spider/EasySpider
1
0
Fork 0
mirror of https://github.com/NaiboWang/EasySpider.git synced 2025-04-12 03:27:08 +08:00
Code Issues Packages Projects Releases Wiki Activity

Screenshots zoom to the maximum size under headless mode

Browse Source
This commit is contained in:
naibo 2025-01-08 12:02:36 +08:00
parent f43bdd236d
commit 119cb99711
2 changed files with 15 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
ExecuteStage/.vscode/launch.json vendored
View File

@ -12,7 +12,7 @@
"justMyCode": false, "justMyCode": false,
// "args": ["--ids", "[7]", "--read_type", "remote", "--headless", "0"] // "args": ["--ids", "[7]", "--read_type", "remote", "--headless", "0"]
// "args": ["--ids", "[9]", "--read_type", "remote", "--headless", "0", "--saved_file_name", "YOUTUBE"] // "args": ["--ids", "[9]", "--read_type", "remote", "--headless", "0", "--saved_file_name", "YOUTUBE"]
"args": ["--ids", "[8]", "--headless", "0", "--user_data", "0", "--keyboard", "0", "args": ["--ids", "[0]", "--headless", "0", "--user_data", "0", "--keyboard", "0",
"--read_type", "remote", "--read_type", "remote",
] ]
// "args": "--ids '[97]' --user_data 1 --server_address http://localhost:8074 --config_folder '/Users/naibo/Documents/EasySpider/ElectronJS/' --headless 0 --read_type remote --config_file_name config.json --saved_file_name" // "args": "--ids '[97]' --user_data 1 --server_address http://localhost:8074 --config_folder '/Users/naibo/Documents/EasySpider/ElectronJS/' --headless 0 --read_type remote --config_file_name config.json --saved_file_name"

20
ExecuteStage/easyspider_executestage.py
View File

@ -73,13 +73,13 @@ desired_capabilities["pageLoadStrategy"] = "none"
class BrowserThread(Thread): class BrowserThread(Thread):
def __init__(self, browser_t, id, service, version, event, saveName, config, option): def __init__(self, browser_t, id, service, version, event, saveName, config, option, commandline_config=""):
Thread.__init__(self) Thread.__init__(self)
self.logs = io.StringIO() self.logs = io.StringIO()
self.log = bool(service.get("recordLog", True)) self.log = bool(service.get("recordLog", True))
self.browser = browser_t self.browser = browser_t
self.option = option self.option = option
self.config = config self.commandline_config = commandline_config
self.version = version self.version = version
self.totalSteps = 0 self.totalSteps = 0
self.id = id self.id = id
@ -1870,7 +1870,15 @@ class BrowserThread(Thread):
width = size["width"] width = size["width"]
height = size["height"] height = size["height"]
# 调整浏览器窗口的大小 # 调整浏览器窗口的大小
self.browser.set_window_size(width, height) if self.commandline_config["headless"] == 1: # 无头模式下，截取整个网页的高度
page_width = self.browser.execute_script(
"return document.body.scrollWidth")
page_height = self.browser.execute_script(
"return document.body.scrollHeight")
self.browser.set_window_size(page_width, page_height)
time.sleep(1)
else:
self.browser.set_window_size(width, height)
element.screenshot("Data/Task_" + str(self.id) + "/" + self.saveName + element.screenshot("Data/Task_" + str(self.id) + "/" + self.saveName +
"/screenshots/" + str(time.time()) + ".png") "/screenshots/" + str(time.time()) + ".png")
# 截图完成后，将浏览器的窗口大小设置为原来的大小 # 截图完成后，将浏览器的窗口大小设置为原来的大小
@ -2183,7 +2191,7 @@ class BrowserThread(Thread):
if __name__ == '__main__': if __name__ == '__main__':
# 如果需要调试程序，请在命令行参数中加入--keyboard 0 来禁用键盘监听以提升调试速度 # 如果需要调试程序，请在命令行参数中加入--keyboard 0 来禁用键盘监听以提升调试速度
# If you need to debug the program, please add --keyboard 0 in the command line parameters to disable keyboard listening to improve debugging speed # If you need to debug the program, please add --keyboard 0 in the command line parameters to disable keyboard listening to improve debugging speed
config = { commandline_config = {
"ids": [0], "ids": [0],
"saved_file_name": "", "saved_file_name": "",
"user_data": False, "user_data": False,
@ -2198,7 +2206,7 @@ if __name__ == '__main__':
"docker_driver": "", "docker_driver": "",
"user_folder": "", "user_folder": "",
} }
c = Config(config) c = Config(commandline_config)
print(c) print(c)
options = webdriver.ChromeOptions() options = webdriver.ChromeOptions()
driver_path = "chromedriver.exe" driver_path = "chromedriver.exe"
@ -2440,7 +2448,7 @@ if __name__ == '__main__':
event = Event() event = Event()
event.set() event.set()
thread = BrowserThread(browser_t, id, service, thread = BrowserThread(browser_t, id, service,
c.version, event, c.saved_file_name, config=config, option=tmp_options[i]) c.version, event, c.saved_file_name, config=config, option=tmp_options[i], commandline_config=c)
print("Thread with task id: ", id, " is created") print("Thread with task id: ", id, " is created")
threads.append(thread) threads.append(thread)
thread.start() thread.start()