1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
|
from .common import InfoExtractor
from ..utils import (
float_or_none,
int_or_none,
orderedSet,
parse_iso8601,
parse_qs,
parse_resolution,
str_or_none,
traverse_obj,
url_or_none,
)
class EpidemicSoundIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?epidemicsound\.com/track/(?P<id>[0-9a-zA-Z]+)'
_TESTS = [{
'url': 'https://www.epidemicsound.com/track/yFfQVRpSPz/',
'md5': 'd98ff2ddb49e8acab9716541cbc9dfac',
'info_dict': {
'id': '45014',
'display_id': 'yFfQVRpSPz',
'ext': 'mp3',
'title': 'Door Knock Door 1',
'alt_title': 'Door Knock Door 1',
'tags': ['foley', 'door', 'knock', 'glass', 'window', 'glass door knock'],
'categories': ['Misc. Door'],
'duration': 1,
'thumbnail': 'https://cdn.epidemicsound.com/curation-assets/commercial-release-cover-images/default-sfx/3000x3000.jpg',
'timestamp': 1415320353,
'upload_date': '20141107',
},
}, {
'url': 'https://www.epidemicsound.com/track/mj8GTTwsZd/',
'md5': 'c82b745890f9baf18dc2f8d568ee3830',
'info_dict': {
'id': '148700',
'display_id': 'mj8GTTwsZd',
'ext': 'mp3',
'title': 'Noplace',
'tags': ['liquid drum n bass', 'energetic'],
'categories': ['drum and bass'],
'duration': 237,
'timestamp': 1694426482,
'thumbnail': 'https://cdn.epidemicsound.com/curation-assets/commercial-release-cover-images/11138/3000x3000.jpg',
'upload_date': '20230911',
'release_timestamp': 1700535606,
'release_date': '20231121',
},
}]
@staticmethod
def _epidemic_parse_thumbnail(url: str):
if not url_or_none(url):
return None
return {
'url': url,
**(traverse_obj(url, ({parse_qs}, {
'width': ('width', 0, {int_or_none}),
'height': ('height', 0, {int_or_none}),
})) or parse_resolution(url)),
}
@staticmethod
def _epidemic_fmt_or_none(f):
if not f.get('format'):
f['format'] = f.get('format_id')
elif not f.get('format_id'):
f['format_id'] = f['format']
if not f['url'] or not f['format']:
return None
if f.get('format_note'):
f['format_note'] = f'track ID {f["format_note"]}'
if f['format'] != 'full':
f['preference'] = -2
return f
def _real_extract(self, url):
video_id = self._match_id(url)
json_data = self._download_json(f'https://www.epidemicsound.com/json/track/{video_id}', video_id)
thumbnails = traverse_obj(json_data, [('imageUrl', 'cover')])
thumb_base_url = traverse_obj(json_data, ('coverArt', 'baseUrl', {url_or_none}))
if thumb_base_url:
thumbnails.extend(traverse_obj(json_data, (
'coverArt', 'sizes', ..., {thumb_base_url.__add__})))
return traverse_obj(json_data, {
'id': ('id', {str_or_none}),
'display_id': ('publicSlug', {str}),
'title': ('title', {str}),
'alt_title': ('oldTitle', {str}),
'duration': ('length', {float_or_none}),
'timestamp': ('added', {parse_iso8601}),
'release_timestamp': ('releaseDate', {parse_iso8601}),
'categories': ('genres', ..., 'tag', {str}),
'tags': ('metadataTags', ..., {str}),
'age_limit': ('isExplicit', {lambda b: 18 if b else None}),
'thumbnails': ({lambda _: thumbnails}, {orderedSet}, ..., {self._epidemic_parse_thumbnail}),
'formats': ('stems', {dict.items}, ..., {
'format': (0, {str_or_none}),
'format_note': (1, 's3TrackId', {str_or_none}),
'format_id': (1, 'stemType', {str}),
'url': (1, 'lqMp3Url', {url_or_none}),
}, {self._epidemic_fmt_or_none}),
})
|