如何用多个图像文件进行压力测试烧瓶应用程序

发布于 2025-02-07 07:20:28 字数 981 浏览 1 评论 0原文

我有一个执行对象检测的烧瓶应用程序。我可以通过使用此方法通过卷曲来获得所需的输出：

curl -X POST -F images=@images/dog.jpeg http://localhost:8000/detections

我希望使用蝗虫对其进行压力测试。我当前的方法是这样的：

class UserBehavior(TaskSet):
    def _get_image_part(self, file_path, file_content_type='image/jpeg'):
        file_name = os.path.basename(file_path)
        file_content = open(file_path, 'rb')
        return file_name, file_content, file_content_type

    @task
    def detections(self):
        url = '/detections'
        
        files = {
            "images": self._get_image_part("images/dog.jpeg"),
        }
        response = self.client.post(url, files=files, verify=False)
        print(response.text)

class MyLocust(HttpUser):
    tasks = [UserBehavior]
    wait_time = between(0.1, 0.2)

如何为相同的测试图像加载列表（数组），以便在提出的每一个请求中，从目录中上传了一个随机选择的图像？

此外，当触发小负载（2-3个用＆lt; 5rps的用户）时，上述片段正常工作，但是在更高的请求或并发用户时，应用程序会引发错误（最常见的是FilenotFoundError）。是合理的吗？

提前致谢

原文

I have a flask app which performs object detection. I'm able to get the desired output via curl by using this:

curl -X POST -F images=@images/dog.jpeg http://localhost:8000/detections

I wish to stress test it using locust. My current approach is something like this:

class UserBehavior(TaskSet):
    def _get_image_part(self, file_path, file_content_type='image/jpeg'):
        file_name = os.path.basename(file_path)
        file_content = open(file_path, 'rb')
        return file_name, file_content, file_content_type

    @task
    def detections(self):
        url = '/detections'
        
        files = {
            "images": self._get_image_part("images/dog.jpeg"),
        }
        response = self.client.post(url, files=files, verify=False)
        print(response.text)

class MyLocust(HttpUser):
    tasks = [UserBehavior]
    wait_time = between(0.1, 0.2)

How do I load a list (array) of test images for the same, so that with every request made, a randomly selected image from a directory is uploaded?

Additionally, the above snippet works fine when a small load is triggered (2-3 users with <5rps) but on higher requests or concurrent users, the application throws error (the most common one being FileNotFoundError). Is it justified?

Thanks in advance

分享到QQ

分享到微博

如果你对这篇内容有疑问，欢迎到本站社区发帖提问参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。

发布评论

需要登录才能够评论，你可以免费注册一个本站的账号。

寄居者 2025-02-14 07:20:28

将图像加载到列表中（作为全局或类级变量），然后在任务中随机选择一个。

您要避免的是每次都从磁盘中重新加载图像，因为确实没有必要。

回复收藏 0 原文

佞臣 2025-02-14 07:20:28

基于 cyberwiz的答案，我通过在目录中创建图像列表，然后从此列表中随机选择图像来解决它每个请求。

这是我的实施：

def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self.image_list = os.listdir('images')

@task
def detections(self):
    url = '/detections'
    image_name = random.choice(self.image_list)
    
    files = {
        "images": self._get_image_part(os.path.join(os.getcwd(), 'images', image_name))
    }
    response = self.client.post(url, files=files, verify=False)
    print(response.text)

Based on Cyberwiz's answer, I solved it by creating a list of images in a directory and then randomly selecting an image from this list for every request.

Here's my implementation:

def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self.image_list = os.listdir('images')

@task
def detections(self):
    url = '/detections'
    image_name = random.choice(self.image_list)
    
    files = {
        "images": self._get_image_part(os.path.join(os.getcwd(), 'images', image_name))
    }
    response = self.client.post(url, files=files, verify=False)
    print(response.text)

回复收藏 0 原文

~没有更多了~