西储大学轴承故障数据 python3使用

转载于：参考链接
西储大学轴承故障数据官网
python2 版本 cwru库github地址
Github数据下载地址
python2 可以直接安装使用CWRU库，该库的功能是下载数据，并且切分成可供训练和评估的训练集和测试集数据

python2安装方式：

pip安装：

$ pip install --user cwru

github下载源代码安装:
`$ python setup.py install
使用

import cwru
data = cwru.CWRU("12DriveEndFault", "1797", 384)

可以使用data.X_train, data.y_train, data.X_test, data.y_test, data.labels, data.nclasses 来训练和评估模型

CWRU的参数：

exp：‘12DriveEndFault’, ‘12FanEndFault’, ‘48DriveEndFault’

rpm：‘1797’, ‘1772’, ‘1750’, ‘1730’

length：信号的长度

python3版本：

由于python3 和python2 版本的差异，对原python2代码修改：
这产生的路径名和原文不同 生成到项目目录下

import os
import glob
import errno
import random
import urllib.request as urllib
import numpy as np
from scipy.io import loadmat


class CWRU:
    def __init__(self, exp, rpm, length):
        if exp not in ('12DriveEndFault', '12FanEndFault', '48DriveEndFault'):
            print("wrong experiment name: {}".format(exp))
            exit(1)
        if rpm not in ('1797', '1772', '1750', '1730'):
            print("wrong rpm value: {}".format(rpm))
            exit(1)
        # root directory of all data
        rdir = os.path.join('Data/CWRU')
        print(rdir)
        fmeta = os.path.join(os.path.dirname(__file__), 'metadata.txt')
        all_lines = open(fmeta).readlines()
        lines = []
        for line in all_lines:
            l = line.split()
            if (l[0] == exp or l[0] == 'NormalBaseline') and l[1] == rpm:
                lines.append(l)
        self.length = length  # sequence length
        self._load_and_slice_data(rdir, lines)
        # shuffle training and test arrays
        self._shuffle()
        self.labels = tuple(line[2] for line in lines)
        self.nclasses = len(self.labels)  # number of classes

    def _mkdir(self, path):
        try:
            os.makedirs(path)
        except OSError as exc:
            if exc.errno == errno.EEXIST and os.path.isdir(path):
                pass
            else:
                print("can't create directory '{}''".format(path))
                exit(1)

    def _download(self, fpath, link):
        print("Downloading to: '{}'".format(fpath))
        urllib.URLopener().retrieve(link, fpath)

    def _load_and_slice_data(self, rdir, infos):
        self.X_train = np.zeros((0, self.length))
        self.X_test = np.zeros((0, self.length))
        self.y_train = []
        self.y_test = []
        for idx, info in enumerate(infos):

            # directory of this file
            fdir = os.path.join(rdir, info[0], info[1])
            self._mkdir(fdir)
            fpath = os.path.join(fdir, info[2] + '.mat')
            if not os.path.exists(fpath):
                self._download(fpath, info[3].rstrip('\n'))

            mat_dict = loadmat(fpath)
            # key = filter(lambda x: 'DE_time' in x, mat_dict.keys())[0]
            fliter_i = filter(lambda x: 'DE_time' in x, mat_dict.keys())
            fliter_list = [item for item in fliter_i]
            key = fliter_list[0]
            time_series = mat_dict[key][:, 0]
            idx_last = -(time_series.shape[0] % self.length)
            clips = time_series[:idx_last].reshape(-1, self.length)
            n = clips.shape[0]
            n_split = int((3 * n / 4))
            self.X_train = np.vstack((self.X_train, clips[:n_split]))
            self.X_test = np.vstack((self.X_test, clips[n_split:]))
            self.y_train += [idx] * n_split
            self.y_test += [idx] * (clips.shape[0] - n_split)

    def _shuffle(self):
        # shuffle training samples
        index = list(range(self.X_train.shape[0]))
        random.Random(0).shuffle(index)
        self.X_train = self.X_train[index]
        self.y_train = tuple(self.y_train[i] for i in index)

        # shuffle test samples
        index = list(range(self.X_test.shape[0]))
        random.Random(0).shuffle(index)
        self.X_test = self.X_test[index]
        self.y_test = tuple(self.y_test[i] for i in index)

由于下载出现302错误，所以自己打包下载了相关链接中的所有数据并且进行了数据转换

#include 
#include 
#include 
#include 
#include 
using namespace std;
char a[100],b[100],c[100],d[100];
void getAllFiles(string path, vector<string>& files)
{
    // 文件句柄
    long hFile = 0;
    // 文件信息
    struct _finddata_t fileinfo;

    string p;

    if ((hFile = _findfirst(p.assign(path).append("\\*").c_str(), &fileinfo)) != -1)
    {
        do
        {
            // 保存文件的全路径
            files.push_back(p.assign(path).append("\\").append(fileinfo.name));

        }
        while (_findnext(hFile, &fileinfo) == 0);   //寻找下一个，成功返回0，否则-1

        _findclose(hFile);
    }
}
void make_dir(string dir)
{
    if (_access(dir.c_str(), 0) == -1)
    {
        cout << dir << " is not existing" << endl;
        int flag = _mkdir(dir.c_str());

        if (flag == 0)
        {
            cout << "make successfully" << endl;
        }
        else
        {
            cout << "make fsiled" << endl;
        }
    }
    else if (_access(dir.c_str(), 0) == 0)
    {
        cout << dir << " exists" << endl;
    }
    else
    {
        cout<<"erro!"<<endl;
    }
}

int CopyFile(string SourceFile,string NewFile)
{
    ifstream in;
    ofstream out;
    in.open(SourceFile,ios::binary);//打开源文件
    if(in.fail())//打开源文件失败
    {
        cout<<"Error 1: Fail to open the source file."<<endl;
        in.close();
        out.close();
        return 0;
    }
    out.open(NewFile,ios::binary);//创建目标文件
    if(out.fail())//创建文件失败
    {
        cout<<"Error 2: Fail to create the new file."<<endl;
        out.close();
        in.close();
        return 0;
    }
    else//复制文件
    {
        out<<in.rdbuf();
        out.close();
        in.close();
        return 1;
    }
}

int main()
{
    freopen("metadata.txt","r",stdin);
    int count1 = 0,count2 =0;
    while(scanf("%s %s %s %s",&a,&b,&c,&d)!=EOF)
    {
        count2++;
        int len=strlen(d),x=0;
        for(int i=0; i<len; i++)
            if(d[i]<='9' && d[i] >='0')
                x=x*10+d[i]-'0';
        string A=string(a),B=string(b),C=string(c);
        cout<<A<<" "<<B<<" "<<C<<" "<<x<<endl;
        string filePath1 = "D:\\ACM_CODE\\ACMtask\\data";
        string filePath2 = "D:\\ACM_CODE\\ACMtask\\datas";
        vector<string> temp;
        int y=0;
        getAllFiles(filePath1, temp);
        for (int i = 0; i < temp.size(); i++ )
        {
            y=0;
            //  cout<
            for(int j=0; j<temp[i].length(); j++)
            {
                if(temp[i][j] >= '0' && temp[i][j] <= '9')
                    y=y*10+temp[i][j]-'0';
                else if(temp[i][j] =='.' && temp[i][j+1] =='m' &&
                        temp[i][j+2] =='a' && temp[i][j+3] =='t')
                {
                    break;
                }
                else
                    y=0;
            }
            if(x == y)
            {
                y=i;
                break;
            }
        }
        cout<<y<<"  "<<temp[y]<<endl;
        string dir=filePath2+"\\"+A;
        make_dir(dir);
        dir=dir+"\\"+B;
        make_dir(dir);
        string source=temp[y];
        string NewFile=dir+"\\"+C+".mat";
        if(CopyFile(source,NewFile))
        {
            cout<<"文件已成功复制..."<<endl;
            count1++;
        }
        else
        {
            cout<<"文件复制失败..."<<endl;

        }
    }
    printf("YES %d %d\n",count1,count2);
    return 0;
}

这里提供下载分类好的数据集合
下载链接