spider/EasySpider
1
0
Fork 0
mirror of https://github.com/NaiboWang/EasySpider.git synced 2025-04-20 22:15:08 +08:00
Code Issues Packages Projects Releases Wiki Activity

Change keyboard module

Browse Source
This commit is contained in:
naibo 2023-07-09 08:59:11 +08:00
parent 261fb13006
commit bb6e2b2881
5 changed files with 36 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
.temp_to_pub/EasySpider_windows_x64/execution_instances/2.json Normal file
View File

File diff suppressed because one or more lines are too long

1
.temp_to_pub/EasySpider_windows_x64/execution_instances/3.json Normal file
View File

File diff suppressed because one or more lines are too long

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

@ -12,7 +12,7 @@
"justMyCode": true, "justMyCode": true,
// "args": ["--id", "[7]", "--read_type", "remote", "--headless", "0"] // "args": ["--id", "[7]", "--read_type", "remote", "--headless", "0"]
// "args": ["--id", "[9]", "--read_type", "remote", "--headless", "0", "--saved_file_name", "YOUTUBE"] // "args": ["--id", "[9]", "--read_type", "remote", "--headless", "0", "--saved_file_name", "YOUTUBE"]
"args": ["--id", "[6]", "--headless", "0", "--user_data", "1"] "args": ["--id", "[3]", "--headless", "0", "--user_data", "1"]
} }
] ]
} }

22
ExecuteStage/easyspider_executestage.py
View File

@ -39,10 +39,11 @@ import os
from commandline_config import Config from commandline_config import Config
import pytesseract import pytesseract
from PIL import Image from PIL import Image
from pynput.keyboard import Key, Listener
# import uuid # import uuid
from threading import Thread, Event from threading import Thread, Event
from myChrome import MyChrome, MyUCChrome from myChrome import MyChrome, MyUCChrome
from utils import check_pause, download_image, get_output_code, isnull, lowercase_tags_in_xpath, myMySQL, new_line, on_press, on_release_creator, write_to_csv, write_to_excel from utils import download_image, get_output_code, isnull, lowercase_tags_in_xpath, myMySQL, new_line, on_press, on_release_creator, write_to_csv, write_to_excel
desired_capabilities = DesiredCapabilities.CHROME desired_capabilities = DesiredCapabilities.CHROME
desired_capabilities["pageLoadStrategy"] = "none" desired_capabilities["pageLoadStrategy"] = "none"
@ -1504,19 +1505,20 @@ if __name__ == '__main__':
threads.append(thread) threads.append(thread)
thread.start() thread.start()
# Set the pause operation # Set the pause operation
if sys.platform != "linux": # if sys.platform != "linux":
Thread(target=check_pause, args=("p", event)).start() # Thread(target=check_pause, args=("p", event)).start()
else: # else:
from pynput.keyboard import Key, Listener time.sleep(3)
print("\n\n----------------------------------")
print("正在运行任务，按键盘p键可暂停任务的执行以便手工操作浏览器如输入验证码；如果想恢复任务的执行，请再次按p键。")
print("Running task, long press 'p' to pause the task for manual operation of the browser such as entering the verification code; If you want to resume the execution of the task, please long press 'p' again.")
print("----------------------------------\n\n")
# 使用监听器监听键盘输入 # 使用监听器监听键盘输入
with Listener(on_press=on_press, on_release=on_release_creator(event)) as listener: with Listener(on_press=on_press, on_release=on_release_creator(event)) as listener:
listener.join() listener.join()
time.sleep(5)
print("\n\n----------------------------------")
print("正在运行任务，长按键盘p键可暂停任务的执行以便手工操作浏览器如输入验证码；如果想恢复任务的执行，请再次长按p键。")
print("Running task, long press 'p' to pause the task for manual operation of the browser such as entering the verification code; If you want to resume the execution of the task, please long press 'p' again.")
print("----------------------------------\n\n")
for thread in threads: for thread in threads:
thread.join() thread.join()

40
ExecuteStage/utils.py
View File

@ -7,7 +7,7 @@ import os
import re import re
import time import time
import uuid import uuid
import keyboard # import keyboard
from openpyxl import Workbook, load_workbook from openpyxl import Workbook, load_workbook
import requests import requests
from urllib.parse import urlparse from urllib.parse import urlparse
@ -24,41 +24,41 @@ def is_valid_url(url):
def lowercase_tags_in_xpath(xpath): def lowercase_tags_in_xpath(xpath):
return re.sub(r"([A-Z]+)(?=[\[\]//]|$)", lambda x: x.group(0).lower(), xpath) return re.sub(r"([A-Z]+)(?=[\[\]//]|$)", lambda x: x.group(0).lower(), xpath)
def on_press(key):
pass
def on_release_creator(event): def on_release_creator(event):
def on_release(key): def on_release(key):
try: try:
if key.char == 'p': # 当按下esc键时，退出监听 if key.char == 'p': # 当按下esc键时，退出监听
if event._flag == False: if event._flag == False:
print("任务执行中，长按p键暂停执行。") print("任务执行中，按p键暂停执行。")
print("Task is running, long press 'p' to pause.") print("Task is running, long press 'p' to pause.")
# 设置Event的值为True，使得线程b可以继续执行 # 设置Event的值为True，使得线程b可以继续执行
event.set() event.set()
else: else:
# 设置Event的值为False，使得线程b暂停执行 # 设置Event的值为False，使得线程b暂停执行
print("任务已暂停，长按p键继续执行...") print("任务已暂停，按p键继续执行...")
print("Task paused, press 'p' to continue...") print("Task paused, press 'p' to continue...")
event.clear() event.clear()
except: except:
pass pass
return on_release return on_release
def check_pause(key, event): def on_press(key):
while True: pass
if keyboard.is_pressed(key): # 按下p键，暂停程序
if event._flag == False: # def check_pause(key, event):
print("任务执行中，长按p键暂停执行。") # while True:
print("Task is running, long press 'p' to pause.") # if keyboard.is_pressed(key): # 按下p键，暂停程序
# 设置Event的值为True，使得线程b可以继续执行 # if event._flag == False:
event.set() # print("任务执行中，长按p键暂停执行。")
else: # print("Task is running, long press 'p' to pause.")
# 设置Event的值为False，使得线程b暂停执行 # # 设置Event的值为True，使得线程b可以继续执行
print("任务已暂停，长按p键继续执行...") # event.set()
print("Task paused, press 'p' to continue...") # else:
event.clear() # # 设置Event的值为False，使得线程b暂停执行
time.sleep(1) # 每秒检查一次 # print("任务已暂停，长按p键继续执行...")
# print("Task paused, press 'p' to continue...")
# event.clear()
# time.sleep(1) # 每秒检查一次
def download_image(url, save_directory): def download_image(url, save_directory):