[extractor/lbry] Extract original quality formats (#7257)

Closes #7251
Authored by: bashonly

yt_dlp/extractor/lbry.py

@@ -1,8 +1,8 @@
 import functools
 import json
+import urllib.parse
 
 from .common import InfoExtractor
-from ..compat import compat_str, compat_urllib_parse_unquote
 from ..utils import (
     ExtractorError,
     HEADRequest,
@@ -12,7 +12,10 @@ from ..utils import (
     int_or_none,
     mimetype2ext,
     parse_qs,
+    traverse_obj,
     try_get,
+    url_or_none,
+    urlhandle_detect_ext,
     urljoin,
 )
 
@@ -52,38 +55,25 @@ class LBRYBaseIE(InfoExtractor):
             '/%s:%s' % (claim_name, claim_id))
 
     def _parse_stream(self, stream, url):
-        stream_value = stream.get('value') or {}
+        stream_type = traverse_obj(stream, ('value', 'stream_type', {str}))
-        stream_type = stream_value.get('stream_type')
+
-        source = stream_value.get('source') or {}
+        info = traverse_obj(stream, {
-        media = stream_value.get(stream_type) or {}
+            'title': ('value', 'title', {str}),
-        signing_channel = stream.get('signing_channel') or {}
+            'thumbnail': ('value', 'thumbnail', 'url', {url_or_none}),
-        channel_name = signing_channel.get('name')
+            'description': ('value', 'description', {str}),
-        channel_claim_id = signing_channel.get('claim_id')
+            'license': ('value', 'license', {str}),
-        channel_url = None
+            'timestamp': ('timestamp', {int_or_none}),
-        if channel_name and channel_claim_id:
+            'release_timestamp': ('value', 'release_time', {int_or_none}),
-            channel_url = self._permanent_url(url, channel_name, channel_claim_id)
+            'tags': ('value', 'tags', ..., {lambda x: x or None}),
+            'duration': ('value', stream_type, 'duration', {int_or_none}),
+            'channel': ('signing_channel', 'value', 'title', {str}),
+            'channel_id': ('signing_channel', 'claim_id', {str}),
+        })
+
+        channel_name = traverse_obj(stream, ('signing_channel', 'name', {str}))
+        if channel_name and info.get('channel_id'):
+            info['channel_url'] = self._permanent_url(url, channel_name, info['channel_id'])
 
-        info = {
-            'thumbnail': try_get(stream_value, lambda x: x['thumbnail']['url'], compat_str),
-            'description': stream_value.get('description'),
-            'license': stream_value.get('license'),
-            'timestamp': int_or_none(stream.get('timestamp')),
-            'release_timestamp': int_or_none(stream_value.get('release_time')),
-            'tags': stream_value.get('tags'),
-            'duration': int_or_none(media.get('duration')),
-            'channel': try_get(signing_channel, lambda x: x['value']['title']),
-            'channel_id': channel_claim_id,
-            'channel_url': channel_url,
-            'ext': determine_ext(source.get('name')) or mimetype2ext(source.get('media_type')),
-            'filesize': int_or_none(source.get('size')),
-        }
-        if stream_type == 'audio':
-            info['vcodec'] = 'none'
-        else:
-            info.update({
-                'width': int_or_none(media.get('width')),
-                'height': int_or_none(media.get('height')),
-            })
         return info
 
 
@@ -186,6 +176,28 @@ class LBRYIE(LBRYBaseIE):
             'license': 'None',
         },
         'params': {'skip_download': True}
+    }, {
+        # original quality format w/higher resolution than HLS formats
+        'url': 'https://odysee.com/@wickedtruths:2/Biotechnological-Invasion-of-Skin-(April-2023):4',
+        'md5': '305b0b3b369bde1b984961f005b67193',
+        'info_dict': {
+            'id': '41fbfe805eb73c8d3012c0c49faa0f563274f634',
+            'ext': 'mp4',
+            'title': 'Biotechnological Invasion of Skin (April 2023)',
+            'description': 'md5:709a2f4c07bd8891cda3a7cc2d6fcf5c',
+            'channel': 'Wicked Truths',
+            'channel_id': '23d2bbf856b0ceed5b1d7c5960bcc72da5a20cb0',
+            'channel_url': 'https://odysee.com/@wickedtruths:23d2bbf856b0ceed5b1d7c5960bcc72da5a20cb0',
+            'timestamp': 1685790036,
+            'upload_date': '20230603',
+            'release_timestamp': 1685617473,
+            'release_date': '20230601',
+            'duration': 1063,
+            'thumbnail': 'https://thumbs.odycdn.com/4e6d39da4df0cfdad45f64e253a15959.webp',
+            'tags': ['smart skin surveillance', 'biotechnology invasion of skin', 'morgellons'],
+            'license': 'None',
+            'protocol': 'https',  # test for direct mp4 download
+        },
     }, {
         'url': 'https://odysee.com/@BrodieRobertson:5/apple-is-tracking-everything-you-do-on:e',
         'only_matching': True,
@@ -221,41 +233,64 @@ class LBRYIE(LBRYBaseIE):
             display_id = display_id.split('/', 2)[-1].replace('/', ':')
         else:
             display_id = display_id.replace(':', '#')
-        display_id = compat_urllib_parse_unquote(display_id)
+        display_id = urllib.parse.unquote(display_id)
         uri = 'lbry://' + display_id
         result = self._resolve_url(uri, display_id, 'stream')
         headers = {'Referer': 'https://odysee.com/'}
-        if result['value'].get('stream_type') in self._SUPPORTED_STREAM_TYPES:
+
+        formats = []
+        stream_type = traverse_obj(result, ('value', 'stream_type', {str}))
+
+        if stream_type in self._SUPPORTED_STREAM_TYPES:
             claim_id, is_live = result['claim_id'], False
             streaming_url = self._call_api_proxy(
                 'get', claim_id, {'uri': uri}, 'streaming url')['streaming_url']
+
+            # GET request returns original video/audio file if available
+            ext = urlhandle_detect_ext(self._request_webpage(
+                streaming_url, display_id, 'Checking for original quality', headers=headers))
+            if ext != 'm3u8':
+                formats.append({
+                    'url': streaming_url,
+                    'format_id': 'original',
+                    'quality': 1,
+                    **traverse_obj(result, ('value', {
+                        'ext': ('source', (('name', {determine_ext}), ('media_type', {mimetype2ext}))),
+                        'filesize': ('source', 'size', {int_or_none}),
+                        'width': ('video', 'width', {int_or_none}),
+                        'height': ('video', 'height', {int_or_none}),
+                    }), get_all=False),
+                    'vcodec': 'none' if stream_type == 'audio' else None,
+                })
+
+            # HEAD request returns redirect response to m3u8 URL if available
             final_url = self._request_webpage(
                 HEADRequest(streaming_url), display_id, headers=headers,
                 note='Downloading streaming redirect url info').geturl()
+
         elif result.get('value_type') == 'stream':
             claim_id, is_live = result['signing_channel']['claim_id'], True
             live_data = self._download_json(
                 'https://api.odysee.live/livestream/is_live', claim_id,
                 query={'channel_claim_id': claim_id},
                 note='Downloading livestream JSON metadata')['data']
-            streaming_url = final_url = live_data.get('VideoURL')
+            final_url = live_data.get('VideoURL')
             # Upcoming videos may still give VideoURL
             if not live_data.get('Live'):
-                streaming_url = final_url = None
+                final_url = None
                 self.raise_no_formats('This stream is not live', True, claim_id)
+
         else:
             raise UnsupportedError(url)
 
-        info = self._parse_stream(result, url)
         if determine_ext(final_url) == 'm3u8':
-            info['formats'] = self._extract_m3u8_formats(
+            formats.extend(self._extract_m3u8_formats(
-                final_url, display_id, 'mp4', 'm3u8_native', m3u8_id='hls', live=is_live, headers=headers)
+                final_url, display_id, 'mp4', m3u8_id='hls', live=is_live, headers=headers))
-        else:
+
-            info['url'] = streaming_url
         return {
-            **info,
+            **self._parse_stream(result, url),
             'id': claim_id,
-            'title': result['value']['title'],
+            'formats': formats,
             'is_live': is_live,
             'http_headers': headers,
         }
@@ -299,14 +334,12 @@ class LBRYChannelIE(LBRYBaseIE):
             if not (stream_claim_name and stream_claim_id):
                 continue
 
-            info = self._parse_stream(item, url)
+            yield {
-            info.update({
+                **self._parse_stream(item, url),
                 '_type': 'url',
                 'id': stream_claim_id,
-                'title': try_get(item, lambda x: x['value']['title']),
                 'url': self._permanent_url(url, stream_claim_name, stream_claim_id),
-            })
+            }
-            yield info
 
     def _real_extract(self, url):
         display_id = self._match_id(url).replace(':', '#')