|
|
@ -1,3 +1,5 @@
|
|
|
|
|
|
|
|
import functools
|
|
|
|
|
|
|
|
import json
|
|
|
|
import random
|
|
|
|
import random
|
|
|
|
import re
|
|
|
|
import re
|
|
|
|
import time
|
|
|
|
import time
|
|
|
@ -6,7 +8,9 @@ from .common import InfoExtractor
|
|
|
|
from ..utils import (
|
|
|
|
from ..utils import (
|
|
|
|
KNOWN_EXTENSIONS,
|
|
|
|
KNOWN_EXTENSIONS,
|
|
|
|
ExtractorError,
|
|
|
|
ExtractorError,
|
|
|
|
|
|
|
|
extract_attributes,
|
|
|
|
float_or_none,
|
|
|
|
float_or_none,
|
|
|
|
|
|
|
|
get_element_html_by_id,
|
|
|
|
int_or_none,
|
|
|
|
int_or_none,
|
|
|
|
parse_filesize,
|
|
|
|
parse_filesize,
|
|
|
|
str_or_none,
|
|
|
|
str_or_none,
|
|
|
@ -17,6 +21,7 @@ from ..utils import (
|
|
|
|
url_or_none,
|
|
|
|
url_or_none,
|
|
|
|
urljoin,
|
|
|
|
urljoin,
|
|
|
|
)
|
|
|
|
)
|
|
|
|
|
|
|
|
from ..utils.traversal import traverse_obj
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class BandcampIE(InfoExtractor):
|
|
|
|
class BandcampIE(InfoExtractor):
|
|
|
@ -459,7 +464,7 @@ class BandcampUserIE(InfoExtractor):
|
|
|
|
},
|
|
|
|
},
|
|
|
|
}, {
|
|
|
|
}, {
|
|
|
|
'url': 'https://coldworldofficial.bandcamp.com/music',
|
|
|
|
'url': 'https://coldworldofficial.bandcamp.com/music',
|
|
|
|
'playlist_mincount': 10,
|
|
|
|
'playlist_mincount': 7,
|
|
|
|
'info_dict': {
|
|
|
|
'info_dict': {
|
|
|
|
'id': 'coldworldofficial',
|
|
|
|
'id': 'coldworldofficial',
|
|
|
|
'title': 'Discography of coldworldofficial',
|
|
|
|
'title': 'Discography of coldworldofficial',
|
|
|
@ -473,12 +478,19 @@ class BandcampUserIE(InfoExtractor):
|
|
|
|
},
|
|
|
|
},
|
|
|
|
}]
|
|
|
|
}]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def _yield_items(self, webpage):
|
|
|
|
|
|
|
|
yield from (
|
|
|
|
|
|
|
|
re.findall(r'<li data-item-id=["\'][^>]+>\s*<a href=["\'](?![^"\'/]*?/merch)([^"\']+)', webpage)
|
|
|
|
|
|
|
|
or re.findall(r'<div[^>]+trackTitle["\'][^"\']+["\']([^"\']+)', webpage))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
yield from traverse_obj(webpage, (
|
|
|
|
|
|
|
|
{functools.partial(get_element_html_by_id, 'music-grid')}, {extract_attributes},
|
|
|
|
|
|
|
|
'data-client-items', {json.loads}, ..., 'page_url', {str}))
|
|
|
|
|
|
|
|
|
|
|
|
def _real_extract(self, url):
|
|
|
|
def _real_extract(self, url):
|
|
|
|
uploader = self._match_id(url)
|
|
|
|
uploader = self._match_id(url)
|
|
|
|
webpage = self._download_webpage(url, uploader)
|
|
|
|
webpage = self._download_webpage(url, uploader)
|
|
|
|
|
|
|
|
|
|
|
|
discography_data = (re.findall(r'<li data-item-id=["\'][^>]+>\s*<a href=["\'](?![^"\'/]*?/merch)([^"\']+)', webpage)
|
|
|
|
|
|
|
|
or re.findall(r'<div[^>]+trackTitle["\'][^"\']+["\']([^"\']+)', webpage))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
return self.playlist_from_matches(
|
|
|
|
return self.playlist_from_matches(
|
|
|
|
discography_data, uploader, f'Discography of {uploader}', getter=lambda x: urljoin(url, x))
|
|
|
|
self._yield_items(webpage), uploader, f'Discography of {uploader}',
|
|
|
|
|
|
|
|
getter=functools.partial(urljoin, url))
|
|
|
|