Update app.py

app.py

@@ -26,12 +26,13 @@ def get_root_domain(url):
     else:
         return domain
 
-def filter_type(_type: str):
+# 网络抓包内容过滤
+def filter_type(type: str):
     types = [
         'application/javascript', 'application/x-javascript', 'text/css', 'webp', 'image/png', 'image/gif',
         'image/jpeg', 'image/x-icon', 'application/octet-stream'
     ]
-    if _type not in types:
+    if type not in types:
         return True
     return False
     
@@ -41,10 +42,10 @@ def main():
     
 @app.get("/chrome")
 def chrome(url:str=None,wait:int=5,header:str=None,cookie:str=None):
-    
+    # 设置日志性能监听参数
     caps = {
         "browserName": "chrome",
-        'goog:loggingPrefs': {'performance': 'ALL'}  # 开启日志性能监听
+        'goog:loggingPrefs': {'performance': 'ALL'}  
     }
     
     # 必须有目标url
@@ -78,6 +79,8 @@ def chrome(url:str=None,wait:int=5,header:str=None,cookie:str=None):
 
     # 设置为无头模式
     options.add_argument('--headless')
+
+    # 开启日志性能监听
     for key, value in caps.items():
         options.set_capability(key, value)
 
@@ -123,22 +126,25 @@ def chrome(url:str=None,wait:int=5,header:str=None,cookie:str=None):
 
     # 完全加载完成时，页面是否有发生过 301 302 跳转过
     is_jump = (target_url != current_url)
-    
+
+    network = []
+    # 从日志性能监听中抓取网络包（Network），并进行过滤
     performance_log = driver.get_log('performance')  # 获取名称为 performance 的日志
     for packet in performance_log:
         message = json.loads(packet.get('message')).get('message')  # 获取message的数据
         if message.get('method') != 'Network.responseReceived':  # 如果method 不是 responseReceived 类型就不往下执行
             continue
         packet_type = message.get('params').get('response').get('mimeType')  # 获取该请求返回的type
-        if not filter_type(_type=packet_type):  # 过滤type
+        if not filter_type(packet_type):  # 过滤type
             continue
         requestId = message.get('params').get('requestId')  # 唯一的请求标识符。相当于该请求的身份证
         url = message.get('params').get('response').get('url')  # 获取 该请求  url
         try:
             resp = driver.execute_cdp_cmd('Network.getResponseBody', {'requestId': requestId})  # selenium调用 cdp
-            print(f'type: {packet_type} url: {url}')
+            request_headers = driver.execute_cdp_cmd('Network.getRequestHeaders', {'requestId': request_id})
+            network.append({"url":url, "type":packet_type})
             print(f'response: {resp}')
-            print()
+            print(f'request_headers: {request_headers}')
         except WebDriverException:  # 忽略异常
             pass
 
@@ -146,7 +152,8 @@ def chrome(url:str=None,wait:int=5,header:str=None,cookie:str=None):
         "url": current_url,
         "page_source": page_source,
         "cookies": cookies,
-        "is_jump": is_jump
+        "is_jump": is_jump,
+        "network": network,
     }
     
     driver.quit()