Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Recipe susbsytem: Add a remove_attributes property to make it easy to specify that some attributes should be removed from all tags

Browse Source
This commit is contained in:
Kovid Goyal 2009-08-15 12:15:24 -06:00
parent d49f21d47e
commit babe84e4e5
1 changed files with 9 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
src/calibre/web/feeds/news.py
View File

@ -210,6 +210,12 @@ class BasicNewsRecipe(Recipe):
#: tags before the first element with `id="content"`. #: tags before the first element with `id="content"`.
remove_tags_before = None remove_tags_before = None
#: List of attributes to remove from all tags
#: For example::
#:
#: remove_attributes = ['style', 'font']
remove_attributes = []
#: Keep only the specified tags and their children. #: Keep only the specified tags and their children.
#: For the format for specifying a tag see :attr:`BasicNewsRecipe.remove_tags`. #: For the format for specifying a tag see :attr:`BasicNewsRecipe.remove_tags`.
#: If this list is not empty, then the `<body>` tag will be emptied and re-filled with #: If this list is not empty, then the `<body>` tag will be emptied and re-filled with
@ -562,6 +568,9 @@ class BasicNewsRecipe(Recipe):
script.extract() script.extract()
for script in list(soup.findAll('noscript')): for script in list(soup.findAll('noscript')):
script.extract() script.extract()
for attr in self.remove_attributes:
for x in soup.findAll(attrs={attr:True}):
del x[attr]
return self.postprocess_html(soup, first_fetch) return self.postprocess_html(soup, first_fetch)