多线程系列：自定义多线程类获取函数返回值

自定义多线程类参考博客: https://blog.csdn.net/xpt211314/article/details/109543014.

这里展示的多线程示例包括：

子线程包括多线程的情况
计算子线程耗时情况
多个线程对应一一对应不同的函数情况
多个线程对应同一个函数，只是输入参数不一样情况。

代码示例

python 复制代码

import threading
import time
from pprint import pprint


class MyThread(threading.Thread):
    def __init__(self, target, args=()):
        super(MyThread, self).__init__()
        self.func = target
        self.args = args
    def run(self):
        self.result = self.func(*self.args)
    def get_result(self):
        threading.Thread.join(self)  # 等待线程执行完毕
        try:
            return self.result
        except Exception as e:
            return e


# 定义分类模型函数
def classify_model_1(image_path):
    # 模拟分类操作
    time.sleep(1)
    print("classify_model_1")

    return {f"模型1_{image_path}分类结果": image_path}

def classify_model_2(image_path):
    # 模拟分类操作
    time.sleep(1)
    print("classify_model_2")

    return {f"模型2_{image_path}分类结果": image_path}

def classify_model_3(image_path):
    # 模拟分类操作
    time.sleep(1)
    print("classify_model_3")

    return {f"模型3_{image_path}分类结果": image_path}

def classify_model_4(image_path):
    # 模拟分类操作
    time.sleep(1)
    print("classify_model_4")

    return {f"模型4_{image_path}分类结果": image_path}

def classify_model_5(image_path):
    # 模拟分类操作
    time.sleep(1)
    print("classify_model_5")

    return {f"模型5_{image_path}分类结果": image_path}



# 定义分割模型函数
def segment_model_1(image_path):
    # 模拟分割操作
    time.sleep(1)
    print("segment_model_1")

    return {f"模型1_{image_path}分割结果": image_path}


def segment_model_2(image_path):
    # 模拟分割操作
    time.sleep(1)
    print("segment_model_2")

    return {f"模型2_{image_path}分割结果": image_path}

def segment_model_3(image_path):
    # 模拟分割操作
    time.sleep(1)
    print("segment_model_3")

    return {f"模型3_{image_path}分割结果": image_path}

def segment_model_4(image_path):
    # 模拟分割操作
    time.sleep(1)
    print("segment_model_4")

    return {f"模型4_{image_path}分割结果": image_path}

def segment_model_5(image_path):
    # 模拟分割操作
    time.sleep(1)
    print("segment_model_5")

    return {f"模型5_{image_path}分割结果": image_path}


def process_image(image_path):
    # 创建线程列表
    model_thread_list = [
        MyThread(classify_model, args=(image_path,)) 
        for classify_model in [
          classify_model_1, classify_model_2, classify_model_3, 
          classify_model_4, classify_model_5
        ]
    ]
    model_thread_list.extend([
        MyThread(segment_model, args=(image_path,))
        for segment_model in [
          segment_model_1, segment_model_2, segment_model_3,
          segment_model_4, segment_model_5
        ]
    ])
    start_time_dict = {}
    run_time_dict = {}
    # 启动线程
    for t in model_thread_list:
        t.start()
        start_time_dict[t.ident] = time.time()

    # 等待所有线程完成
    for t in model_thread_list:
        t.join()
        run_time_dict[t.ident] = time.time() - start_time_dict[t.ident]

    # 获取结果
    model_result_dict = {}
    for t in model_thread_list:
      model_result_dict.update(t.get_result()) 
    # print(results)
    model_result_dict.update(run_time_dict)

    return model_result_dict


if __name__ == '__main__':
    tic = time.time()
    image_path_list = [
      "image1.jpg",
      "image2.jpg",
    ]
    image_thread_list = [
      MyThread(process_image, args=(image_path,))
      for image_path in image_path_list 
    ]
    for t in image_thread_list: 
      t.start()
    result_dict = {}
    for t in image_thread_list:
        t.join()
        result_dict.update(t.get_result())

    pprint(result_dict)
    toc = time.time()
    print("程序运行时间：", toc - tic)

结果示例

python 复制代码

# segment_model_2
# segment_model_5
# segment_model_2
# segment_model_1
# segment_model_1
# segment_model_3
# segment_model_4
# classify_model_5
# segment_model_4
# segment_model_5
# segment_model_3
# classify_model_3
# classify_model_2
# classify_model_3
# classify_model_4
# classify_model_1
# classify_model_1
# classify_model_2
# classify_model_5
# classify_model_4
# {88328: 1.0205564498901367,
#  110924: 1.0205564498901367,
#  132612: 1.0205564498901367,
#  139284: 1.0205564498901367,
#  140332: 1.0205564498901367,
#  140392: 1.0205564498901367,
#  140404: 1.0205564498901367,
#  140756: 1.0205564498901367,
#  140920: 1.0205564498901367,
#  141028: 1.0205564498901367,
#  141076: 1.0205564498901367,
#  141252: 1.0205564498901367,
#  141416: 1.0205564498901367,
#  141628: 1.0205564498901367,
#  141964: 1.0205564498901367,
#  142716: 1.0205564498901367,
#  143008: 1.0205564498901367,
#  143148: 1.0205564498901367,
#  143312: 1.0205564498901367,
#  144176: 1.0205564498901367,
#  '模型1_image1.jpg分割结果': 'image1.jpg',
#  '模型1_image1.jpg分类结果': 'image1.jpg',
#  '模型1_image2.jpg分割结果': 'image2.jpg',
#  '模型1_image2.jpg分类结果': 'image2.jpg',
#  '模型2_image1.jpg分割结果': 'image1.jpg',
#  '模型2_image1.jpg分类结果': 'image1.jpg',
#  '模型2_image2.jpg分割结果': 'image2.jpg',
#  '模型2_image2.jpg分类结果': 'image2.jpg',
#  '模型3_image1.jpg分割结果': 'image1.jpg',
#  '模型3_image1.jpg分类结果': 'image1.jpg',
#  '模型3_image2.jpg分割结果': 'image2.jpg',
#  '模型3_image2.jpg分类结果': 'image2.jpg',
#  '模型4_image1.jpg分割结果': 'image1.jpg',
#  '模型4_image1.jpg分类结果': 'image1.jpg',
#  '模型4_image2.jpg分割结果': 'image2.jpg',
#  '模型4_image2.jpg分类结果': 'image2.jpg',
#  '模型5_image1.jpg分割结果': 'image1.jpg',
#  '模型5_image1.jpg分类结果': 'image1.jpg',
#  '模型5_image2.jpg分割结果': 'image2.jpg',
#  '模型5_image2.jpg分类结果': 'image2.jpg'}
# 程序运行时间： 1.0205564498901367