pictuga
/
morss
1
1
Fork 0
Code Issues Pull Requests Actions Packages Releases Activity

readabilite: remove code duplicate

master
pictuga 2022-01-23 11:41:32 +01:00
parent fe5dbf1ce0
commit e81f6b173f
1 changed files with 2 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
morss/readabilite.py
View File

@ -207,8 +207,10 @@ def clean_root(root, keep_threshold=None):
def clean_node(node, keep_threshold=None): def clean_node(node, keep_threshold=None):
parent = node.getparent() parent = node.getparent()
# remove comments
if (isinstance(node, lxml.html.HtmlComment) if (isinstance(node, lxml.html.HtmlComment)
or isinstance(node, lxml.html.HtmlProcessingInstruction)): or isinstance(node, lxml.html.HtmlProcessingInstruction)):
parent.remove(node)
return return
if parent is None: if parent is None:
@ -242,11 +244,6 @@ def clean_node(node, keep_threshold=None):
parent.remove(node) parent.remove(node)
return return
# remove comments
if isinstance(node, lxml.html.HtmlComment) or isinstance(node, lxml.html.HtmlProcessingInstruction):
parent.remove(node)
return
# remove if too many kids & too high link density # remove if too many kids & too high link density
wc = count_words(node.text_content()) wc = count_words(node.text_content())
if wc != 0 and len(list(node.iter())) > 3: if wc != 0 and len(list(node.iter())) > 3: