v4l3n71n/web_scrap
1
0
Fork 0
Code Issues Pull Requests Projects Releases 10 Wiki Activity

finish menu with album import

Browse Source
This commit is contained in:
Valentin CZERYBA 2023-08-25 23:46:43 +02:00
parent 1fc9c48d2c
commit 6401692d0d
2 changed files with 7 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
import_export_canalblog.py
View File

@ -170,7 +170,7 @@ if __name__ == '__main__':
import_parser.add_argument("--no-create", help="No create post", dest="create", default="store_false", action="store_true")
import_parser.add_argument("--no-update", help="No update post", dest="update", default="store_false", action="store_true")
import_parser.add_argument("--no-image", help="No image add or update", dest="image", default="store_false", action="store_true")
import_parser.add_argument("--author", help="Define author for page album", default="")
import_parser.add_argument("--author-album", dest=author, help="Define author for page album", default="")
@ -264,7 +264,6 @@ if __name__ == '__main__':
for i in wordpress:
importWp.setUrl(i)
importWp.fromFile(files=args.file.split(","))
exit(0)
menuWp = WPMenu(name="Thread-1", basic=basic, wordpress=args.wordpress, logger=logger, parser=args.parser, ssl_canalblog=ssl_canalblog, ssl_wordpress=ssl_wordpress)
menuWp.fromFile("{0}".format(args.file.split(",")[0]))
if len(args.directory) > 0:

10
lib/WPMenu.py
View File

@ -91,9 +91,11 @@ class WPMenu:
search = "posts"
post_type = "post"
if len(articletitle) == 0:
articletitle = soup.find("div", class_="albumbody").find("h2").get_text()
search = "pages"
post_type = "page"
articletitle = soup.find_all("div", class_="albumbody")
if len(articletitle) > 0:
articletitle = articletitle[0].find("h2").get_text()
search = "pages"
post_type = "page"
exist = False
for index in range(1,10):
@ -171,7 +173,7 @@ class WPMenu:
title = title[::-1]
self._logger.info("{0} link {1} title {2}".format(self._name, link, title))
if link == "index.html":
if second_title == "archives" or second_title == "albums":
if second_title == "albums":
idMenu = self._getIdFromPost(href)
else:
idMenu = self._getId(title)