Working devtools response catch

This commit is contained in:
Анатолий Богомолов 2023-12-17 19:57:44 +10:00
parent 6783437582
commit c19c91bb81
1 changed files with 42 additions and 11 deletions

View File

@ -2,6 +2,8 @@ import time
import logging import logging
import uuid import uuid
import os import os
import json
import re
from selenium import webdriver from selenium import webdriver
from selenium.webdriver.chrome.options import Options from selenium.webdriver.chrome.options import Options
@ -28,6 +30,9 @@ class StolichkiDriver(webdriver.Chrome):
self, options: Options = None, service: Service = None, keep_alive: bool = True self, options: Options = None, service: Service = None, keep_alive: bool = True
) -> None: ) -> None:
self.last_resp_index = 0
self.last_resp_url = ""
# assert os.environ.get("TWOCAPTCA_KEY") is not None, "Can't fins environment variable TWOCAPTCHA_KEY" # assert os.environ.get("TWOCAPTCA_KEY") is not None, "Can't fins environment variable TWOCAPTCHA_KEY"
if options is None: if options is None:
@ -42,6 +47,8 @@ class StolichkiDriver(webdriver.Chrome):
options.add_experimental_option("useAutomationExtension", False) options.add_experimental_option("useAutomationExtension", False)
options.page_load_strategy = "eager" options.page_load_strategy = "eager"
options.capabilities["goog:loggingPrefs"] = {"performance": "ALL"}
# self.__solver = TwoCaptcha(os.environ.get("TWOCAPTCA_KEY")) # self.__solver = TwoCaptcha(os.environ.get("TWOCAPTCA_KEY"))
super().__init__(options, service, keep_alive) super().__init__(options, service, keep_alive)
@ -117,6 +124,30 @@ class StolichkiDriver(webdriver.Chrome):
self.add_cookie(new_cookie) self.add_cookie(new_cookie)
def get_network_response(self, url_mask):
logs = self.get_log("performance")[self.last_resp_index:]
url_mask = re.compile(url_mask)
for log in filter(lambda log: self.__filter_logs(log), logs):
message = json.loads(log.get("message"))['message']
request_id = message["params"]["requestId"]
resp_url = message["params"]["response"]["url"]
if re.fullmatch(url_mask, resp_url) and self.last_resp_url != resp_url:
self.last_resp_index = logs.index(log)
self.last_resp_url = resp_url
body = self.execute_cdp_cmd("Network.getResponseBody", {"requestId": request_id})
return json.loads(body['body'])
return None
def __filter_logs(self, log):
message = json.loads(log.get("message"))['message']
return (
message.get("method") == "Network.responseReceived"
and "json" in message["params"]["response"]["mimeType"]
)
def __handle_captcha(self) -> None: def __handle_captcha(self) -> None:
for attempt in range(5): for attempt in range(5):
logging.info(f"Trying to solve captcha {attempt + 1}/5") logging.info(f"Trying to solve captcha {attempt + 1}/5")