Update youtubeExtractor.ts

This commit is contained in:
CK 2026-03-05 20:20:44 +05:30 committed by GitHub
parent 25de499931
commit da39ca47bc
No known key found for this signature in database
GPG key ID: B5690EEEBB952194

View file

@ -18,26 +18,15 @@ interface InnertubeFormat {
qualityLabel?: string; qualityLabel?: string;
audioQuality?: string; audioQuality?: string;
audioSampleRate?: string; audioSampleRate?: string;
audioChannels?: number;
approxDurationMs?: string;
initRange?: { start: string; end: string }; initRange?: { start: string; end: string };
indexRange?: { start: string; end: string }; indexRange?: { start: string; end: string };
} }
interface StreamingData {
formats?: InnertubeFormat[];
adaptiveFormats?: InnertubeFormat[];
hlsManifestUrl?: string;
expiresInSeconds?: string;
}
interface PlayerResponse { interface PlayerResponse {
streamingData?: StreamingData; streamingData?: {
videoDetails?: { formats?: InnertubeFormat[];
videoId?: string; adaptiveFormats?: InnertubeFormat[];
title?: string; hlsManifestUrl?: string;
lengthSeconds?: string;
isLive?: boolean;
}; };
playabilityStatus?: { playabilityStatus?: {
status?: string; status?: string;
@ -53,12 +42,9 @@ interface StreamCandidate {
height: number; height: number;
fps: number; fps: number;
ext: 'mp4' | 'webm' | 'm4a' | 'other'; ext: 'mp4' | 'webm' | 'm4a' | 'other';
initRange?: { start: string; end: string };
indexRange?: { start: string; end: string };
bitrate: number; bitrate: number;
audioSampleRate?: string; audioSampleRate?: string;
mimeType: string; mimeType: string;
itag?: number;
} }
interface HlsVariant { interface HlsVariant {
@ -68,25 +54,32 @@ interface HlsVariant {
bandwidth: number; bandwidth: number;
} }
export interface TrailerPlaybackSource {
videoUrl: string; // best video (may be muxed, DASH video-only, HLS, or progressive)
audioUrl: string | null; // separate audio URL if adaptive, null if muxed/HLS
quality: string;
isMuxed: boolean; // true if videoUrl already contains audio
isDash: boolean; // true if we should build a DASH manifest
durationSeconds: number;
}
export interface YouTubeExtractionResult { export interface YouTubeExtractionResult {
source: TrailerPlaybackSource; /** Primary playable URL — HLS manifest, progressive muxed, or video-only adaptive */
videoUrl: string;
/** Separate audio URL when adaptive video-only is used. null for HLS/progressive. */
audioUrl: string | null;
quality: string;
videoId: string; videoId: string;
title?: string;
} }
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
// Client definitions // Constants — matching the Kotlin extractor exactly
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
// Used for all GET requests (watch page, HLS manifest fetch)
const DEFAULT_USER_AGENT =
'Mozilla/5.0 (Linux; Android 12; Android TV) AppleWebKit/537.36 ' +
'(KHTML, like Gecko) Chrome/133.0.0.0 Safari/537.36';
const DEFAULT_HEADERS: Record<string, string> = {
'accept-language': 'en-US,en;q=0.9',
'user-agent': DEFAULT_USER_AGENT,
};
const PREFERRED_ADAPTIVE_CLIENT = 'android_vr';
const REQUEST_TIMEOUT_MS = 20000;
interface ClientDef { interface ClientDef {
key: string; key: string;
id: string; id: string;
@ -96,21 +89,22 @@ interface ClientDef {
priority: number; priority: number;
} }
const PREFERRED_ADAPTIVE_CLIENT = 'android_vr'; // Matching the Kotlin extractor client list exactly (versions updated to current)
const CLIENTS: ClientDef[] = [ const CLIENTS: ClientDef[] = [
{ {
key: 'android_vr', key: 'android_vr',
id: '28', id: '28',
version: '1.62.27', version: '1.62.27',
userAgent: 'com.google.android.apps.youtube.vr.oculus/1.62.27 (Linux; U; Android 12L; eureka-user Build/SQ3A.220605.009.A1) gzip', userAgent:
'com.google.android.apps.youtube.vr.oculus/1.62.27 ' +
'(Linux; U; Android 12; en_US; Quest 3; Build/SQ3A.220605.009.A1) gzip',
context: { context: {
clientName: 'ANDROID_VR', clientName: 'ANDROID_VR',
clientVersion: '1.62.27', clientVersion: '1.62.27',
deviceMake: 'Oculus', deviceMake: 'Oculus',
deviceModel: 'Quest 3', deviceModel: 'Quest 3',
osName: 'Android', osName: 'Android',
osVersion: '12L', osVersion: '12',
platform: 'MOBILE', platform: 'MOBILE',
androidSdkVersion: 32, androidSdkVersion: 32,
hl: 'en', hl: 'en',
@ -122,7 +116,8 @@ const CLIENTS: ClientDef[] = [
key: 'android', key: 'android',
id: '3', id: '3',
version: '20.10.38', version: '20.10.38',
userAgent: 'com.google.android.youtube/20.10.38 (Linux; U; Android 14; en_US) gzip', userAgent:
'com.google.android.youtube/20.10.38 (Linux; U; Android 14; en_US) gzip',
context: { context: {
clientName: 'ANDROID', clientName: 'ANDROID',
clientVersion: '20.10.38', clientVersion: '20.10.38',
@ -139,7 +134,8 @@ const CLIENTS: ClientDef[] = [
key: 'ios', key: 'ios',
id: '5', id: '5',
version: '20.10.1', version: '20.10.1',
userAgent: 'com.google.ios.youtube/20.10.1 (iPhone16,2; U; CPU iOS 17_4 like Mac OS X)', userAgent:
'com.google.ios.youtube/20.10.1 (iPhone16,2; U; CPU iOS 17_4 like Mac OS X)',
context: { context: {
clientName: 'IOS', clientName: 'IOS',
clientVersion: '20.10.1', clientVersion: '20.10.1',
@ -154,13 +150,6 @@ const CLIENTS: ClientDef[] = [
}, },
]; ];
const REQUEST_TIMEOUT_MS = 20000;
const DEFAULT_UA = 'Mozilla/5.0 (Linux; Android 12; Android TV) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/133.0.0.0 Safari/537.36';
const DEFAULT_HEADERS: Record<string, string> = {
'accept-language': 'en-US,en;q=0.9',
'user-agent': DEFAULT_UA,
};
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
// Helpers // Helpers
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
@ -191,26 +180,16 @@ function getMimeBase(mimeType?: string): string {
return (mimeType ?? '').split(';')[0].trim(); return (mimeType ?? '').split(';')[0].trim();
} }
function getCodecs(mimeType?: string): string {
const m = (mimeType ?? '').match(/codecs="?([^"]+)"?/i);
return m ? m[1].trim() : '';
}
function getExt(mimeType?: string): 'mp4' | 'webm' | 'm4a' | 'other' { function getExt(mimeType?: string): 'mp4' | 'webm' | 'm4a' | 'other' {
const base = getMimeBase(mimeType); const base = getMimeBase(mimeType);
if (base.includes('mp4') || base === 'audio/mp4') return 'mp4'; if (base === 'video/mp4' || base === 'audio/mp4') return 'mp4';
if (base.includes('webm')) return 'webm'; if (base.includes('webm')) return 'webm';
if (base === 'audio/mp4' || base.includes('m4a')) return 'm4a'; if (base.includes('m4a')) return 'm4a';
return 'other'; return 'other';
} }
function containerScore(ext: string): number { function containerScore(ext: string): number {
switch (ext) { return ext === 'mp4' || ext === 'm4a' ? 0 : ext === 'webm' ? 1 : 2;
case 'mp4':
case 'm4a': return 0;
case 'webm': return 1;
default: return 2;
}
} }
function videoScore(height: number, fps: number, bitrate: number): number { function videoScore(height: number, fps: number, bitrate: number): number {
@ -222,24 +201,10 @@ function audioScore(bitrate: number, sampleRate: number): number {
} }
function parseQualityLabel(label?: string): number { function parseQualityLabel(label?: string): number {
if (!label) return 0; const m = (label ?? '').match(/(\d{2,4})p/);
const m = label.match(/(\d{2,4})p/);
return m ? parseInt(m[1], 10) : 0; return m ? parseInt(m[1], 10) : 0;
} }
function isMuxedFormat(f: InnertubeFormat): boolean {
const codecs = getCodecs(f.mimeType);
return (!!f.qualityLabel && !!f.audioQuality) || codecs.includes(',');
}
function isVideoOnly(f: InnertubeFormat): boolean {
return !!(f.qualityLabel && !f.audioQuality && f.mimeType?.startsWith('video/'));
}
function isAudioOnly(f: InnertubeFormat): boolean {
return !!(f.audioQuality && !f.qualityLabel && f.mimeType?.startsWith('audio/'));
}
function summarizeUrl(url: string): string { function summarizeUrl(url: string): string {
try { try {
const u = new URL(url); const u = new URL(url);
@ -249,6 +214,24 @@ function summarizeUrl(url: string): string {
} }
} }
function sortCandidates(items: StreamCandidate[]): StreamCandidate[] {
return [...items].sort((a, b) => {
if (b.score !== a.score) return b.score - a.score;
const ca = containerScore(a.ext), cb = containerScore(b.ext);
if (ca !== cb) return ca - cb;
return a.priority - b.priority;
});
}
function pickBestForClient(
items: StreamCandidate[],
preferredClient: string,
): StreamCandidate | null {
const fromPreferred = items.filter(c => c.client === preferredClient);
const pool = fromPreferred.length > 0 ? fromPreferred : items;
return sortCandidates(pool)[0] ?? null;
}
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
// Watch page — extract API key + visitor data dynamically // Watch page — extract API key + visitor data dynamically
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
@ -262,31 +245,29 @@ async function fetchWatchConfig(videoId: string): Promise<WatchConfig> {
const controller = new AbortController(); const controller = new AbortController();
const timer = setTimeout(() => controller.abort(), REQUEST_TIMEOUT_MS); const timer = setTimeout(() => controller.abort(), REQUEST_TIMEOUT_MS);
try { try {
const res = await fetch(`https://www.youtube.com/watch?v=${videoId}&hl=en`, { const res = await fetch(
headers: DEFAULT_HEADERS, `https://www.youtube.com/watch?v=${videoId}&hl=en`,
signal: controller.signal, { headers: DEFAULT_HEADERS, signal: controller.signal },
}); );
clearTimeout(timer); clearTimeout(timer);
if (!res.ok) { if (!res.ok) {
logger.warn('YouTubeExtractor', `Watch page fetch failed: ${res.status}`); logger.warn('YouTubeExtractor', `Watch page ${res.status}`);
return { apiKey: null, visitorData: null }; return { apiKey: null, visitorData: null };
} }
const html = await res.text(); const html = await res.text();
const apiKeyMatch = html.match(/"INNERTUBE_API_KEY":"([^"]+)"/); const apiKey = html.match(/"INNERTUBE_API_KEY":"([^"]+)"/)?.[1] ?? null;
const visitorMatch = html.match(/"VISITOR_DATA":"([^"]+)"/); const visitorData = html.match(/"VISITOR_DATA":"([^"]+)"/)?.[1] ?? null;
return { logger.info('YouTubeExtractor', `Watch page: apiKey=${apiKey ? 'found' : 'missing'} visitorData=${visitorData ? 'found' : 'missing'}`);
apiKey: apiKeyMatch?.[1] ?? null, return { apiKey, visitorData };
visitorData: visitorMatch?.[1] ?? null,
};
} catch (err) { } catch (err) {
clearTimeout(timer); clearTimeout(timer);
logger.warn('YouTubeExtractor', 'Watch page fetch error:', err); logger.warn('YouTubeExtractor', 'Watch page error:', err);
return { apiKey: null, visitorData: null }; return { apiKey: null, visitorData: null };
} }
} }
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
// Player API request // Player API
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
async function fetchPlayerResponse( async function fetchPlayerResponse(
@ -298,8 +279,7 @@ async function fetchPlayerResponse(
const controller = new AbortController(); const controller = new AbortController();
const timer = setTimeout(() => controller.abort(), REQUEST_TIMEOUT_MS); const timer = setTimeout(() => controller.abort(), REQUEST_TIMEOUT_MS);
// Use dynamic API key if available, otherwise omit (deprecated but harmless) const endpoint = apiKey
const url = apiKey
? `https://www.youtube.com/youtubei/v1/player?key=${encodeURIComponent(apiKey)}&prettyPrint=false` ? `https://www.youtube.com/youtubei/v1/player?key=${encodeURIComponent(apiKey)}&prettyPrint=false`
: `https://www.youtube.com/youtubei/v1/player?prettyPrint=false`; : `https://www.youtube.com/youtubei/v1/player?prettyPrint=false`;
@ -316,19 +296,24 @@ async function fetchPlayerResponse(
const body = JSON.stringify({ const body = JSON.stringify({
videoId, videoId,
context: { client: client.context },
contentCheckOk: true, contentCheckOk: true,
racyCheckOk: true, racyCheckOk: true,
context: { client: client.context },
playbackContext: { playbackContext: {
contentPlaybackContext: { html5Preference: 'HTML5_PREF_WANTS' }, contentPlaybackContext: { html5Preference: 'HTML5_PREF_WANTS' },
}, },
}); });
try { try {
const res = await fetch(url, { method: 'POST', headers, body, signal: controller.signal }); const res = await fetch(endpoint, {
method: 'POST',
headers,
body,
signal: controller.signal,
});
clearTimeout(timer); clearTimeout(timer);
if (!res.ok) { if (!res.ok) {
logger.warn('YouTubeExtractor', `[${client.key}] player API HTTP ${res.status}`); logger.warn('YouTubeExtractor', `[${client.key}] HTTP ${res.status}`);
return null; return null;
} }
return await res.json() as PlayerResponse; return await res.json() as PlayerResponse;
@ -344,14 +329,17 @@ async function fetchPlayerResponse(
} }
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
// HLS manifest parsing — pick best variant by height then bandwidth // HLS manifest parsing
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
async function parseBestHlsVariant(manifestUrl: string): Promise<HlsVariant | null> { async function parseBestHlsVariant(manifestUrl: string): Promise<HlsVariant | null> {
const controller = new AbortController(); const controller = new AbortController();
const timer = setTimeout(() => controller.abort(), REQUEST_TIMEOUT_MS); const timer = setTimeout(() => controller.abort(), REQUEST_TIMEOUT_MS);
try { try {
const res = await fetch(manifestUrl, { headers: DEFAULT_HEADERS, signal: controller.signal }); const res = await fetch(manifestUrl, {
headers: DEFAULT_HEADERS,
signal: controller.signal,
});
clearTimeout(timer); clearTimeout(timer);
if (!res.ok) return null; if (!res.ok) return null;
const text = await res.text(); const text = await res.text();
@ -364,14 +352,16 @@ async function parseBestHlsVariant(manifestUrl: string): Promise<HlsVariant | nu
const nextLine = lines[i + 1]; const nextLine = lines[i + 1];
if (!nextLine || nextLine.startsWith('#')) continue; if (!nextLine || nextLine.startsWith('#')) continue;
// Parse attributes // Parse attribute list
const attrs: Record<string, string> = {}; const attrs: Record<string, string> = {};
const attrStr = line.substring(line.indexOf(':') + 1);
let key = '', val = '', inKey = true, inQuote = false; let key = '', val = '', inKey = true, inQuote = false;
for (const ch of attrStr) { for (const ch of line.substring(line.indexOf(':') + 1)) {
if (inKey) { if (ch === '=') inKey = false; else key += ch; continue; } if (inKey) { if (ch === '=') inKey = false; else key += ch; continue; }
if (ch === '"') { inQuote = !inQuote; continue; } if (ch === '"') { inQuote = !inQuote; continue; }
if (ch === ',' && !inQuote) { attrs[key.trim()] = val.trim(); key = ''; val = ''; inKey = true; continue; } if (ch === ',' && !inQuote) {
if (key.trim()) attrs[key.trim()] = val.trim();
key = ''; val = ''; inKey = true; continue;
}
val += ch; val += ch;
} }
if (key.trim()) attrs[key.trim()] = val.trim(); if (key.trim()) attrs[key.trim()] = val.trim();
@ -381,14 +371,17 @@ async function parseBestHlsVariant(manifestUrl: string): Promise<HlsVariant | nu
const height = parseInt(res2[1] ?? '0', 10) || 0; const height = parseInt(res2[1] ?? '0', 10) || 0;
const bandwidth = parseInt(attrs['BANDWIDTH'] ?? '0', 10) || 0; const bandwidth = parseInt(attrs['BANDWIDTH'] ?? '0', 10) || 0;
// Absolutize URL
let variantUrl = nextLine; let variantUrl = nextLine;
if (!variantUrl.startsWith('http')) { if (!variantUrl.startsWith('http')) {
try { variantUrl = new URL(variantUrl, manifestUrl).toString(); } catch { /* keep as-is */ } try { variantUrl = new URL(variantUrl, manifestUrl).toString(); } catch { /* keep */ }
} }
const candidate: HlsVariant = { url: variantUrl, width, height, bandwidth }; const candidate: HlsVariant = { url: variantUrl, width, height, bandwidth };
if (!best || height > best.height || (height === best.height && bandwidth > best.bandwidth)) { if (
!best ||
candidate.height > best.height ||
(candidate.height === best.height && candidate.bandwidth > best.bandwidth)
) {
best = candidate; best = candidate;
} }
} }
@ -401,13 +394,13 @@ async function parseBestHlsVariant(manifestUrl: string): Promise<HlsVariant | nu
} }
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
// Format collection — tries ALL clients, collects from all // Format collection — tries ALL clients, collects from all (matching Kotlin)
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
interface CollectedFormats { interface CollectedFormats {
progressive: StreamCandidate[]; // muxed video+audio progressive: StreamCandidate[];
adaptiveVideo: StreamCandidate[]; // video-only adaptive adaptiveVideo: StreamCandidate[];
adaptiveAudio: StreamCandidate[]; // audio-only adaptive adaptiveAudio: StreamCandidate[];
hlsManifests: Array<{ clientKey: string; priority: number; url: string }>; hlsManifests: Array<{ clientKey: string; priority: number; url: string }>;
} }
@ -435,16 +428,17 @@ async function collectAllFormats(
const sd = resp.streamingData; const sd = resp.streamingData;
if (!sd) continue; if (!sd) continue;
// Collect HLS manifest URL if present
if (sd.hlsManifestUrl) { if (sd.hlsManifestUrl) {
hlsManifests.push({ clientKey: client.key, priority: client.priority, url: sd.hlsManifestUrl }); hlsManifests.push({ clientKey: client.key, priority: client.priority, url: sd.hlsManifestUrl });
} }
let clientProgressive = 0, clientVideo = 0, clientAudio = 0; let nProg = 0, nVid = 0, nAud = 0;
// Progressive (muxed) formats // Progressive (muxed) formats — matching Kotlin: skip non-video mimeTypes
for (const f of (sd.formats ?? [])) { for (const f of (sd.formats ?? [])) {
if (!f.url || !f.mimeType?.startsWith('video/')) continue; if (!f.url) continue;
const mimeBase = getMimeBase(f.mimeType);
if (f.mimeType && !mimeBase.startsWith('video/')) continue;
const height = f.height ?? parseQualityLabel(f.qualityLabel); const height = f.height ?? parseQualityLabel(f.qualityLabel);
const fps = f.fps ?? 0; const fps = f.fps ?? 0;
const bitrate = f.bitrate ?? f.averageBitrate ?? 0; const bitrate = f.bitrate ?? f.averageBitrate ?? 0;
@ -456,13 +450,10 @@ async function collectAllFormats(
height, height,
fps, fps,
ext: getExt(f.mimeType), ext: getExt(f.mimeType),
initRange: f.initRange,
indexRange: f.indexRange,
bitrate, bitrate,
mimeType: f.mimeType, mimeType: f.mimeType ?? '',
itag: f.itag,
}); });
clientProgressive++; nProg++;
} }
// Adaptive formats // Adaptive formats
@ -482,13 +473,10 @@ async function collectAllFormats(
height, height,
fps, fps,
ext: getExt(f.mimeType), ext: getExt(f.mimeType),
initRange: f.initRange,
indexRange: f.indexRange,
bitrate, bitrate,
mimeType: f.mimeType, mimeType: f.mimeType ?? '',
itag: f.itag,
}); });
clientVideo++; nVid++;
} else if (mimeBase.startsWith('audio/')) { } else if (mimeBase.startsWith('audio/')) {
const bitrate = f.bitrate ?? f.averageBitrate ?? 0; const bitrate = f.bitrate ?? f.averageBitrate ?? 0;
const sampleRate = parseFloat(f.audioSampleRate ?? '0') || 0; const sampleRate = parseFloat(f.audioSampleRate ?? '0') || 0;
@ -500,18 +488,15 @@ async function collectAllFormats(
height: 0, height: 0,
fps: 0, fps: 0,
ext: getExt(f.mimeType), ext: getExt(f.mimeType),
initRange: f.initRange,
indexRange: f.indexRange,
bitrate, bitrate,
audioSampleRate: f.audioSampleRate, audioSampleRate: f.audioSampleRate,
mimeType: f.mimeType, mimeType: f.mimeType ?? '',
itag: f.itag,
}); });
clientAudio++; nAud++;
} }
} }
logger.info('YouTubeExtractor', `[${client.key}] progressive=${clientProgressive} video=${clientVideo} audio=${clientAudio} hls=${sd.hlsManifestUrl ? 1 : 0}`); logger.info('YouTubeExtractor', `[${client.key}] progressive=${nProg} video=${nVid} audio=${nAud} hls=${sd.hlsManifestUrl ? 1 : 0}`);
} catch (err) { } catch (err) {
logger.warn('YouTubeExtractor', `[${client.key}] Failed:`, err); logger.warn('YouTubeExtractor', `[${client.key}] Failed:`, err);
} }
@ -520,80 +505,6 @@ async function collectAllFormats(
return { progressive, adaptiveVideo, adaptiveAudio, hlsManifests }; return { progressive, adaptiveVideo, adaptiveAudio, hlsManifests };
} }
// ---------------------------------------------------------------------------
// Sorting and selection
// ---------------------------------------------------------------------------
function sortCandidates(items: StreamCandidate[]): StreamCandidate[] {
return [...items].sort((a, b) => {
if (b.score !== a.score) return b.score - a.score;
const ca = containerScore(a.ext), cb = containerScore(b.ext);
if (ca !== cb) return ca - cb;
return a.priority - b.priority;
});
}
function pickBestForClient(items: StreamCandidate[], preferredClient: string): StreamCandidate | null {
const fromPreferred = items.filter(c => c.client === preferredClient);
const pool = fromPreferred.length > 0 ? fromPreferred : items;
return sortCandidates(pool)[0] ?? null;
}
// ---------------------------------------------------------------------------
// DASH manifest builder
// ---------------------------------------------------------------------------
async function buildDashManifest(
video: StreamCandidate,
audio: StreamCandidate,
videoId: string,
durationSeconds: number,
): Promise<string | null> {
try {
const FileSystem = await import('expo-file-system/legacy');
if (!FileSystem.cacheDirectory) return null;
const esc = (s: string) =>
s.replace(/&/g, '&amp;').replace(/"/g, '&quot;').replace(/</g, '&lt;').replace(/>/g, '&gt;');
const dur = `PT${durationSeconds}S`;
const vMime = getMimeBase(video.mimeType);
const aMime = getMimeBase(audio.mimeType);
const vCodec = getCodecs(video.mimeType);
const aCodec = getCodecs(audio.mimeType);
const vInit = video.initRange ? `${video.initRange.start}-${video.initRange.end}` : '0-0';
const vIdx = video.indexRange ? `${video.indexRange.start}-${video.indexRange.end}` : '0-0';
const aInit = audio.initRange ? `${audio.initRange.start}-${audio.initRange.end}` : '0-0';
const aIdx = audio.indexRange ? `${audio.indexRange.start}-${audio.indexRange.end}` : '0-0';
const mpd = `<?xml version="1.0" encoding="UTF-8"?>
<MPD xmlns="urn:mpeg:dash:schema:mpd:2011" type="static" mediaPresentationDuration="${dur}" minBufferTime="PT2S" profiles="urn:mpeg:dash:profile:isoff-on-demand:2011">
<Period duration="${dur}">
<AdaptationSet id="1" mimeType="${vMime}" codecs="${vCodec}" subsegmentAlignment="true" subsegmentStartsWithSAP="1">
<Representation id="v1" bandwidth="${video.bitrate}" height="${video.height}">
<BaseURL>${esc(video.url)}</BaseURL>
<SegmentBase indexRange="${vIdx}"><Initialization range="${vInit}"/></SegmentBase>
</Representation>
</AdaptationSet>
<AdaptationSet id="2" mimeType="${aMime}" codecs="${aCodec}" lang="en" subsegmentAlignment="true" subsegmentStartsWithSAP="1">
<Representation id="a1" bandwidth="${audio.bitrate}" audioSamplingRate="${audio.audioSampleRate ?? '44100'}">
<BaseURL>${esc(audio.url)}</BaseURL>
<SegmentBase indexRange="${aIdx}"><Initialization range="${aInit}"/></SegmentBase>
</Representation>
</AdaptationSet>
</Period>
</MPD>`;
const path = `${FileSystem.cacheDirectory}trailer_${videoId}.mpd`;
await FileSystem.writeAsStringAsync(path, mpd, { encoding: FileSystem.EncodingType.UTF8 });
logger.info('YouTubeExtractor', `DASH manifest: ${path} (video=${video.itag} ${video.height}p, audio=${audio.itag})`);
return path;
} catch (err) {
logger.warn('YouTubeExtractor', 'DASH manifest write failed:', err);
return null;
}
}
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
// Public API // Public API
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
@ -602,12 +513,15 @@ export class YouTubeExtractor {
/** /**
* Extract a playable source from a YouTube video ID or URL. * Extract a playable source from a YouTube video ID or URL.
* *
* Strategy: * Matches the Kotlin InAppYouTubeExtractor approach:
* 1. Fetch watch page to get dynamic API key + visitor data * 1. Fetch watch page for dynamic API key + visitor data
* 2. Try ALL clients in parallel collect formats from all that succeed * 2. Try ALL clients, collect formats from all that succeed
* 3. On Android: prefer DASH (adaptive video from android_vr + best audio) * 3. Pick best HLS variant (by resolution/bandwidth) as primary
* falling back to best HLS manifest, then best progressive * 4. Fall back to best progressive (muxed) if no HLS
* 4. On iOS: prefer HLS manifest (AVPlayer native), then progressive (muxed) *
* Note: Unlike the Kotlin version, we do not return separate videoUrl/audioUrl
* for adaptive streams react-native-video cannot merge two sources. HLS
* provides the best quality without needing a separate audio track.
*/ */
static async extract( static async extract(
videoIdOrUrl: string, videoIdOrUrl: string,
@ -622,30 +536,33 @@ export class YouTubeExtractor {
const effectivePlatform = platform ?? (Platform.OS === 'android' ? 'android' : 'ios'); const effectivePlatform = platform ?? (Platform.OS === 'android' ? 'android' : 'ios');
logger.info('YouTubeExtractor', `Extracting videoId=${videoId} platform=${effectivePlatform}`); logger.info('YouTubeExtractor', `Extracting videoId=${videoId} platform=${effectivePlatform}`);
// Step 1: fetch watch page for dynamic API key + visitor data // Step 1: watch page for dynamic API key + visitor data
const watchConfig = await fetchWatchConfig(videoId); const { apiKey, visitorData } = await fetchWatchConfig(videoId);
if (watchConfig.apiKey) {
logger.info('YouTubeExtractor', `Got apiKey and visitorData from watch page`);
} else {
logger.warn('YouTubeExtractor', `Could not get apiKey from watch page — proceeding without`);
}
// Step 2: collect formats from all clients // Step 2: collect formats from all clients
const { progressive, adaptiveVideo, adaptiveAudio, hlsManifests } = const { progressive, adaptiveVideo, adaptiveAudio, hlsManifests } =
await collectAllFormats(videoId, watchConfig.apiKey, watchConfig.visitorData); await collectAllFormats(videoId, apiKey, visitorData);
logger.info('YouTubeExtractor', `Totals: progressive=${progressive.length} adaptiveVideo=${adaptiveVideo.length} adaptiveAudio=${adaptiveAudio.length} hls=${hlsManifests.length}`); logger.info('YouTubeExtractor',
`Totals: progressive=${progressive.length} adaptiveVideo=${adaptiveVideo.length} ` +
`adaptiveAudio=${adaptiveAudio.length} hls=${hlsManifests.length}`
);
if (progressive.length === 0 && adaptiveVideo.length === 0 && hlsManifests.length === 0) { if (progressive.length === 0 && adaptiveVideo.length === 0 && hlsManifests.length === 0) {
logger.warn('YouTubeExtractor', `No usable formats for videoId=${videoId}`); logger.warn('YouTubeExtractor', `No usable formats for videoId=${videoId}`);
return null; return null;
} }
// Step 3: pick best HLS variant (by resolution/bandwidth) // Step 3: pick best HLS variant across all manifests
let bestHls: (HlsVariant & { manifestUrl: string }) | null = null; let bestHls: (HlsVariant & { manifestUrl: string }) | null = null;
for (const { url, priority } of hlsManifests.sort((a, b) => a.priority - b.priority)) { for (const { url } of hlsManifests.sort((a, b) => a.priority - b.priority)) {
const variant = await parseBestHlsVariant(url); const variant = await parseBestHlsVariant(url);
if (variant && (!bestHls || variant.height > bestHls.height || (variant.height === bestHls.height && variant.bandwidth > bestHls.bandwidth))) { if (
variant &&
(!bestHls ||
variant.height > bestHls.height ||
(variant.height === bestHls.height && variant.bandwidth > bestHls.bandwidth))
) {
bestHls = { ...variant, manifestUrl: url }; bestHls = { ...variant, manifestUrl: url };
} }
} }
@ -655,95 +572,54 @@ export class YouTubeExtractor {
const bestAdaptiveAudio = pickBestForClient(adaptiveAudio, PREFERRED_ADAPTIVE_CLIENT); const bestAdaptiveAudio = pickBestForClient(adaptiveAudio, PREFERRED_ADAPTIVE_CLIENT);
if (bestHls) logger.info('YouTubeExtractor', `Best HLS: ${bestHls.height}p ${bestHls.bandwidth}bps`); if (bestHls) logger.info('YouTubeExtractor', `Best HLS: ${bestHls.height}p ${bestHls.bandwidth}bps`);
if (bestProgressive) logger.info('YouTubeExtractor', `Best progressive: ${bestProgressive.height}p score=${bestProgressive.score}`); if (bestProgressive) logger.info('YouTubeExtractor', `Best progressive: ${bestProgressive.height}p`);
if (bestAdaptiveVideo) logger.info('YouTubeExtractor', `Best adaptive video: ${bestAdaptiveVideo.height}p client=${bestAdaptiveVideo.client}`); if (bestAdaptiveVideo) logger.info('YouTubeExtractor', `Best adaptive video: ${bestAdaptiveVideo.height}p client=${bestAdaptiveVideo.client}`);
if (bestAdaptiveAudio) logger.info('YouTubeExtractor', `Best adaptive audio: bitrate=${bestAdaptiveAudio.bitrate} client=${bestAdaptiveAudio.client}`); if (bestAdaptiveAudio) logger.info('YouTubeExtractor', `Best adaptive audio: ${bestAdaptiveAudio.bitrate}bps client=${bestAdaptiveAudio.client}`);
// Placeholder duration (refined below) // Step 4: select final source
const durationSeconds = 300; // Priority: HLS > progressive muxed
// (No DASH/MPD — react-native-video cannot merge separate video+audio streams)
let source: TrailerPlaybackSource | null = null; if (bestHls) {
logger.info('YouTubeExtractor', `Using HLS: ${summarizeUrl(bestHls.manifestUrl)}`);
if (effectivePlatform === 'android') { return {
// Android priority: DASH adaptive > HLS > progressive videoUrl: bestHls.manifestUrl,
if (bestAdaptiveVideo && bestAdaptiveAudio) { audioUrl: null,
const mpdPath = await buildDashManifest(bestAdaptiveVideo, bestAdaptiveAudio, videoId, durationSeconds); quality: `${bestHls.height}p`,
if (mpdPath) { videoId,
source = { };
videoUrl: mpdPath,
audioUrl: null,
quality: `${bestAdaptiveVideo.height}p`,
isMuxed: true, // MPD contains both tracks
isDash: true,
durationSeconds,
};
}
}
if (!source && bestHls) {
source = {
videoUrl: bestHls.manifestUrl,
audioUrl: null,
quality: `${bestHls.height}p`,
isMuxed: true,
isDash: false,
durationSeconds,
};
}
if (!source && bestProgressive) {
source = {
videoUrl: bestProgressive.url,
audioUrl: null,
quality: `${bestProgressive.height}p`,
isMuxed: true,
isDash: false,
durationSeconds,
};
}
} else {
// iOS priority: HLS (native AVPlayer support) > progressive
if (bestHls) {
source = {
videoUrl: bestHls.manifestUrl,
audioUrl: null,
quality: `${bestHls.height}p`,
isMuxed: true,
isDash: false,
durationSeconds,
};
}
if (!source && bestProgressive) {
source = {
videoUrl: bestProgressive.url,
audioUrl: null,
quality: `${bestProgressive.height}p`,
isMuxed: true,
isDash: false,
durationSeconds,
};
}
} }
if (!source) { if (bestProgressive) {
logger.warn('YouTubeExtractor', `Could not build a playable source for videoId=${videoId}`); logger.info('YouTubeExtractor', `Using progressive: ${summarizeUrl(bestProgressive.url)}`);
return null; return {
videoUrl: bestProgressive.url,
audioUrl: null,
quality: `${bestProgressive.height}p`,
videoId,
};
} }
logger.info('YouTubeExtractor', `Final source: ${summarizeUrl(source.videoUrl)} quality=${source.quality} dash=${source.isDash}`); // Last resort: video-only adaptive (no audio, but beats nothing)
if (bestAdaptiveVideo) {
logger.warn('YouTubeExtractor', `Using video-only adaptive (no audio): ${bestAdaptiveVideo.height}p`);
return {
videoUrl: bestAdaptiveVideo.url,
audioUrl: null,
quality: `${bestAdaptiveVideo.height}p`,
videoId,
};
}
return { logger.warn('YouTubeExtractor', `No playable source for videoId=${videoId}`);
source, return null;
videoId,
title: undefined,
};
} }
/** Convenience: returns just the best playable URL or null. */
static async getBestStreamUrl( static async getBestStreamUrl(
videoIdOrUrl: string, videoIdOrUrl: string,
platform?: 'android' | 'ios', platform?: 'android' | 'ios',
): Promise<string | null> { ): Promise<string | null> {
const result = await this.extract(videoIdOrUrl, platform); const result = await this.extract(videoIdOrUrl, platform);
return result?.source.videoUrl ?? null; return result?.videoUrl ?? null;
} }
static parseVideoId(input: string): string | null { static parseVideoId(input: string): string | null {