spider/EasySpider
1
0
Fork 0
mirror of https://github.com/NaiboWang/EasySpider.git synced 2025-04-22 11:01:52 +08:00
Code Issues Packages Projects Releases Wiki Activity

V0.3.5

Browse Source
This commit is contained in:
naibo 2023-07-09 01:41:26 +08:00
parent feb3a037b2
commit b8f8c190ba
2 changed files with 31 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

39
.temp_to_pub/compress.py
View File

@ -14,8 +14,13 @@ import py7zr
def compress_folder_to_7z(folder_path, output_file):
if os.path.exists(output_file):
os.remove(output_file)
with py7zr.SevenZipFile(output_file, 'w') as archive:
archive.writeall(folder_path, output_file)
# with py7zr.SevenZipFile(output_file, 'w') as archive:
# archive.writeall(folder_path, output_file)
# 压缩文件夹
try:
subprocess.call(["7z", "a", output_file, folder_path])
except:
subprocess.call(["7za", "a", output_file, folder_path])
def compress_folder_to_7z_split(folder_path, output_file):
if os.path.exists(output_file):
@ -54,11 +59,13 @@ if __name__ == "__main__":
shutil.rmtree("./EasySpider_windows_x64/Data")
if os.path.exists("./EasySpider_windows_x64/execution_instances"):
shutil.rmtree("./EasySpider_windows_x64/execution_instances")
os.remove("./EasySpider_windows_x64/config.json")
os.remove("./EasySpider_windows_x64/mysql_config.json")
if os.path.exists("./EasySpider_windows_x64/config.json"):
os.remove("./EasySpider_windows_x64/config.json")
if os.path.exists("./EasySpider_windows_x64/mysql_config.json"):
os.remove("./EasySpider_windows_x64/mysql_config.json")
os.mkdir("./EasySpider_windows_x64/Data")
os.mkdir("./EasySpider_windows_x64/execution_instances")
compress_folder_to_7z_split("./EasySpider_windows_x64", file_name)
# compress_folder_to_7z_split("./EasySpider_windows_x64", file_name)
print(f"Compress {file_name} Split successfully!")
compress_folder_to_7z("./EasySpider_windows_x64", file_name)
print(f"Compress {file_name} successfully!")
@ -70,20 +77,28 @@ if __name__ == "__main__":
shutil.rmtree("./EasySpider_windows_x86/Data")
if os.path.exists("./EasySpider_windows_x86/execution_instances"):
shutil.rmtree("./EasySpider_windows_x86/execution_instances")
os.remove("./EasySpider_windows_x86/config.json")
os.remove("./EasySpider_windows_x86/mysql_config.json")
if os.path.exists("./EasySpider_windows_x86/config.json"):
os.remove("./EasySpider_windows_x86/config.json")
if os.path.exists("./EasySpider_windows_x86/mysql_config.json"):
os.remove("./EasySpider_windows_x86/mysql_config.json")
os.mkdir("./EasySpider_windows_x86/Data")
os.mkdir("./EasySpider_windows_x86/execution_instances")
compress_folder_to_7z("./EasySpider_windows_x64", file_name)
compress_folder_to_7z_split("./EasySpider_windows_x86", file_name)
print(f"Compress {file_name} Split successfully!")
compress_folder_to_7z("./EasySpider_windows_x86", file_name)
print(f"Compress {file_name} successfully!")
elif sys.platform == "linux" and platform.architecture()[0] == "64bit":
file_name = f"EasySpider_{easyspider_version}_Linux_x64.7z"
if os.path.exists("./EasySpider_Linux_x64/user_data"):
shutil.rmtree("./EasySpider_Linux_x64/user_data")
shutil.rmtree("./EasySpider_Linux_x64/Data")
shutil.rmtree("./EasySpider_Linux_x64/execution_instances")
shutil.rmtree("./EasySpider_Linux_x64/config.json")
shutil.rmtree("./EasySpider_Linux_x64/mysql_config.json")
if os.path.exists("./EasySpider_Linux_x64/Data"):
shutil.rmtree("./EasySpider_Linux_x64/Data")
if os.path.exists("./EasySpider_Linux_x64/execution_instances"):
shutil.rmtree("./EasySpider_Linux_x64/execution_instances")
if os.path.exists("./EasySpider_Linux_x64/config.json"):
os.remove("./EasySpider_Linux_x64/config.json")
if os.path.exists("./EasySpider_Linux_x64/mysql_config.json"):
os.remove("./EasySpider_Linux_x64/mysql_config.json")
os.mkdir("./EasySpider_Linux_x64/Data")
os.mkdir("./EasySpider_Linux_x64/execution_instances")
# compress_folder_to_7z("./EasySpider_Linux_x64", file_name)

5
ExecuteStage/easyspider_executestage.py
View File

@ -1144,7 +1144,10 @@ class BrowserThread(Thread):
# 提取数据事件
def getData(self, para, loopElement, isInLoop=True, parentPath="", index=0):
pageHTML = etree.HTML(self.browser.page_source)
try:
pageHTML = etree.HTML(self.browser.page_source)
except:
pageHTML = ""
if loopElement != "": # 只在数据在循环中提取时才需要获取循环元素
try:
loopElementOuterHTML = loopElement.get_attribute('outerHTML')