From 01824d275bfa7efbaca274b38c1ddc2b03f12f5d Mon Sep 17 00:00:00 2001 From: dirkf Date: Wed, 19 Jan 2022 13:24:33 +0000 Subject: [PATCH] Additional tweaks: allow any .ndr.de, simplify quote match --- youtube_dl/extractor/ndr.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/youtube_dl/extractor/ndr.py b/youtube_dl/extractor/ndr.py index 1996d4f96..26627f8b0 100644 --- a/youtube_dl/extractor/ndr.py +++ b/youtube_dl/extractor/ndr.py @@ -28,7 +28,7 @@ class NDRBaseIE(InfoExtractor): class NDRIE(NDRBaseIE): IE_NAME = 'ndr' IE_DESC = 'NDR.de - Norddeutscher Rundfunk' - _VALID_URL = r'https?://(?:\w+\.)?ndr\.de/(?:[^/]+/)*(?P[^/?#]+),[\da-z]+\.html' + _VALID_URL = r'https?://(?:\w+\.)*ndr\.de/(?:[^/]+/)*(?P[^/?#]+),[\da-z]+\.html' _TESTS = [{ # httpVideo, same content id 'url': 'http://www.ndr.de/fernsehen/Party-Poette-und-Parade,hafengeburtstag988.html', @@ -202,7 +202,7 @@ class NJoyIE(NDRBaseIE): def _extract_embed(self, webpage, display_id, url=None): # find tell-tale URL with the actual ID, or ... video_id = self._search_regex( - (r'''\bsrc\s*=\s*(?:"|')?(?:/\w+)+/([a-z]+\d+)(?!\.)\b''', + (r'''\bsrc\s*=\s*["']?(?:/\w+)+/([a-z]+\d+)(?!\.)\b''', r']+id="pp_([\da-z]+)"', ), webpage, 'NDR id', default=None) @@ -322,7 +322,7 @@ class NDREmbedBaseIE(InfoExtractor): class NDREmbedIE(NDREmbedBaseIE): IE_NAME = 'ndr:embed' - _VALID_URL = r'https?://(?:www\.)?ndr\.de/(?:[^/]+/)*(?P[\da-z]+)-(?:(?:ard)?player|externalPlayer)\.html' + _VALID_URL = r'https?://(?:\w+\.)*ndr\.de/(?:[^/]+/)*(?P[\da-z]+)-(?:(?:ard)?player|externalPlayer)\.html' _TESTS = [{ 'url': 'http://www.ndr.de/fernsehen/sendungen/ndr_aktuell/ndraktuell28488-player.html', 'md5': '8b9306142fe65bbdefb5ce24edb6b0a9',