533 lines
19 KiB
Python
533 lines
19 KiB
Python
from selenium import webdriver as _root_webdriver
|
|
from selenium.webdriver.chrome.options import Options
|
|
from selenium.webdriver.chrome.service import Service
|
|
from selenium.webdriver.common.by import By
|
|
from selenium.webdriver.common.keys import Keys
|
|
from selenium.webdriver.common.alert import Alert
|
|
from selenium.webdriver.common.action_chains import ActionChains
|
|
from selenium.webdriver.support import expected_conditions as EC
|
|
from selenium.webdriver.support.wait import WebDriverWait
|
|
from selenium.common.exceptions import *
|
|
import tempfile
|
|
import platform
|
|
import random
|
|
import json
|
|
import time
|
|
import sys
|
|
import os
|
|
|
|
# You can set the drive path and browser location through the environment variables,
|
|
# and select the webdriver type.
|
|
# It has preferential priority through environmental variable settings.
|
|
SELENIUM_WIRE = os.environ.get('SELENIUM_WIRE') or ''
|
|
# Attention!!!
|
|
# Must set the environment before importing selenium.
|
|
# Otherwise, these settings will be invalid.
|
|
SELENIUM_BROWSER_CHOOSE = os.environ.get('SELENIUM_BROWSER_CHOOSE') or 'Chrome'
|
|
SELENIUM_BROWSER_DRIVER = os.environ.get('SELENIUM_BROWSER_DRIVER') or ''
|
|
SELENIUM_BROWSER_BINARY = os.environ.get('SELENIUM_BROWSER_BINARY') or ''
|
|
# Whether to turn on headless mode.
|
|
HEADLESS_BROWSER_ENABLE = os.environ.get('HEADLESS_BROWSER_ENABLE') == '1' or os.environ.get('HEADLESS') == '1'
|
|
# Choose Selenium or Selenium-Wire, default: Selenium.
|
|
if SELENIUM_WIRE:
|
|
exec('from seleniumwire import webdriver as _root_webdriver')
|
|
# Choose options and service.
|
|
if SELENIUM_BROWSER_CHOOSE:
|
|
exec('from selenium.webdriver.%s.options import Options' % (SELENIUM_BROWSER_CHOOSE or 'Chrome').lower())
|
|
exec('from selenium.webdriver.%s.service import Service' % (SELENIUM_BROWSER_CHOOSE or 'Chrome').lower())
|
|
# Use the chrome driver by default, such a code writing method is for grammar prompts.
|
|
DriverChoose = _root_webdriver.Chrome
|
|
if SELENIUM_BROWSER_CHOOSE:
|
|
exec('DriverChoose=_root_webdriver.%s' % (SELENIUM_BROWSER_CHOOSE or 'Chrome').capitalize())
|
|
|
|
|
|
class BrowserMobileEmulation(dict):
|
|
"""
|
|
Mobile emulation parameters.
|
|
"""
|
|
def __init__(self, w=540, h=960, user_agent=None):
|
|
du = 'Mozilla/5.0 (Linux; U; Android 13; zh-cn; 2109119BC Build/TKQ1.220829.002) ' \
|
|
'AppleWebKit/537.36 (KHTML, like Gecko) Version/4.0 ' \
|
|
'Chrome/98.0.4758.102 MQQBrowser/13.6 Mobile Safari/537.36'
|
|
user_agent = user_agent or du
|
|
super().__init__({'w': w, 'h': h, 'user_agent': user_agent})
|
|
self.w = self.h = self.user_agent = None
|
|
|
|
def __setattr__(self, key, value):
|
|
pass
|
|
|
|
def __getitem__(self, item):
|
|
try:
|
|
return super().__getitem__(item)
|
|
except KeyError:
|
|
return None
|
|
|
|
def __getattr__(self, item):
|
|
try:
|
|
return super().__getitem__(item)
|
|
except KeyError:
|
|
return None
|
|
|
|
|
|
class DefaultChromeLocation:
|
|
"""
|
|
Set default chrome and driver location.
|
|
"""
|
|
xDriver = '/opt/google/chrome/chromedriver'
|
|
xChrome = '/opt/google/chrome/chrome'
|
|
wDriver = 'C:/Program Files/Google/Chrome/Application/chromedriver.exe'
|
|
wChrome = 'C:/Program Files/Google/Chrome/Application/chrome.exe'
|
|
|
|
|
|
class PositionTab:
|
|
"""
|
|
Position for switch tab.
|
|
"""
|
|
Prev = 'Go-Prev'
|
|
Next = 'Go-Next'
|
|
|
|
|
|
class ColorUtils:
|
|
"""
|
|
Color utils.
|
|
"""
|
|
@staticmethod
|
|
def hex2rgb(color):
|
|
color = color[1:].upper()
|
|
for x in color:
|
|
if x not in '0123456789ABCDEF':
|
|
raise Exception('Found invalid hexa character {0}.'.format(x))
|
|
if len(color) == 6 or len(color) == 8:
|
|
color = '#' + color[0:6]
|
|
elif len(color) == 3:
|
|
color = '#' + color[0] * 2 + color[1] * 2 + color[2] * 2
|
|
else:
|
|
raise Exception('Hexa string should be 3, 6 or 8 digits. if 8 digits, last 2 are ignored.')
|
|
hexcolor = color[1:]
|
|
r, g, b = int(hexcolor[0:2], 16), int(hexcolor[2:4], 16), int(hexcolor[4:6], 16)
|
|
return r, g, b
|
|
|
|
|
|
class Browser(DriverChoose):
|
|
"""
|
|
Browser web driver.
|
|
"""
|
|
def __init__(
|
|
self,
|
|
driver: str = None,
|
|
binary: str = None,
|
|
headless: bool = False,
|
|
lang: str = None,
|
|
mute: bool = False,
|
|
no_images: bool = False,
|
|
user_agent: str = None,
|
|
http_proxy: str = None,
|
|
home: str = None,
|
|
window_size: str = None,
|
|
mobile_emulation: BrowserMobileEmulation = None,
|
|
option_arguments: list = None,
|
|
req_interceptor=None,
|
|
res_interceptor=None,
|
|
):
|
|
choose = SELENIUM_BROWSER_CHOOSE.capitalize()
|
|
self.platform = platform.uname().system
|
|
default_driver = None
|
|
default_binary = None
|
|
headless = HEADLESS_BROWSER_ENABLE or headless
|
|
if choose == 'Chrome':
|
|
if self.platform == 'Linux':
|
|
default_driver = DefaultChromeLocation.xDriver
|
|
default_binary = DefaultChromeLocation.xChrome
|
|
else:
|
|
default_driver = DefaultChromeLocation.wDriver
|
|
default_binary = DefaultChromeLocation.wChrome
|
|
driver = SELENIUM_BROWSER_DRIVER or driver or default_driver
|
|
binary = SELENIUM_BROWSER_BINARY or binary or default_binary
|
|
if self.platform == 'Linux' and not window_size: window_size = '1920x1080'
|
|
if self.platform == 'Windows' and headless and not window_size: window_size = '1920x1080'
|
|
# Initialization settings.
|
|
if (isinstance(option_arguments, list)) is False: option_arguments = []
|
|
cdplist = []
|
|
service = Service()
|
|
options = Options()
|
|
self.cdplist = cdplist
|
|
# Delete prompt information of chrome being controlled.
|
|
hasattr(options, 'add_experimental_option') and options.add_experimental_option("excludeSwitches", ['enable-automation', 'enable-logging'])
|
|
# Mobile emulation parameter setting start.
|
|
if mobile_emulation:
|
|
if hasattr(options, 'add_experimental_option') is False:
|
|
raise Exception('Do not support mobile emulation currently.')
|
|
self.w_browser = mobile_emulation.w + 14
|
|
self.h_browser = mobile_emulation.h + 0
|
|
self.w_inner_window = mobile_emulation.w + 0
|
|
self.h_inner_window = mobile_emulation.h - 86
|
|
self.mobile_emulation_screen_w = mobile_emulation.w
|
|
self.mobile_emulation_screen_h = mobile_emulation.h
|
|
window_size = '%s,%s' % (self.w_browser, self.h_browser)
|
|
options.add_experimental_option(
|
|
'mobileEmulation', {
|
|
'deviceMetrics': {
|
|
'width': self.w_inner_window,
|
|
'height': self.h_inner_window,
|
|
'pixelRatio': 2.75,
|
|
'touch': True
|
|
},
|
|
'userAgent': mobile_emulation.user_agent
|
|
}
|
|
)
|
|
cdplist.append([
|
|
'Emulation.setUserAgentOverride', {
|
|
'userAgent': mobile_emulation.user_agent,
|
|
'userAgentMetadata': {
|
|
'platform': 'Android' if mobile_emulation.user_agent.find('iPhone') == -1 else 'iPhone',
|
|
'mobile': True,
|
|
'platformVersion': '',
|
|
'architecture': '',
|
|
'model': ''
|
|
}
|
|
}
|
|
])
|
|
else:
|
|
self.w_browser = 0
|
|
self.h_browser = 0
|
|
self.w_inner_window = 0
|
|
self.h_inner_window = 0
|
|
self.mobile_emulation_screen_w = 0
|
|
self.mobile_emulation_screen_h = 0
|
|
# Mobile emulation parameter setting end.
|
|
# Set browser and webdriver path.
|
|
if driver:
|
|
service.path = driver
|
|
if binary:
|
|
options.binary_location = binary
|
|
# Add webdriver option arguments.
|
|
for i in option_arguments:
|
|
options.add_argument(i)
|
|
# Set headless mode.
|
|
if self.platform == 'Linux' or headless:
|
|
options.add_argument('--headless')
|
|
# Set no-sandbox mode.
|
|
if self.platform == 'Linux':
|
|
options.add_argument('--no-sandbox')
|
|
options.add_argument('--disable-dev-shm-usage')
|
|
options.add_argument('--disable-gpu')
|
|
# Set language of browser, default is zh-CN.
|
|
if lang:
|
|
options.add_argument('--lang=%s' % (lang or 'zh-CN'))
|
|
hasattr(options, 'set_preference') and options.set_preference('intl.accept_languages', lang or 'zh-CN')
|
|
# Set mute.
|
|
if mute:
|
|
options.add_argument('--mute-audio=true')
|
|
hasattr(options, 'set_preference') and print('Warning: Do not support mute audio currently.', file=sys.stderr)
|
|
# Set no images mode.
|
|
if no_images:
|
|
options.add_argument('--blink-settings=imagesEnabled=false')
|
|
hasattr(options, 'set_preference') and print('Warning: Do not support disable images currently.', file=sys.stderr)
|
|
# Set default user agent.
|
|
if user_agent:
|
|
options.add_argument('--user-agent=%s' % user_agent)
|
|
hasattr(options, 'set_preference') and options.set_preference('general.useragent.override', user_agent)
|
|
# Set http proxy for browser.
|
|
if http_proxy:
|
|
options.add_argument('--proxy-server=http://%s' % http_proxy)
|
|
# Set browser window size before startup.
|
|
if window_size:
|
|
options.add_argument('--window-size=%s' % window_size.replace("\x20", '').replace('x', ','))
|
|
else:
|
|
options.add_argument('--start-maximized')
|
|
# Start the browser.
|
|
super().__init__(service=service, options=options)
|
|
# Selenium-Wire backend optimization start.
|
|
try:
|
|
self.backend.master.options.add_option('ssl_insecure', bool, True, 'Do not verify upstream server SSL/TLS certificates.')
|
|
self.backend.master.options.add_option('upstream_cert', bool, False, 'Connect to upstream server to look up certificate details.')
|
|
self.backend.master.options.add_option('http2', bool, False, 'Enable/disable HTTP/2 support.')
|
|
except AttributeError:
|
|
pass
|
|
# Selenium-Wire backend optimization end.
|
|
if mobile_emulation:
|
|
cdplist.append(['Emulation.setFocusEmulationEnabled', {'enabled': True}])
|
|
cdplist.append(['Emulation.setTouchEmulationEnabled', {'enabled': True, 'maxTouchPoints': 5}])
|
|
cdplist.append(['Emulation.setEmitTouchEventsForMouse', {'enabled': True, 'configuration': 'mobile'}])
|
|
# Set the request and response interceptor.
|
|
if req_interceptor:
|
|
hasattr(self, 'backend') or print('Warning: Can not use the interceptor, because not extends Seleniun-Wire.', file=sys.stderr)
|
|
self.request_interceptor = req_interceptor
|
|
if res_interceptor:
|
|
hasattr(self, 'backend') or print('Warning: Can not use the interceptor, because not extends Seleniun-Wire.', file=sys.stderr)
|
|
self.response_interceptor = res_interceptor
|
|
# Sync set http proxy for Selenium-Wire backend.
|
|
if http_proxy:
|
|
self.proxy = {'http': 'http://%s' % http_proxy, 'https': 'https://%s' % http_proxy}
|
|
# Set browser window size after startup, by default, there will be full screen display window.
|
|
if window_size:
|
|
self.set_window_size(*window_size.replace("\x20", '').replace('x', ',').split(','))
|
|
else:
|
|
self.maximize_window()
|
|
# Sets a sticky timeout to implicitly wait for an element to be found.
|
|
self.implicitly_wait(10)
|
|
# Set the amount of time to wait for a page load to complete.
|
|
self.set_page_load_timeout(25)
|
|
# Open the default page.
|
|
home and self.open(home)
|
|
|
|
@staticmethod
|
|
def wait(secs: int | float = 1):
|
|
"""
|
|
Will sleep waiting.
|
|
"""
|
|
number_int = int(secs)
|
|
number_float = secs - number_int
|
|
for i in range(number_int):
|
|
time.sleep(1)
|
|
else:
|
|
time.sleep(number_float)
|
|
|
|
def open(self, url=None):
|
|
"""
|
|
Open the URL, simulate into the URL in the address bar and jump, the new page has no Referrer.
|
|
"""
|
|
self.update_cdp_command()
|
|
return self.get(url)
|
|
|
|
def turn(self, url=None):
|
|
"""
|
|
Simulation "window.location.href" jumps, the new page has Referrer.
|
|
"""
|
|
return self.execute_script('window.location.href=%s;' % json.dumps(url, indent=None, ensure_ascii=True), None)
|
|
|
|
def find(self, path):
|
|
"""
|
|
Use XPath to find an element.
|
|
"""
|
|
ele = self.find_element(By.XPATH, path)
|
|
self.element_prominent(ele, '#f8be5f')
|
|
return ele
|
|
|
|
def find_mult(self, path):
|
|
"""
|
|
Use XPath to find elements.
|
|
"""
|
|
ele = self.find_elements(By.XPATH, path)
|
|
len(ele) > 0 and [self.element_prominent(e, '#f8be5f') for e in ele]
|
|
return ele
|
|
|
|
def find_element_by(self, sentence):
|
|
"""
|
|
Custom find element, pass into a tuple or list.
|
|
"""
|
|
ele = self.find_element(*sentence)
|
|
self.element_prominent(ele, '#f8be5f')
|
|
return ele
|
|
|
|
def click(self, element):
|
|
"""
|
|
Click element for desktop version.
|
|
"""
|
|
self.element_prominent(element, '#ff0000')
|
|
self.action_chains().reset_actions()
|
|
self.action_chains().click(element).perform()
|
|
self.wait(0.1)
|
|
|
|
def touch(self, x, y):
|
|
"""
|
|
Click on the coordinates for Mobile edition.
|
|
"""
|
|
self.action_chains().reset_actions()
|
|
self.action_chains().move_by_offset(x, y).click().perform()
|
|
self.wait(0.1)
|
|
|
|
def input(self, element, content):
|
|
"""
|
|
Enter the content to the element.
|
|
"""
|
|
self.element_prominent(element, '#00b6f1')
|
|
self.action_chains().reset_actions()
|
|
self.action_chains().send_keys_to_element(element, content).perform()
|
|
self.wait(0.1)
|
|
|
|
def mouse(self, element):
|
|
"""
|
|
Park the mouse here.
|
|
"""
|
|
self.element_prominent(element, '#49dc07')
|
|
self.action_chains().reset_actions()
|
|
self.action_chains().move_to_element(element).perform()
|
|
|
|
def tab_create(self, url=None):
|
|
"""
|
|
Create a new tab and open the URL.
|
|
"""
|
|
self.switch_to.new_window('tab')
|
|
self.update_cdp_command()
|
|
url and self.open(url)
|
|
|
|
def tab_switch(self, tab: int | str):
|
|
"""
|
|
Switch the browser tab page
|
|
"""
|
|
handles = self.window_handles
|
|
lengths = len(handles)
|
|
current = handles.index(self.current_window_handle)
|
|
if isinstance(tab, int):
|
|
handle = tab
|
|
elif tab == PositionTab.Prev:
|
|
handle = (current - 1)
|
|
elif tab in PositionTab.Next:
|
|
handle = (current + 1) % lengths
|
|
else:
|
|
handle = None
|
|
self.switch_to.window(handles[handle])
|
|
self.wait(0.2)
|
|
self.update_cdp_command()
|
|
|
|
def tab_cancel(self):
|
|
"""
|
|
Close the current browser tab page.
|
|
"""
|
|
handles = self.window_handles
|
|
if len(handles):
|
|
current = handles.index(self.current_window_handle)
|
|
self.close()
|
|
current > 0 and self.switch_to.window(handles[current - 1])
|
|
self.wait(0.2)
|
|
|
|
def tab_cancel_all(self):
|
|
"""
|
|
Close all the browser tab page.
|
|
"""
|
|
handles = self.window_handles
|
|
for i in handles:
|
|
self.tab_cancel()
|
|
|
|
def frame_switch_to(self, element_of_frame):
|
|
"""
|
|
Switch frame to the specified frame element.
|
|
"""
|
|
self.switch_to.frame(element_of_frame)
|
|
self.wait(0.2)
|
|
|
|
def frame_switch_to_default(self):
|
|
"""
|
|
Switch to the default frame.
|
|
"""
|
|
self.switch_to.default_content()
|
|
self.wait(0.2)
|
|
|
|
def scroll(self):
|
|
"""
|
|
Scroll page.
|
|
:return:
|
|
"""
|
|
self.action_chains().reset_actions()
|
|
self.action_chains().scroll_by_amount(0, self.execute_script('return document.documentElement.clientHeight;')).perform()
|
|
self.wait(0.8)
|
|
|
|
def scroll_to(self, pos: int | str):
|
|
"""
|
|
Scroll to the specified location.
|
|
"""
|
|
if isinstance(pos, int) and pos > 0:
|
|
self.execute_script('window.scrollTo(0, arguments[0]);', pos)
|
|
elif pos == 0:
|
|
self.execute_script('window.scrollTo(0, 0);')
|
|
elif pos == 0 - 1:
|
|
self.execute_script('window.scrollTo(0, document.body.scrollHeight);')
|
|
else:
|
|
pass
|
|
self.wait(0.8)
|
|
|
|
def scroll_to_element(self, element):
|
|
"""
|
|
Scroll to the specified element location.
|
|
"""
|
|
self.action_chains().reset_actions()
|
|
self.action_chains().scroll_to_element(element).perform()
|
|
self.wait(0.8)
|
|
|
|
def element_force_display(self, element):
|
|
"""
|
|
Make hidden element visible and interactive.
|
|
"""
|
|
self.execute_script(
|
|
'let e=arguments[0];e.style.display="inline-block";e.style.visibility="visible";e.setAttribute("hidden","false");', element
|
|
)
|
|
|
|
def element_prominent(self, element, color='#ff0000', dura=2500):
|
|
"""
|
|
Make the element highlight.
|
|
"""
|
|
if not element:
|
|
return False
|
|
high = ColorUtils.hex2rgb(color)
|
|
r = high[0]
|
|
g = high[1]
|
|
b = high[2]
|
|
self.execute_script('''
|
|
let e=arguments[0];
|
|
try{
|
|
let o=[e.style.background||null,e.style.border||null];
|
|
e.style.border="1px solid %s";e.style.background="rgba(%s,%s,%s,0.2)";
|
|
if(!e.prominent){
|
|
e.prominent=true;
|
|
setTimeout(function(args){try{args[0].prominent=null;args[0].style.background=args[1][0];args[0].style.border=args[1][1]}catch(e){}},%s,[e,o]);
|
|
}
|
|
}catch(e){}''' % (color, r, g, b, dura), element
|
|
)
|
|
|
|
def webdriver_wait(self, timeout: float, poll_frequency: float = 0.5, ignored_exceptions=None):
|
|
"""
|
|
Return WebDriverWait object.
|
|
"""
|
|
return WebDriverWait(
|
|
driver=self,
|
|
timeout=timeout,
|
|
poll_frequency=poll_frequency,
|
|
ignored_exceptions=ignored_exceptions
|
|
)
|
|
|
|
def current_alert(self):
|
|
"""
|
|
Return current alert object.
|
|
"""
|
|
return Alert(self)
|
|
|
|
def window_inner_size(self):
|
|
"""
|
|
Get the page window inner size.
|
|
"""
|
|
size = self.execute_script('return [window.innerWidth, window.innerHeight];')
|
|
return {'w': size[0] or 0, 'h': size[1] or 0}
|
|
|
|
def action_chains(self):
|
|
"""
|
|
Return ActionChains object.
|
|
"""
|
|
return ActionChains(self)
|
|
|
|
def screenshot(self) -> bytes:
|
|
"""
|
|
Screenshot as bytes.
|
|
"""
|
|
return self.get_screenshot_as_png()
|
|
|
|
def update_cdp_command(self) -> None:
|
|
for cmd in self.cdplist:
|
|
self.execute_cdp_cmd(*cmd)
|
|
|
|
|
|
class WebDriver(Browser):
|
|
"""
|
|
Get a browser driver object.
|
|
"""
|
|
def __init__(self):
|
|
super().__init__(lang='zh-CN')
|
|
|
|
|
|
if __name__ == '__main__':
|
|
# e.g. Test it can work normally.
|
|
driver = WebDriver()
|
|
driver.open('https://www.hao123.com/')
|
|
driver.wait()
|
|
driver.quit()
|