[Selenium] 自动侦测浏览器版本并下载对应的浏览器驱动

昨天在群里聊天时，有同学说 Appium 官方支持自动下载兼容的浏览器驱动，想来Selenium也有类似的方法，于是在网上搜索一番。参考了Medium上一篇文章的方法，对步骤进行改进，增加了对多浏览器的支持。

首先，先想好大致上的几个步骤

1. 识别本地浏览器版本
2. 下载对应浏览器版本的驱动
3. 解压到对应文件夹
4. 记录到mapping.json文件中

接下来就是撸起袖子开干

定义好目录结构

|— config

​ |— mapping.json: 浏览器驱动配置信息

|— driver: 存放浏览器驱动

|— utils

​ |— driver_util.py: 封装的工具包

|— test_search.py: 测试脚本

数据准备

导入第三方库，定义好路径名称等常量

import json
import os
import zipfile
import shutil
import requests
import pathlib
from win32com import client as win_client

# 工作目录（当前路径调试时需加上.parent）
BASE_DIR = str(pathlib.Path.cwd())
# BASE_DIR = str(pathlib.Path.cwd().parent)

CHROME_DRIVER_BASE_URL = "https://chromedriver.storage.googleapis.com"
EDGE_DRIVER_BASE_URL = "https://msedgedriver.azureedge.net"
CHROME_DRIVER_ZIP = "chromedriver_win32.zip"
EDGE_DRIVER_ZIP = "edgedriver_win64.zip"
CHROME_DRIVER = "chromedriver.exe"
EDGE_DRIVER = "msedgedriver.exe"

BROWSER_DRIVER_DIR = str(pathlib.PurePath(BASE_DIR, "driver"))
DRIVER_MAPPING_FILE = os.path.join(BASE_DIR, "config", "mapping.json")

第一步，获取浏览器的版本

Chrome 浏览器有些小版本没有对应版本号的浏览器驱动，需要借助 Query API 查询对应大版本LATEST RELEASE版本，再根据查询对应的浏览器驱动

新版Edge 浏览器每个版本号官网都有对应的驱动下载

Latest Version API
https://chromedriver.storage.googleapis.com/LATEST_RELEASE_{version}Download Chrome Driver API
https://chromedriver.storage.googleapis.com/{version}/chromedriver_win32.zip
https://msedgedriver.azureedge.net/{version}/edgedriver_win64.zip

代码如下

def get_browser_version(file_path):
    """
    获取浏览器版本
    :param file_path: 浏览器文件路径
    :return: 浏览器大版本号
    """
    # 判断路径文件是否存在
    if not os.path.isfile(file_path):
        raise FileNotFoundError(f"{file_path} is not found.")
    win_obj = win_client.Dispatch('Scripting.FileSystemObject')
    version = win_obj.GetFileVersion(file_path)

    return version.strip()

def get_browser_major_version(file_path):
    """
    获取浏览器大版本号
    :param file_path: 浏览器文件路径
    :return: 浏览器大版本号
    """
    browser_ver = get_browser_version(file_path)
    browser_major_ver = browser_ver.split(".")[0]

    return browser_major_ver

def get_latest_browser_version(browser_major_ver):
    """
    获取匹配大版本的最新release版本
    :param browser_major_ver: 浏览器大版本号
    :return: 最新release版本号
    """
    latest_api = f"{CHROME_DRIVER_BASE_URL}/LATEST_RELEASE_{browser_major_ver}"
    resp = requests.get(latest_api)
    latest_driver_version = resp.text.strip()

    return latest_driver_version

第二步，下载浏览器驱动

def download_browser_driver(latest_driver_version, browser_name):
    """
    下载浏览器驱动压缩包
    :param browser_name: 浏览器名称
    :param latest_driver_version: 浏览器的版本号
    """
    download_api = None
    if browser_name == "Chrome":
        download_api = f"{CHROME_DRIVER_BASE_URL}/{latest_driver_version}/{CHROME_DRIVER_ZIP}"
    elif browser_name == "Edge":
        download_api = f"{EDGE_DRIVER_BASE_URL}/{latest_driver_version}/{EDGE_DRIVER_ZIP}"

    download_dir = os.path.join(str(BROWSER_DRIVER_DIR), os.path.basename(download_api))
    # 下载，设置超时时间20s
    resp = requests.get(download_api, stream=True, timeout=20)

    if resp.status_code == 200:
        with open(download_dir, 'wb') as fo:
            fo.write(resp.content)
    else:
        raise Exception("Download chrome driver failed")

第三步，解驱动压缩包

解压后将原压缩包删除

def unzip_driver(browser_major_ver, browser_name):
    """
    解压驱动压缩包
    :param browser_name: 浏览器名称
    :param browser_major_ver: 浏览器大版本号
    :return: 驱动文件路径
    """
    file_path = None
    driver_path = None

    if browser_name == "Chrome":
        file_path = os.path.join(BROWSER_DRIVER_DIR, os.path.basename(CHROME_DRIVER_ZIP))
        driver_path = os.path.join(BROWSER_DRIVER_DIR, browser_major_ver, CHROME_DRIVER)
    elif browser_name == "Edge":
        file_path = os.path.join(BROWSER_DRIVER_DIR, os.path.basename(EDGE_DRIVER_ZIP))
        driver_path = os.path.join(BROWSER_DRIVER_DIR, browser_major_ver, EDGE_DRIVER)
    browser_driver_dir = os.path.join(BROWSER_DRIVER_DIR, browser_major_ver)

    # 解压到指定目录
    with zipfile.ZipFile(file_path, 'r') as zip_ref:
        zip_ref.extractall(browser_driver_dir)

    return driver_path

def remove_driver_zip(browser_name):
    """
    删除下载的驱动压缩包
    :param browser_name: 浏览器名称
    """
    file_path = None
    if browser_name == "Chrome":
        file_path = os.path.join(BROWSER_DRIVER_DIR, os.path.basename(CHROME_DRIVER_ZIP))
    elif browser_name == "Edge":
        file_path = os.path.join(BROWSER_DRIVER_DIR, os.path.basename(EDGE_DRIVER_ZIP))
    os.remove(file_path)

第四步，读写配置文件信息

def read_driver_mapping_json():
    """
    读取 mapping_json
    :return: 字典格式
    """
    if os.path.exists(DRIVER_MAPPING_FILE):
        with open(DRIVER_MAPPING_FILE) as fo:
            try:
                driver_mapping_dict = json.load(fo)
            # mapping.json内容为空时，返回空字典
            except json.decoder.JSONDecodeError:
                driver_mapping_dict = {}
    else:
        raise FileNotFoundError(f"{DRIVER_MAPPING_FILE} is not found")

    return driver_mapping_dict

def write_driver_mapping_json(browser_major_ver, latest_driver_version, driver_path, browser_name):
    """
    写入 mapping_json
    :param browser_major_ver: 浏览器大版本号
    :param latest_driver_version: 浏览器驱动版本号
    :param driver_path: 驱动存放路径
    :param browser_name: 浏览器名称
    """
    mapping_dict = read_driver_mapping_json()
    # 版本号在dict中（浏览器名不在dict中）
    if browser_major_ver in mapping_dict:

        mapping_dict[browser_major_ver][browser_name] = {
                            "driver_path": driver_path,
                            "driver_version": latest_driver_version
                }
    # 大版本号不在dict中，且字典不为空
    elif browser_major_ver not in mapping_dict and mapping_dict:
        mapping_dict[browser_major_ver] = {
            browser_name:
                {
                    "driver_path": driver_path,
                    "driver_version": latest_driver_version
                }
        }
    # 字典为空
    else:
        mapping_dict = {
            browser_major_ver:
                {
                    browser_name:
                        {
                            "driver_path": driver_path,
                            "driver_version": latest_driver_version
                        }
                }
        }
        mapping_dict.update(mapping_dict)

    with open(DRIVER_MAPPING_FILE, 'w') as fo:
        json.dump(mapping_dict, fo)

综合

将以上步骤整合到automatic_discover_driver函数中，通过调用该函数返回浏览器驱动路径

def automatic_discover_driver(browser_path, browser_name="Chrome"):
    """
    侦测浏览器驱动是否在mapping.json有记录，否则下载该驱动
    :param browser_path: 浏览器路径
    :param browser_name: 浏览器名称
    """
    browser_maj_ver = get_browser_major_version(browser_path)
    # Chrome需要获取大版本号对应的latest release version
    # Edge 可直接用当前浏览器版本号
    if browser_name == "Chrome":
        latest_browser_ver = get_latest_browser_version(browser_maj_ver)
    elif browser_name == "Edge":
        latest_browser_ver = get_browser_version(browser_path)
    else:
        raise Exception(f"{browser_name} is not found")

    # 读取mapping.json内容
    mapping_dict = read_driver_mapping_json()

    # json为空 或版本号不在mapping_dict中 或浏览器名不在mapping_dict中
    if not mapping_dict or \
            browser_maj_ver not in mapping_dict or \
            browser_name not in mapping_dict[browser_maj_ver]:

        # 下载浏览器驱动压缩包
        download_browser_driver(latest_browser_ver, browser_name)
        # 解压浏览器驱动压缩包，并返回驱动路径
        driver_path = unzip_driver(browser_maj_ver, browser_name)
        # 将浏览器大版本号、浏览器名、驱动路径、对应的浏览器版本号信息写入到mapping.json中
        write_driver_mapping_json(browser_maj_ver, latest_browser_ver, driver_path, browser_name)

        # 删除浏览器驱动压缩包
        remove_driver_zip(browser_name)

    # 返回浏览器驱动的路径
    mapping_dict = read_driver_mapping_json()
    return mapping_dict[browser_maj_ver][browser_name]["driver_path"]

测试

创建一个test_search.py文件验证是否可以自动下载对应的浏览器驱动

import pytest
from time import sleep
from selenium import webdriver
from utils.driver_util import automatic_discover_driver as automatic

class TestSearch:
    _CHROME_PATH = r"C:\Program Files (x86)\Google\Chrome\Application\chrome.exe"
    _EDGE_PATH = r"C:\Program Files (x86)\Microsoft\Edge\Application\msedge.exe"
    _browser = "Edge"

    def setup(self):
        driver_path = automatic(self._EDGE_PATH, self._browser)
        if self._browser == "Chrome":
            self.driver = webdriver.Chrome(driver_path)
        elif self._browser == "Edge":
            self.driver = webdriver.Edge(driver_path)

    def teardown(self):
        self.driver.close()
        self.driver.quit()

    def test_search_bing(self):
        self.driver.get("https://cn.bing.com/")
        self.driver.find_element_by_id("sb_form_q").send_keys("selenium")
        self.driver.find_element_by_id("sb_go_par").click()
        sleep(3)

if __name__ == '__main__':
    pytest.main()

实测，成功打开浏览器！

详细代码：https://github.com/felixzfq/AutomaticDiscoverBrowserDriver

参考资料

• Automatic discovery of compatible Chromedriver
• [Python][Selenium] 人生苦短，把麻煩的 Chrome Browser Driver Version Mapping 自動化