[common] remove unwanted query params from unsigned akamai manifest URLs

This commit is contained in:
Remita Amine 2020-12-19 20:08:44 +01:00
parent 942b8ca3be
commit 22feed08a1
2 changed files with 9 additions and 3 deletions

View file

@ -2605,6 +2605,13 @@ class InfoExtractor(object):
return entries return entries
def _extract_akamai_formats(self, manifest_url, video_id, hosts={}): def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
signed = 'hdnea=' in manifest_url
if not signed:
# https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
manifest_url = re.sub(
r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
'', manifest_url).strip('?')
formats = [] formats = []
hdcore_sign = 'hdcore=3.7.0' hdcore_sign = 'hdcore=3.7.0'
@ -2630,7 +2637,7 @@ class InfoExtractor(object):
formats.extend(m3u8_formats) formats.extend(m3u8_formats)
http_host = hosts.get('http') http_host = hosts.get('http')
if http_host and m3u8_formats and 'hdnea=' not in m3u8_url: if http_host and m3u8_formats and not signed:
REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+' REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',') qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
qualities_length = len(qualities) qualities_length = len(qualities)

View file

@ -33,8 +33,7 @@ class NRKBaseIE(InfoExtractor):
def _extract_nrk_formats(self, asset_url, video_id): def _extract_nrk_formats(self, asset_url, video_id):
if re.match(r'https?://[^/]+\.akamaihd\.net/i/', asset_url): if re.match(r'https?://[^/]+\.akamaihd\.net/i/', asset_url):
return self._extract_akamai_formats( return self._extract_akamai_formats(asset_url, video_id)
re.sub(r'(?:b=\d+-\d+|__a__=off)&?', '', asset_url), video_id)
asset_url = re.sub(r'(?:bw_(?:low|high)=\d+|no_audio_only)&?', '', asset_url) asset_url = re.sub(r'(?:bw_(?:low|high)=\d+|no_audio_only)&?', '', asset_url)
formats = self._extract_m3u8_formats( formats = self._extract_m3u8_formats(
asset_url, video_id, 'mp4', 'm3u8_native', fatal=False) asset_url, video_id, 'mp4', 'm3u8_native', fatal=False)