Mirrors/yt-dlc
1
0
Fork 0
mirror of https://github.com/blackjack4494/yt-dlc.git synced 2024-11-15 15:32:31 +01:00
Code Issues Projects Releases Wiki Activity

[utlis] add extract_attributes for extracting html tags attributes

Browse Source
This commit is contained in:
remitamine 2015-09-11 04:44:17 +01:00
parent 1721fef28b
commit 689fb748ee
1 changed files with 8 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
youtube_dl/utils.py
View File

@ -248,6 +248,14 @@ def get_element_by_attribute(attribute, value, html):
return unescapeHTML(res) return unescapeHTML(res)
def extract_attributes(attributes_str, attributes_regex=r'(?s)\s*([^\s=]+)\s*=\s*["\']([^"\']+)["\']'):
attributes = re.findall(attributes_regex, attributes_str)
attributes_dict = {}
if attributes:
attributes_dict = {attribute_name: attribute_value for (attribute_name, attribute_value) in attributes}
return attributes_dict
def clean_html(html): def clean_html(html):
"""Clean an HTML snippet into a readable string""" """Clean an HTML snippet into a readable string"""