python开线程处理-爱代码爱编程

# -*- coding: utf-8 -*-

import os
import time
import numpy as np
import threading
from imutils import paths
import xml.etree.ElementTree as ET

def parse_xml(xml_file):
    tree = ET.parse(xml_file)
    root = tree.getroot()
    size = root.find('size')
    w = int(size.find('width').text)
    h = int(size.find('height').text)

    bbox = []
    # 这里需要注意下，迭代某个节点，需要用 root.iter('对应的节点')
    for obj in root.iter('object'):
        cls = obj.find('name').text
        xmlbox = obj.find('bndbox')
        b = (float(xmlbox.find('xmin').text), float(xmlbox.find('ymin').text), float(xmlbox.find('xmax').text),
             float(xmlbox.find('ymax').text))
        bbox.append([val for val in b])
    return bbox


def writeTxt(input_list, save_root, start=0, out_list=[]):
    for index, in_file in enumerate(input_list):
        ret_box = parse_xml(in_file)
        fname = os.path.split(in_file)[1].split('.')[0] + '.txt'
        txt_path = os.path.join(save_root, fname)
        with open(txt_path, 'w') as f:
            for box in ret_box:
                f.write(' '.join([str(val) for val in box]))
        out_list[start+index] = ret_box

def run():
    xml_root = r'D:\data\test-data\trainxml'
    save_root = r'D:\data\test-data\thread-txt'
    xml_all = list(paths.list_files(xml_root, '.xml'))
    xml_num = len(xml_all)
    out_ret = [0] * xml_num
    thread_num = 8
    each_num = np.ceil(xml_num / thread_num).astype(np.int)
    for i in range(thread_num):
        start_index = i * each_num
        end_idex = i * each_num + each_num
        if i == (thread_num-1):
            writeTxt(xml_all[start_index:end_idex], save_root, start_index, out_ret)
        else:
            writeTxt(xml_all[start_index:], save_root, start_index, out_ret)

def thread_run():
    xml_root = r'D:\data\test-data\trainxml'
    save_root = r'D:\data\test-data\thread-txt'
    xml_all = list(paths.list_files(xml_root, '.xml'))
    xml_num = len(xml_all)
    out_ret = [0] * xml_num
    thread_num = 8
    each_num = np.ceil(xml_num / thread_num).astype(np.int)
    thread_list =[]
    for i in range(thread_num):
        start_index = i * each_num
        end_idex = i * each_num + each_num
        if i == (thread_num-1):
            t = threading.Thread(target=writeTxt, args=(xml_all[start_index:end_idex], save_root, start_index, out_ret))
        else:
            t = threading.Thread(target=writeTxt, args=(xml_all[start_index:], save_root, start_index, out_ret))
        thread_list.append(t)
        t.start()

    for child_thread in thread_list:
        child_thread.join()
    print(out_ret)
    np_array = np.concatenate(out_ret, 0)

if __name__ == '__main__':
    start_time = time.time()
    run()
    end_time = time.time()
    print(f'thread time:{end_time - start_time}')

下图是不开线程和开线程结果对比，开线程会快到2倍

不开线程

开线程

参考：

https://www.liujiangblog.com/course/python/79