断'}, | |
| # 'cwru': {'id': 6, 'description': 'Case Western Reserve University 轴承数据集'}, | |
| # 'nasa': {'id': 7, 'description': 'NASA 进行性退化数据集'}, | |
| # 'phm': {'id': 8, 'description': 'PHM 竞赛数据集'}, | |
| # } | |
| # for keyword, info in domain_keywords.items(): | |
| # if keyword in filename_lower or keyword in path_lower or keyword in dataset_name.lower(): | |
| # file_info['Domain_id'] = info['id'] | |
| # file_info['Domain_description'] = info['description'] | |
| # break | |
| # # 尝试识别任务类型 | |
| # task_keywords = { | |
| # 'Fault_Diagnosis': ['fault', 'failure', 'diagnosis', 'diagnostic'], | |
| # 'Anomaly_Detection': ['anomaly', 'outlier', 'detection'], | |
| # 'Remaining_Life': ['remaining', 'life', 'rul', 'prognostics', 'prognosis'], | |
| # 'Digital_Twin_Prediction': ['digital', 'twin', 'simulation'] | |
| # } | |
| # for task, keywords in task_keywords.items(): | |
| # if any(kw in filename_lower or kw in path_lower or kw in dataset_name.lower() for kw in keywords): | |
| # file_info[task] = 'Yes' | |
| # # 如果是剩余寿命任务，尝试填充RUL标签信息 | |
| # if task == 'Remaining_Life' and file_info['Remaining_Life'] == 'Yes': | |
| # file_info['RUL_label'] = 'RUL' | |
| # file_info['RUL_label_description'] = '设备剩余使用寿命(天/小时/循环)' | |
| datasets.append(file_info) | |
| return datasets | |
| def save_to_csv(datasets, output_file): | |
| """ | |
| 将数据集信息保存到CSV文件 | |
| Args: | |
| datasets (list): 数据集信息列表 | |
| output_file (str): 输出文件路径 | |
| """ | |
| if not datasets: | |
| print("没有找到数据集") | |
| return | |
| columns = [ | |
| 'Id', 'Dataset_id', 'Name', 'Description', 'TYPE', 'File', 'Visiable', | |
| 'Label', 'Label_Description', 'RUL_label', 'RUL_label_description', 'Fault level','Domain_id', 'Domain_description', | |
| 'Sample_rate', 'Sample_lenth', 'Channel', 'Fault_Diagnosis', | |
| 'Anomaly_Detection', 'Remaining_Life', 'Digital_Twin_Prediction', | |
| ] | |
| # # 处理可能包含非ASCII字符的文件路径 | |
| # for item in datasets: | |
| # if 'File' in item and item['File']: | |
| # try: | |
| # # 确保文件路径是有效的 UTF-8 字符串 | |
| # item['File'] = item['File'].encode('utf-8', errors='ignore').decode('utf-8') | |
| # except Exception as e: | |
| # print(f"处理文件路径时出错: {e}, 路径: {item['File']}") | |
| # item['File'] = repr(item['File'])[1:-1] # 使用字符串表示作为后备 | |
| df = pd.DataFrame(datasets, columns=columns) | |
| df.to_csv(output_file, index=False, encoding='utf-8-sig') | |
| print(f"找到 {len(datasets)} 个数据集文件") # , encoding='utf-8-sig' | |
| print(f"数据已保存到 {output_file}") | |
| def main(): | |
| # 指定根目录和输出文件 | |
| root_dir = '/home/user/data/PHMbenchdata/raw' | |
| output_file = os.path.join(root_dir, 'dataset_info2.csv') | |
| print(f"开始遍历目录: {root_dir}") | |
| # 遍历目录收集信息 | |
| datasets = explore_datasets(root_dir) | |
| # 保存到CSV | |
| save_to_csv(datasets, output_file) | |
| if __name__ == "__main__": | |
| main() | |