From 35779eda7a2ad52f2bbbe2e1251ad8ddb4bac7d1 Mon Sep 17 00:00:00 2001 From: vallovic Date: Fri, 19 Feb 2021 14:10:10 +0000 Subject: [PATCH 1/8] Dealing with RTP latest changes --- youtube_dl/extractor/rtp.py | 75 +++++++++++++++++++++++++++---------- 1 file changed, 55 insertions(+), 20 deletions(-) diff --git a/youtube_dl/extractor/rtp.py b/youtube_dl/extractor/rtp.py index 02986f442..665c86f14 100644 --- a/youtube_dl/extractor/rtp.py +++ b/youtube_dl/extractor/rtp.py @@ -6,22 +6,27 @@ from ..utils import ( determine_ext, js_to_json, ) +from ..compat import ( + compat_b64decode, + compat_urllib_parse_unquote, +) + +import re class RTPIE(InfoExtractor): - _VALID_URL = r'https?://(?:www\.)?rtp\.pt/play/p(?P[0-9]+)/(?P[^/?#]+)/?' + _VALID_URL = r'https?://(?:www\.)?rtp\.pt/play/(.*\/)?p(?P[0-9]+)/(?P[^/?#]+)/?' _TESTS = [{ - 'url': 'http://www.rtp.pt/play/p405/e174042/paixoes-cruzadas', - 'md5': 'e736ce0c665e459ddb818546220b4ef8', + 'url': 'https://www.rtp.pt/play/p117/e476527/os-contemporaneos', 'info_dict': { - 'id': 'e174042', - 'ext': 'mp3', - 'title': 'Paixões Cruzadas', - 'description': 'As paixões musicais de António Cartaxo e António Macedo', + 'id': 'e476527', + 'ext': 'mp4', + 'title': 'Os Contemporâneos Episódio 1 - RTP Play - RTP', + 'description': 'Os Contemporâneos, um programa de humor com um olhar na sociedade portuguesa!', 'thumbnail': r're:^https?://.*\.jpg', }, }, { - 'url': 'http://www.rtp.pt/play/p831/a-quimica-das-coisas', + 'url': 'https://www.rtp.pt/play/p510/aleixo-fm', 'only_matching': True, }] @@ -29,30 +34,60 @@ class RTPIE(InfoExtractor): video_id = self._match_id(url) webpage = self._download_webpage(url, video_id) - title = self._html_search_meta( - 'twitter:title', webpage, display_name='title', fatal=True) + title = self._html_search_regex(r'(.+?)', webpage, 'title') + + # Get JS object + js_object = self._search_regex(r'(?s)RTPPlayer *\( *({.+?}) *\);', webpage, 'player config') + + json_string_for_config = '' + + # Verify JS object since it isn't pure JSON and maybe it needs some decodings + for line in js_object.splitlines(): + stripped_line = line.strip() + + # If JS object key is 'file' + if re.match('file ?:', stripped_line): + if 'decodeURIComponent' in stripped_line: + # 1) The file URL is inside object and with HLS encoded... + hls_encoded = re.match(r"[^[]*\[([^]]*)\]", stripped_line).groups()[0] + hls_encoded = hls_encoded.replace('"', '').replace('\'', '').replace(',', '') + decoded_file_url = compat_b64decode( + compat_urllib_parse_unquote( + hls_encoded.replace('"', '').replace(',', ''))).decode('utf-8') + + # Insert the decoded HLS file URL into pure JSON string + json_string_for_config += '\nfile: "' + decoded_file_url + '",' + else: + # 2) ... or it's a direct M3U8 file + json_string_for_config += '\n' + line + + elif not stripped_line.startswith("//") and not re.match('fileKey ?:', stripped_line): + # Ignore commented lines and 'fileKey' entry since it is no longer supported by RTP + json_string_for_config += '\n' + line + + # Finally send pure JSON string for JSON parsing + config = self._parse_json(json_string_for_config, video_id, js_to_json) + + # config = self._parse_json(self._search_regex( + # r'(?s)RTPPlayer ?\( ?({.+?})\);', webpage, + # 'player config'), video_id, js_to_json) - config = self._parse_json(self._search_regex( - r'(?s)RTPPlayer\(({.+?})\);', webpage, - 'player config'), video_id, js_to_json) file_url = config['file'] ext = determine_ext(file_url) + if ext == 'm3u8': - file_key = config.get('fileKey') + # Download via m3u8 file formats = self._extract_m3u8_formats( file_url, video_id, 'mp4', 'm3u8_native', - m3u8_id='hls', fatal=file_key) - if file_key: - formats.append({ - 'url': 'https://cdn-ondemand.rtp.pt' + file_key, - 'preference': 1, - }) + m3u8_id='hls') + self._sort_formats(formats) else: formats = [{ 'url': file_url, 'ext': ext, }] + if config.get('mediaType') == 'audio': for f in formats: f['vcodec'] = 'none' From a85625977d39b4ad5bb9b399773700fa20db7587 Mon Sep 17 00:00:00 2001 From: vallovic Date: Sat, 20 Feb 2021 20:42:33 +0000 Subject: [PATCH 2/8] 'Estudo em Casa' wasn't working since RTP has a lot of ways of dealing with their code --- youtube_dl/extractor/rtp.py | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/youtube_dl/extractor/rtp.py b/youtube_dl/extractor/rtp.py index 665c86f14..02716c172 100644 --- a/youtube_dl/extractor/rtp.py +++ b/youtube_dl/extractor/rtp.py @@ -58,7 +58,7 @@ class RTPIE(InfoExtractor): # Insert the decoded HLS file URL into pure JSON string json_string_for_config += '\nfile: "' + decoded_file_url + '",' else: - # 2) ... or it's a direct M3U8 file + # 2) ... or the file URL is not encoded so keep it that way json_string_for_config += '\n' + line elif not stripped_line.startswith("//") and not re.match('fileKey ?:', stripped_line): @@ -68,11 +68,12 @@ class RTPIE(InfoExtractor): # Finally send pure JSON string for JSON parsing config = self._parse_json(json_string_for_config, video_id, js_to_json) - # config = self._parse_json(self._search_regex( - # r'(?s)RTPPlayer ?\( ?({.+?})\);', webpage, - # 'player config'), video_id, js_to_json) + # Check if file URL is directly a string or is still inside object + if isinstance(config['file'], str): + file_url = config['file'] + else: + file_url = config['file']['hls'] - file_url = config['file'] ext = determine_ext(file_url) if ext == 'm3u8': From b955e0a5dc208abaa5f7d64e7228c46bb8a3af63 Mon Sep 17 00:00:00 2001 From: vallovic Date: Sun, 21 Feb 2021 14:05:42 +0000 Subject: [PATCH 3/8] RTP devs like to try out different approaches --- youtube_dl/extractor/rtp.py | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/youtube_dl/extractor/rtp.py b/youtube_dl/extractor/rtp.py index 02716c172..5ba26268e 100644 --- a/youtube_dl/extractor/rtp.py +++ b/youtube_dl/extractor/rtp.py @@ -15,11 +15,11 @@ import re class RTPIE(InfoExtractor): - _VALID_URL = r'https?://(?:www\.)?rtp\.pt/play/(.*\/)?p(?P[0-9]+)/(?P[^/?#]+)/?' + _VALID_URL = r'https?://(?:(?:(?:www\.)?rtp\.pt/play/(?P.*/)?p(?P[0-9]+)/(?Pe[0-9]+/)?)|(?:arquivos\.rtp\.pt/conteudos/))(?P[^/?#]+)/?' _TESTS = [{ 'url': 'https://www.rtp.pt/play/p117/e476527/os-contemporaneos', 'info_dict': { - 'id': 'e476527', + 'id': 'os-contemporaneos', 'ext': 'mp4', 'title': 'Os Contemporâneos Episódio 1 - RTP Play - RTP', 'description': 'Os Contemporâneos, um programa de humor com um olhar na sociedade portuguesa!', @@ -51,9 +51,12 @@ class RTPIE(InfoExtractor): # 1) The file URL is inside object and with HLS encoded... hls_encoded = re.match(r"[^[]*\[([^]]*)\]", stripped_line).groups()[0] hls_encoded = hls_encoded.replace('"', '').replace('\'', '').replace(',', '') - decoded_file_url = compat_b64decode( - compat_urllib_parse_unquote( - hls_encoded.replace('"', '').replace(',', ''))).decode('utf-8') + if 'atob' in stripped_line: + decoded_file_url = compat_b64decode( + compat_urllib_parse_unquote( + hls_encoded.replace('"', '').replace(',', ''))).decode('utf-8') + else: + decoded_file_url = compat_urllib_parse_unquote(hls_encoded) # Insert the decoded HLS file URL into pure JSON string json_string_for_config += '\nfile: "' + decoded_file_url + '",' From 7f40887b29f97d4729e344b04e76456f02c0c33a Mon Sep 17 00:00:00 2001 From: vallovic Date: Tue, 13 Apr 2021 21:03:59 +0100 Subject: [PATCH 4/8] Ignore more comments in RTP source code --- youtube_dl/extractor/rtp.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/youtube_dl/extractor/rtp.py b/youtube_dl/extractor/rtp.py index 5ba26268e..1c487749b 100644 --- a/youtube_dl/extractor/rtp.py +++ b/youtube_dl/extractor/rtp.py @@ -64,8 +64,8 @@ class RTPIE(InfoExtractor): # 2) ... or the file URL is not encoded so keep it that way json_string_for_config += '\n' + line - elif not stripped_line.startswith("//") and not re.match('fileKey ?:', stripped_line): - # Ignore commented lines and 'fileKey' entry since it is no longer supported by RTP + elif not stripped_line.startswith("//") and not re.match('fileKey ?:', stripped_line) and not re.match('.*extraSettings ?:', stripped_line): + # Ignore commented lines, 'fileKey' entry since it is no longer supported by RTP and also 'extraSettings' json_string_for_config += '\n' + line # Finally send pure JSON string for JSON parsing From ff47d11269d6cffb36c8b86cf8f5e8effa881670 Mon Sep 17 00:00:00 2001 From: vallovic Date: Wed, 21 Apr 2021 16:24:26 +0100 Subject: [PATCH 5/8] Ignore EVEN more comments in RTP source code --- youtube_dl/extractor/rtp.py | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/youtube_dl/extractor/rtp.py b/youtube_dl/extractor/rtp.py index 1c487749b..2f9360e12 100644 --- a/youtube_dl/extractor/rtp.py +++ b/youtube_dl/extractor/rtp.py @@ -34,6 +34,10 @@ class RTPIE(InfoExtractor): video_id = self._match_id(url) webpage = self._download_webpage(url, video_id) + + # Remove JS multi-line comments from webpage source + webpage = re.sub(r'(\/\*.*\*\/)', '', webpage, flags=re.DOTALL) + title = self._html_search_regex(r'(.+?)', webpage, 'title') # Get JS object From d30180d4f3e1c0f1f01f7a3aa3d3fa250ce23b15 Mon Sep 17 00:00:00 2001 From: vallovic Date: Fri, 23 Apr 2021 23:02:04 +0100 Subject: [PATCH 6/8] Consider multi-part videos in filename output --- youtube_dl/extractor/rtp.py | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/youtube_dl/extractor/rtp.py b/youtube_dl/extractor/rtp.py index 2f9360e12..da30a7789 100644 --- a/youtube_dl/extractor/rtp.py +++ b/youtube_dl/extractor/rtp.py @@ -40,6 +40,14 @@ class RTPIE(InfoExtractor): title = self._html_search_regex(r'(.+?)', webpage, 'title') + # Replace irrelevant text in title + title = title.replace(' - RTP Play - RTP', '') + + # Check if it's a video split in parts, if so add part number to title + part = self._html_search_regex(r'section\-parts.*(.+?).*', webpage, 'part', default=None) + if part: + title = f'{title} {part}' + # Get JS object js_object = self._search_regex(r'(?s)RTPPlayer *\( *({.+?}) *\);', webpage, 'player config') From 5a7243b741f2f9a746b0955de8bae2f558efd074 Mon Sep 17 00:00:00 2001 From: vallovic Date: Thu, 6 May 2021 23:20:20 +0100 Subject: [PATCH 7/8] Working around new URLs --- youtube_dl/extractor/rtp.py | 3 +++ 1 file changed, 3 insertions(+) diff --git a/youtube_dl/extractor/rtp.py b/youtube_dl/extractor/rtp.py index da30a7789..d421f2f1a 100644 --- a/youtube_dl/extractor/rtp.py +++ b/youtube_dl/extractor/rtp.py @@ -70,6 +70,9 @@ class RTPIE(InfoExtractor): else: decoded_file_url = compat_urllib_parse_unquote(hls_encoded) + # Workaround for new behaviour + decoded_file_url = decoded_file_url.replace('streaming-vod.rtp.pt/hls/', 'streaming-ondemand.rtp.pt/').replace('.mp4/', '/') + # Insert the decoded HLS file URL into pure JSON string json_string_for_config += '\nfile: "' + decoded_file_url + '",' else: From 526ef898bf1bd769f0a0f93d36312544ab0d2e1e Mon Sep 17 00:00:00 2001 From: Pedro Ferreira Date: Thu, 19 Aug 2021 14:56:53 +0200 Subject: [PATCH 8/8] Fix once again RTP downloads --- youtube_dl/extractor/rtp.py | 76 ++++++++++++++++--------------------- 1 file changed, 33 insertions(+), 43 deletions(-) diff --git a/youtube_dl/extractor/rtp.py b/youtube_dl/extractor/rtp.py index d421f2f1a..2e6329755 100644 --- a/youtube_dl/extractor/rtp.py +++ b/youtube_dl/extractor/rtp.py @@ -3,6 +3,7 @@ from __future__ import unicode_literals from .common import InfoExtractor from ..utils import ( + ExtractorError, determine_ext, js_to_json, ) @@ -14,15 +15,22 @@ from ..compat import ( import re +def decode_b64_url(code): + decoded_url = re.match(r"[^[]*\[([^]]*)\]", code).groups()[0] + return compat_b64decode( + compat_urllib_parse_unquote( + decoded_url.replace('"', '').replace('\'', '').replace(',', ''))).decode('utf-8') + + class RTPIE(InfoExtractor): _VALID_URL = r'https?://(?:(?:(?:www\.)?rtp\.pt/play/(?P.*/)?p(?P[0-9]+)/(?Pe[0-9]+/)?)|(?:arquivos\.rtp\.pt/conteudos/))(?P[^/?#]+)/?' _TESTS = [{ - 'url': 'https://www.rtp.pt/play/p117/e476527/os-contemporaneos', + 'url': 'https://www.rtp.pt/play/p9165/e562949/por-do-sol', 'info_dict': { - 'id': 'os-contemporaneos', + 'id': 'por-do-sol', 'ext': 'mp4', - 'title': 'Os Contemporâneos Episódio 1 - RTP Play - RTP', - 'description': 'Os Contemporâneos, um programa de humor com um olhar na sociedade portuguesa!', + 'title': 'Pôr do Sol Episódio 1 - de 16 Ago 2021', + 'description': 'Madalena Bourbon de Linhaça vive atormentada pelo segredo que esconde desde 1990. Matilde Bourbon de Linhaça sonha fugir com o seu amor proibido. O en', 'thumbnail': r're:^https?://.*\.jpg', }, }, { @@ -50,60 +58,42 @@ class RTPIE(InfoExtractor): # Get JS object js_object = self._search_regex(r'(?s)RTPPlayer *\( *({.+?}) *\);', webpage, 'player config') - json_string_for_config = '' + full_url = None - # Verify JS object since it isn't pure JSON and maybe it needs some decodings + # Verify JS object since it isn't pure JSON and maybe it needs some tuning for line in js_object.splitlines(): stripped_line = line.strip() - # If JS object key is 'file' - if re.match('file ?:', stripped_line): - if 'decodeURIComponent' in stripped_line: - # 1) The file URL is inside object and with HLS encoded... - hls_encoded = re.match(r"[^[]*\[([^]]*)\]", stripped_line).groups()[0] - hls_encoded = hls_encoded.replace('"', '').replace('\'', '').replace(',', '') - if 'atob' in stripped_line: - decoded_file_url = compat_b64decode( - compat_urllib_parse_unquote( - hls_encoded.replace('"', '').replace(',', ''))).decode('utf-8') - else: - decoded_file_url = compat_urllib_parse_unquote(hls_encoded) - - # Workaround for new behaviour - decoded_file_url = decoded_file_url.replace('streaming-vod.rtp.pt/hls/', 'streaming-ondemand.rtp.pt/').replace('.mp4/', '/') - - # Insert the decoded HLS file URL into pure JSON string - json_string_for_config += '\nfile: "' + decoded_file_url + '",' + # key == 'fileKey', then we found what we wanted + if re.match(r'fileKey:', stripped_line): + if re.match(r'fileKey: *""', stripped_line): + raise ExtractorError("Episode not found (probably removed)", expected=True) + url = decode_b64_url(stripped_line) + if 'mp3' in url: + full_url = 'https://cdn-ondemand.rtp.pt' + url else: - # 2) ... or the file URL is not encoded so keep it that way - json_string_for_config += '\n' + line + full_url = 'https://streaming-vod.rtp.pt/dash{}/manifest.mpd'.format(url) - elif not stripped_line.startswith("//") and not re.match('fileKey ?:', stripped_line) and not re.match('.*extraSettings ?:', stripped_line): - # Ignore commented lines, 'fileKey' entry since it is no longer supported by RTP and also 'extraSettings' + elif not stripped_line.startswith("//") and not re.match('file *:', stripped_line) and not re.match('.*extraSettings ?:', stripped_line): + # Ignore commented lines, `extraSettings` and `f`. The latter seems to some random unrelated video. json_string_for_config += '\n' + line + if not full_url: + raise ExtractorError("No valid media source found in page") + # Finally send pure JSON string for JSON parsing config = self._parse_json(json_string_for_config, video_id, js_to_json) + full_url = full_url.replace('drm-dash', 'dash') + ext = determine_ext(full_url) - # Check if file URL is directly a string or is still inside object - if isinstance(config['file'], str): - file_url = config['file'] - else: - file_url = config['file']['hls'] - - ext = determine_ext(file_url) - - if ext == 'm3u8': - # Download via m3u8 file - formats = self._extract_m3u8_formats( - file_url, video_id, 'mp4', 'm3u8_native', - m3u8_id='hls') - + if ext == 'mpd': + # Download via mpd file + formats = self._extract_mpd_formats(full_url, video_id) self._sort_formats(formats) else: formats = [{ - 'url': file_url, + 'url': full_url, 'ext': ext, }]