如何使用Python自动化Google PageSpeed Insights测试

Question

有没有一种方法可以自动检查Google Page Speed分数？

Answer 1

因此，我想出了如何使用文档中包含的Google Page Speed API来执行此操作。

TL：DR的解释是，您可以使用以下URL设置，在Cloud Console中启用Google Pagespeed API后替换括号中的值（如果在浏览器中，您还必须已登录Authenticated Google Cloud User）。

https://www.googleapis.com/pagespeedonline/v2/runPagespeed?url={YOUR_SITE_URL}/&filter_third_party_resources=true&locale=en_US&screenshot=false&strategy=desktop&key={YOUR_API_KEY}

从上面的链接可以看到，您将需要一个Google Pagespeed API密钥。这些是从头开始的安装说明。如果您的项目已经在Cloud Console上，则可以跳过前几个步骤。

1. 转到云端https://console.developers.google.com
2. 如有必要，注册一个帐户。
3. 建立专案
4. 转到菜单> API管理器>凭据
5. 创建凭证按钮> API密钥
6. 复制键，然后单击关闭
7. 菜单>仪表板>启用API
8. 使用搜索找到PageSpeed见解API，然后单击它
9. 使用标题附近的“ |> ENABLE”按钮

拥有API密钥后，您可以替换URL中的值，它看起来应该像这样：

https://www.googleapis.com/pagespeedonline/v2/runPagespeed?url=https://www.example.com/&filter_third_party_resources=true&locale=en_US&screenshot=false&strategy=desktop&key=FaKeApIKey29nS8U22mM

网址中的参数strategy = desktop可以更改为strategy = mobile。对于移动设备，您可以获得速度和可用性得分。JSON的开头如下所示：

{
"kind": "pagespeedonline#result",
"id": "https://www.example.com/fake/page”,
"responseCode": 200,
"title": "Example Domain",
"ruleGroups": {
    "SPEED": {
       "score": 100
    },
   "USABILITY": {
      "score": 100
    }
   },
  ....continued...

因此，我使用Python和Python单元测试将其自动化。

import requests
import json
import unittest
from globes import *

api_key = '' # Add API key. Found here: https://console.developers.google.com/apis/credentials/key/
base = 'http://example.com'
locale_code = 'en_US'

def get_insights_json(self, page_url, local, device_type, api_key, speed_or_useability, expected_score):
    url = 'https://www.googleapis.com/pagespeedonline/v2/runPagespeed?url=' + page_url + '&filter_third_party_resources=true&locale=' + local + '&screenshot=false&strategy=' + device_type + '&key=' + api_key
    # print "Getting :: " + url
    r = requests.get(url)
    return_code = r.status_code
    try: self.assertEqual(return_code, 200)
    except AssertionError, e: self.verificationErrors.append(str(page_url) + " did not return 200")
    return_text = r.text
    return_json = json.loads(return_text)
    score = return_json['ruleGroups'][speed_or_useability]['score']
    print 'Getting ' + speed_or_useability + ' for '  + page_url + ' and got a score of ' + str(score)
    try: self.assertTrue(int(score) >= expected_score)
    except AssertionError, e: self.verificationErrors.append(str(page_url) + ' expected ' + device_type + ' speed score to be greater than ' + str(expected_score) + ', instead got ' + str(score) )

class TestAllAPIs(unittest.TestCase):

    def setUp(self):
        self.verificationErrors = []
        self.maxDiff = None

    def tearDown(self):
        self.assertEqual([], self.verificationErrors)

    def test_desktop_speed(self):
        current_page =  base + '' # You could add to the url to test other pages, I tend to do this is a loop using a list I set up by base url.
        device_type = 'desktop'
        target = 'SPEED'
        get_insights_json(self, current_page, locale_code, device_type, api_key, target, 80)

    def test_mobile_speed(self):
        current_page =  base + ''
        device_type = 'mobile'
        target = 'SPEED'
        get_insights_json(self, current_page, locale_code, device_type, api_key, target, 80)

    def test_mobile_useability(self):
        current_page =  base + ''
        device_type = 'mobile'
        target = 'USABILITY'
        get_insights_json(self, current_page, locale_code, device_type, api_key, target, 80)

if __name__ == "__main__":
    unittest.main()

对我来说，有个谜是为什么浏览器需要通过Google进行身份验证才能从URL获取JSON，而Python请求却不需要。