三、python中最基础的文件处理汇总

最新推荐文章于 2022-09-19 19:14:41 发布

原创最新推荐文章于 2022-09-19 19:14:41 发布 · 418 阅读

0 ·

CC 4.0 BY-SA版权

深度学习同时被 3 个专栏收录

14 篇文章

订阅专栏

Python

9 篇文章

订阅专栏

加载数据与文件处理

6 篇文章

订阅专栏

本文介绍了如何将文件夹内的图片和txt文件按比例划分为train和val，复制到多个文件夹，并将多个标签合并到一个txt文件，进行图片压缩，以及将内容整理到一个Excel文件的操作步骤。

文章目录

前言
一、分配train和val的训练数据
二、将一个文件夹的文件按比例复制给多个文件夹
三、多个txt文件的标签内容合并在一个txt文件
四、图片压缩（有损压缩和无损压缩）
五、多个txt文件、图片内容转到1个excel文件

前言

最基本的文件操作。

提示：以下是本篇文章正文内容，下面案例可供参考

一、分配train和val的训练数据

一个文件夹的文件按比例分为train和val数据。


'''train:val=1.5:8.5'''

import os ,random,shutil #文件操作模块
import glob

def moveFile(all_img_dir,val_img_dir,all_txt_dir,val_txt_dir):

    pathdir=os.listdir(all_img_dir)
    filenumber=len(pathdir)
    ratio=0.15      #分配比例
    picknumber=int(filenumber*ratio)  #抽取的文件数量 int类型
    sample=random.sample(pathdir,picknumber)  #随机抽取

    for name in sample:
        txtName=name[:-3]+"txt"
        shutil.move(os.path.join(all_img_dir,name),os.path.join(val_img_dir,name))   #从A移到B
        shutil.move(os.path.join(all_txt_dir, txtName), os.path.join(val_txt_dir, txtName))  # 从A移到B

    return

if __name__ == '__main__':

    all_img_dir=r"./data/text_recog/zyx_data/submit_data/recog/low_score_v3p1/*.png"
    val_img_dir=r"./data/text_recog/txt4val/img_gt1"
    all_txt_dir=r"./data/text_recog/zyx_data/submit_data/recog/low_score_v3p1/*.txt"
    val_txt_dir=r"./data/text_recog/txt4val/label1"
    moveFile(all_img_dir,val_img_dir,all_txt_dir,val_txt_dir)

二、将一个文件夹的文件按比例复制给多个文件夹

#!/usr/bin/env python
# -*- coding:utf-8 -*-

import os
import random
import glob
import shutil
from shutil import copy2
trainpath = r"/data2/enducation/paper_recog_total/train-paper-recog/Recognization/deep-text-recognition-SHENG/data/text_recog/zyx_data/submit_data/recog/low_score_v3p1" #??????????????????txt?????
trainfiles_img=glob.glob(os.path.join(trainpath,"*.png"))

num_train = len(trainfiles_img)
print( "num_train: " + str(num_train) )
index_list = list(range(num_train))
print(index_list)
random.shuffle(index_list)
num = 0
trainDir_img = r'/data/text_recog/txt4train/img_gt2'#?????????7??????????
validDir_img = r'/data/text_recog/txt4val/img_gt2'#?????????3??????????

trainDir_txt=r'/data/text_recog/txt4train/label2'
validDir_txt=r'/data/text_recog/txt4val/label2'
for i in index_list:
    filepPath=trainfiles_img[i]
    files,fileName=os.path.split(filepPath)
    txtname=fileName[:-3]+"txt"
    txtpath=os.path.join(trainpath,txtname)


    if num < num_train*0.85:
        copy2(filepPath, trainDir_img)
        copy2(txtpath,trainDir_txt)
        print(f"train: {str(fileName)} -------> {str(txtname)}")

    else:
        copy2(filepPath, validDir_img)
        copy2(txtpath, validDir_txt)
        print(f"val: {str(fileName)} -------> {str(txtname)}")
    num += 1
    print(f"f{num} move success!")

分给10个文件夹`。




#!/usr/bin/env python
# -*- coding:utf-8 -*-

import os
import random
import glob
import shutil
from shutil import copy2
trainpath = r"/data2/datas/ocr/text_recog/ocr_data_2021.6.30/train"
trainfiles_img=glob.glob(os.path.join(trainpath,"*.png"))

num_train = len(trainfiles_img)
print( "num_train: " + str(num_train) )
index_list = list(range(num_train))
# print(index_list)
random.shuffle(index_list)
num = 0
trainDir_img = r'/data2/datas/ocr/text_recog/ocr_data_2021.6.30/all_train'

for i in index_list:
    filepPath=trainfiles_img[i]
    files,fileName=os.path.split(filepPath)
    txtname=fileName[:-3]+"txt"
    txtpath=os.path.join(trainpath,txtname)
    # print(os.path.join(trainDir_img,"0"))

    if num < num_train*0.10:
        copy2(filepPath, os.path.join(trainDir_img,"0"))
        copy2(txtpath,os.path.join(trainDir_img,"0"))
        print(f"0: {str(fileName)} -------> {str(txtname)}")
    elif num_train*0.10<=num < num_train*0.20:
        copy2(filepPath, os.path.join(trainDir_img, "1"))
        copy2(txtpath, os.path.join(trainDir_img, "1"))
        print(f"1: {str(fileName)} -------> {str(txtname)}")
    elif num_train*0.20<=num < num_train*0.30:
        copy2(filepPath, os.path.join(trainDir_img, "2"))
        copy2(txtpath, os.path.join(trainDir_img, "2"))
        print(f"2: {str(fileName)} -------> {str(txtname)}")
    elif num_train*0.30<=num < num_train*0.40:
        copy2(filepPath, os.path.join(trainDir_img, "3"))
        copy2(txtpath, os.path.join(trainDir_img, "3"))
        print(f"3: {str(fileName)} -------> {str(txtname)}")
    elif num_train*0.40<=num < num_train*0.50:
        copy2(filepPath, os.path.join(trainDir_img, "4"))
        copy2(txtpath, os.path.join(trainDir_img, "4"))
        print(f"4: {str(fileName)} -------> {str(txtname)}")
    elif num_train*0.50<=num < num_train*0.60:
        copy2(filepPath, os.path.join(trainDir_img, "5"))
        copy2(txtpath, os.path.join(trainDir_img, "5"))
        print(f"5: {str(fileName)} -------> {str(txtname)}")
    elif num_train*0.60<=num < num_train*0.70:
        copy2(filepPath, os.path.join(trainDir_img, "6"))
        copy2(txtpath, os.path.join(trainDir_img, "6"))
        print(f"6: {str(fileName)} -------> {str(txtname)}")
    elif num_train*0.70<=num < num_train*0.80:
        copy2(filepPath, os.path.join(trainDir_img, "7"))
        copy2(txtpath, os.path.join(trainDir_img, "7"))
        print(f"7: {str(fileName)} -------> {str(txtname)}")
    elif num_train*0.80<=num < num_train*0.90:
        copy2(filepPath, os.path.join(trainDir_img, "8"))
        copy2(txtpath, os.path.join(trainDir_img, "8"))
        print(f"8: {str(fileName)} -------> {str(txtname)}")
    else:
        copy2(filepPath, os.path.join(trainDir_img, "9"))
        copy2(txtpath, os.path.join(trainDir_img, "9"))
        print(f"9: {str(fileName)} -------> {str(txtname)}")
    num += 1
    print(f"f{num} move success!")

三、多个txt文件的标签内容合并在一个txt文件

import os

def Combine_txt():
    dirpath_label = r"./data/text_recog/txt4train/label"
    dirpath_img = r"/data/text_recog/txt4train/img_gt"
    files = os.listdir(dirpath_label)
    res = ""
    for file in files:
        if file.endswith(".txt"):
            name = file[:-3] + "png"
            img_path = os.path.join(dirpath_img, name)
            txtpath = os.path.join(dirpath_label, file)
            with open(txtpath, "r", encoding="utf-8") as file:
                content = file.read()
                text = img_path + "\t" + content
            file.close()

        input = "%s\n" % (text)
        res += input

    outpath = r"./data/text_recog/txt4train/img_gt/gt.txt"
    with open(outpath, "w", encoding="utf-8") as outfile:
        outfile.write(res)
        outfile.close()
    print(len(res))

if __name__ == '__main__':
    Combine_txt()

四、图片压缩（有损压缩和无损压缩）

import cv2
import os

image_path=r'.\data2'
# image_savepath=r'./compress_img'
#取值范围：0~9，数值越小，压缩比越低。
def Compress_img(proportion):
    i = 0
    for num in os.listdir(image_path):
        for img_name in os.listdir(os.path.join(image_path,num)):
            image_file=os.path.join(image_path,num,img_name)
            image=cv2.imread(image_file)
            # cv2.imwrite(f'./compress_jpg_img/{proportion}/{i}.jpg',image,params=[cv2.IMWRITE_JPEG_QUALITY, proportion])
            '''[cv2.IMWRITE_JPEG_QUAITY,50]可以实现图像的压缩。
                其取值为[0,100].0时图像可以得到极大地压缩，但是图像的品质会被大大降低。'''

            cv2.imwrite(f'./compress_jpg_img/{proportion}/{i}.jpg', image,params=[cv2.IMWRITE_PNG_COMPRESSION, proportion])
            '''[cv2.IMWRITE_PNG_COMPRESSION,0]是调整PNG图像的压缩比。
                为0时，压缩比最小，图像的品质最好。它的压缩范围为[0，9]'''
            print(f'{proportion}---{num}---{img_name}Saved successfully!')
            i+=1

if __name__ == '__main__':
    #压缩范围4-9  PNG  取值范围：0~9，数值越小，压缩比越低。
    #压缩范围40-100  JPG 取值范围：0~100，数值越小，压缩比越高，图片质量损失越严重。
    list=[45,50,55,60,65,70,75,80,85,90]
    for proportions  in  list:
        if not os.path.exists(rf'./compress_jpg_img/{proportions}'):
            os.makedirs(rf'./compress_jpg_img/{proportions}')
        Compress_img(proportions)

五、多个txt文件、图片内容转到1个excel文件

转之前：
在这里插入图片描述

# !/usr/bin/python
# -*- codding: cp936 -*-

from openpyxl import load_workbook
from openpyxl.drawing.image import Image
import os

excel_address = r"E:\enducate\practice\movefile\txt2excel\train.xlsx"
img_path=r"E:\enducate\practice\movefile\val\img_gt"
txtfile = r'E:\enducate\practice\movefile\val\label'

wb = load_workbook(excel_address)
sht = wb.worksheets[0]

for i,img_name in enumerate( os.listdir(img_path)):
    num=img_name[:-3]
    txtname=num+"txt"

    sht.cell(i + 1, 1, num)

    with open(os.path.join(txtfile,txtname),"r", encoding='utf-8') as f:
        for line in f.readlines():
                # strip ??????????
                line = line.strip('\n')
                # ?','???'\t',????????????????????????
                line = line.replace("\t",",")
                line = line.split(',')
                # ??7???
                for index in range(len(line)):
                    sht.cell(i+1, 2, line[index])
                # ????



    img_address_1 = os.path.join(img_path,img_name)
    img = Image(img_address_1)
    #??????
    # img.width = 19.0
    # img.height = 39.0
    sht.add_image(img, f'C{i+1}')

    sht.column_dimensions['A'].width = 60.0
    sht.column_dimensions['B'].width = 60.0
    sht.column_dimensions['C'].width = 60.0
    sht.row_dimensions[i+1].height = 60.0


    wb.save(excel_address)

转之后：
在这里插入图片描述