| 方案 | 優點 | 缺點 | 適用場景 |
|---|---|---|---|
| webbrowser.open() | 跨平台、零依賴 | 無法控制現有窗口、無事件 | 簡單打開 URL |
| Selenium WebDriver | 標準化、多瀏覽器、強大選擇器 | 需驅動程序、資源佔用大 | 測試自動化、爬蟲 |
| pywin32 COM (本文) | 原生控制 IE、事件模型、無額外依賴 | 僅限 Windows/IE、IE 已停止更新 | legacy 系統集成、企業內部系統 |
| Playwright/Puppeteer | 現代、無頭、高性能 | 需 Node.js/Python 綁定 | 現代 Web 自動化首選 |
import win32com.client
import pythoncom
import time
class IEAutomation:
def __init__(self, visible=True):
# DispatchWithEvents 支援事件處理
self.ie = win32com.client.DispatchWithEvents(
'InternetExplorer.Application',
self._EventHandler
)
self.ie.Visible = visible
self._nav_complete = False
self._events = []
class _EventHandler:
def OnDocumentComplete(self, pDisp, URL):
print(f"[DocumentComplete] {URL}")
# 只有頂層文檔完成才標記完成
if pDisp == self.ie.Document:
self._nav_complete = True
def OnDownloadBegin(self):
print("[DownloadBegin]")
def OnDownloadComplete(self):
print("[DownloadComplete]")
def OnNavigateError(self, pDisp, URL, Frame, StatusCode, Cancel):
print(f"[NavigateError] {URL} - {StatusCode}")
def navigate(self, url, timeout=30):
self._nav_complete = False
self.ie.Navigate(url)
# 消息泵處理 COM 事件
start = time.time()
while not self._nav_complete:
if time.time() - start > timeout:
raise TimeoutError(f"Navigation timeout: {url}")
pythoncom.PumpWaitingMessages()
time.sleep(0.1)
def get_html(self):
return self.ie.Document.documentElement.outerHTML
def execute_script(self, script):
return self.ie.Document.parentWindow.execScript(script)
def click_element(self, element_id):
elem = self.ie.Document.getElementById(element_id)
if elem:
elem.click()
def fill_form(self, field_name, value):
elem = self.ie.Document.getElementsByName(field_name)
if elem.Length > 0:
elem.Item(0).value = value
def quit(self):
self.ie.Quit()
# 使用示例
if __name__ == '__main__':
ie = IEAutomation(visible=True)
try:
ie.navigate("https://www.jb51.net")
html = ie.get_html()
print(f"Page length: {len(html)}")
# ie.fill_form('q', 'python')
# ie.click_element('su')
finally:
ie.quit()
DispatchWithEvents 創建事件接收器,映射 IE 的 DWebBrowserEvents2 接口:
OnDocumentComplete:文檔加載完成 (含 iframe,需判斷 pDisp == Document)OnDownloadBegin/Complete:下載狀態OnNavigateError:導航錯誤OnQuit:瀏覽器關閉pythoncom.PumpWaitingMessages() 處理 COM 事件隊列,必須在循環中調用,否則事件不會觸發。
通過 ie.Document 訪問 MSHTML DOM,API 與 JavaScript 類似但區分大小寫:
doc = ie.Document
doc.getElementById('id')
doc.getElementsByName('name')[0]
doc.getElementsByTagName('input')
doc.forms[0].submit()
from selenium import webdriver
from selenium.webdriver.common.by import By
driver = webdriver.Edge() # 或 Chrome, Firefox
driver.get("https://www.jb51.net")
print(driver.page_source)
driver.quit()
# pip install playwright && playwright install
from playwright.sync_api import sync_playwright
with sync_playwright() as p:
browser = p.chromium.launch(headless=False)
page = browser.new_page()
page.goto("https://www.jb51.net")
print(page.content())
browser.close()
雖然 IE 已退役,但大量企業內部系統仍依賴 IE/ActiveX,掌握 COM 自動化技能對維護 legacy 系統仍有價值。