[mixcloud] Simplify url extraction

On the tracks I tested the server number in the url from the webpage is valid
for the mp3 or the m4a file and any other number is invalid, it's a
waste of time to check them.
This commit is contained in:
Jaime Marquínez Ferrándiz 2015-04-17 19:02:49 +02:00
parent d8e7ef04dc
commit c5826a491b

View file

@ -1,7 +1,6 @@
from __future__ import unicode_literals from __future__ import unicode_literals
import re import re
import itertools
from .common import InfoExtractor from .common import InfoExtractor
from ..compat import ( from ..compat import (
@ -46,20 +45,16 @@ class MixcloudIE(InfoExtractor):
}, },
}] }]
def _get_url(self, track_id, template_url, server_number): def _check_url(self, url, track_id, ext):
boundaries = (1, 30)
for nr in server_numbers(server_number, boundaries):
url = template_url % nr
try: try:
# We only want to know if the request succeed # We only want to know if the request succeed
# don't download the whole file # don't download the whole file
self._request_webpage( self._request_webpage(
HEADRequest(url), track_id, HEADRequest(url), track_id,
'Checking URL %d/%d ...' % (nr, boundaries[-1])) 'Trying %s URL' % ext)
return url return True
except ExtractorError: except ExtractorError:
pass return False
return None
def _real_extract(self, url): def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url) mobj = re.match(self._VALID_URL, url)
@ -72,14 +67,9 @@ class MixcloudIE(InfoExtractor):
preview_url = self._search_regex( preview_url = self._search_regex(
r'\s(?:data-preview-url|m-preview)="([^"]+)"', webpage, 'preview url') r'\s(?:data-preview-url|m-preview)="([^"]+)"', webpage, 'preview url')
song_url = preview_url.replace('/previews/', '/c/originals/') song_url = preview_url.replace('/previews/', '/c/originals/')
server_number = int(self._search_regex(r'stream(\d+)', song_url, 'server number')) if not self._check_url(song_url, track_id, 'mp3'):
template_url = re.sub(r'(stream\d*)', 'stream%d', song_url) song_url = song_url.replace('.mp3', '.m4a').replace('originals/', 'm4a/64/')
final_song_url = self._get_url(track_id, template_url, server_number) if not self._check_url(song_url, track_id, 'm4a'):
if final_song_url is None:
self.to_screen('Trying with m4a extension')
template_url = template_url.replace('.mp3', '.m4a').replace('originals/', 'm4a/64/')
final_song_url = self._get_url(track_id, template_url, server_number)
if final_song_url is None:
raise ExtractorError('Unable to extract track url') raise ExtractorError('Unable to extract track url')
PREFIX = ( PREFIX = (
@ -107,7 +97,7 @@ class MixcloudIE(InfoExtractor):
return { return {
'id': track_id, 'id': track_id,
'title': title, 'title': title,
'url': final_song_url, 'url': song_url,
'description': description, 'description': description,
'thumbnail': thumbnail, 'thumbnail': thumbnail,
'uploader': uploader, 'uploader': uploader,
@ -115,35 +105,3 @@ class MixcloudIE(InfoExtractor):
'view_count': view_count, 'view_count': view_count,
'like_count': like_count, 'like_count': like_count,
} }
def server_numbers(first, boundaries):
""" Server numbers to try in descending order of probable availability.
Starting from first (i.e. the number of the server hosting the preview file)
and going further and further up to the higher boundary and down to the
lower one in an alternating fashion. Namely:
server_numbers(2, (1, 5))
# Where the preview server is 2, min number is 1 and max is 5.
# Yields: 2, 3, 1, 4, 5
Why not random numbers or increasing sequences? Since from what I've seen,
full length files seem to be hosted on servers whose number is closer to
that of the preview; to be confirmed.
"""
zip_longest = getattr(itertools, 'zip_longest', None)
if zip_longest is None:
# python 2.x
zip_longest = itertools.izip_longest
if len(boundaries) != 2:
raise ValueError("boundaries should be a two-element tuple")
min, max = boundaries
highs = range(first + 1, max + 1)
lows = range(first - 1, min - 1, -1)
rest = filter(
None, itertools.chain.from_iterable(zip_longest(highs, lows)))
yield first
for n in rest:
yield n