How to strip CSS/style nodes? #100
Comments
|
@Chichilele - It looks like the implementation for so I think you are just getting the text content of your Perhaps simply stripping the markup of a tag is the intended behavior, and we need a new |
|
I had exactly the same problem as this; I hacked it as follows: diff --git a/markdownify/__init__.py b/markdownify/__init__.py
index bf765ec..cb27487 100644
--- a/markdownify/__init__.py
+++ b/markdownify/__init__.py
@@ -117,6 +117,9 @@ class MarkdownConverter(object):
'table', 'thead', 'tbody', 'tfoot',
'tr', 'td', 'th']
+ if node.name == "style":
+ return ""
+
if is_nested_node(node):
for el in node.children:
# Only extract (remove) whitespace-only text node if any of theThis is likely by no means perfect, but worked 100% correctly for the multiple inputs I was trying to convert. |
|
Another simple workaround: from markdownify import markdownify as md
import re
my_html = '...'
stripped_html = re.sub(r'<style>(.*)</style>', '', my_html, flags=re.DOTALL)
markdown = md(stripped_html) |
Yeah, it's not about stripping content, but stripping tags. Generally, the body of the elements is actual content being marked up; style is a bit of an outlier because the body isn't really content. A new option makes sense. We could probably also do with a README update, though; I don't think people realize that you can write custom |
|
@matthewwithanm I suggest you apply a bugfix along the lines of what was suggested by @aytey. I read the readme section about custom converters and it starts with "If you have a special usecase". The thing is, I do not think this style-issue is a special use case at all - in fact, I cannot think of a use case where the current behaviour is desirable. Maybe I am missing something? |
|
@tlk Yeah, I'm not saying the library shouldn't handle this. All I'm saying is that, if you find a case where a tag isn't being handled like you want, you don't need a patch like @aytey's suggestion, or to preprocess like you suggested. You just need to subclass and add a |
|
I'm glad this opened a conversation. Although I went down another route for my project, here is what I would have used to solve this issue (based on the suggestions above): from markdownify import MarkdownConverter
class MarkdownConverterNoStyle(MarkdownConverter):
def convert_style(self, el, text, convert_as_inline):
return ''
html = '<!DOCTYPE html><html><head><style>body {background-color: powderblue;}h1 {color: blue;}p {color: red;}</style></head><body><p>This is text</p></body></html>'
MarkdownConverterNoStyle().convert(html)Returns a nice The @tlk I share your opinion that style should be left out by default |
|
@matthewwithanm Got it - good point! Come to think of it, the script tag would probably also need to be dealt with. PR #112 will make the markdownify library ignore both the script tag and the style tag. |
|
Fixed in 0.12.1. |
I may be misusing the option
stripbut I can't get rid of the style bitsThe text was updated successfully, but these errors were encountered: