X-Git-Url: https://git.immae.eu/?a=blobdiff_plain;f=server%2Fhelpers%2Fyoutube-dl.ts;h=c2aa8870ae3b5d7627899077b6138241b750ab35;hb=c6c0fa6cd8fe8f752463d8982c3dbcd448739c4e;hp=a5ab92df0b865240bb3f101b5e698c1f9e7aef14;hpb=84929846e743f4b5f3971b386ea2c447db8c416f;p=github%2FChocobozzz%2FPeerTube.git diff --git a/server/helpers/youtube-dl.ts b/server/helpers/youtube-dl.ts index a5ab92df0..c2aa8870a 100644 --- a/server/helpers/youtube-dl.ts +++ b/server/helpers/youtube-dl.ts @@ -1,73 +1,132 @@ -import { truncate } from 'lodash' -import { CONSTRAINTS_FIELDS, VIDEO_CATEGORIES } from '../initializers' +import { CONSTRAINTS_FIELDS, VIDEO_CATEGORIES, VIDEO_LANGUAGES, VIDEO_LICENCES } from '../initializers/constants' import { logger } from './logger' import { generateVideoImportTmpPath } from './utils' import { join } from 'path' -import { root } from './core-utils' -import { ensureDir, writeFile, remove } from 'fs-extra' +import { peertubeTruncate, root } from './core-utils' +import { ensureDir, remove, writeFile } from 'fs-extra' import * as request from 'request' import { createWriteStream } from 'fs' +import { CONFIG } from '@server/initializers/config' export type YoutubeDLInfo = { name?: string description?: string category?: number + language?: string licence?: number nsfw?: boolean tags?: string[] thumbnailUrl?: string - originallyPublishedAt?: string + fileExt?: string + originallyPublishedAt?: Date } +export type YoutubeDLSubs = { + language: string + filename: string + path: string +}[] + const processOptions = { maxBuffer: 1024 * 1024 * 10 // 10MB } function getYoutubeDLInfo (url: string, opts?: string[]): Promise { - return new Promise(async (res, rej) => { - const args = opts || [ '-j', '--flat-playlist' ] + return new Promise((res, rej) => { + let args = opts || [ '-j', '--flat-playlist' ] + args = wrapWithProxyOptions(args) - const youtubeDL = await safeGetYoutubeDL() - youtubeDL.getInfo(url, args, processOptions, (err, info) => { - if (err) return rej(err) - if (info.is_live === true) return rej(new Error('Cannot download a live streaming.')) + safeGetYoutubeDL() + .then(youtubeDL => { + youtubeDL.getInfo(url, args, processOptions, (err, info) => { + if (err) return rej(err) + if (info.is_live === true) return rej(new Error('Cannot download a live streaming.')) - const obj = buildVideoInfo(normalizeObject(info)) - if (obj.name && obj.name.length < CONSTRAINTS_FIELDS.VIDEOS.NAME.min) obj.name += ' video' + const obj = buildVideoInfo(normalizeObject(info)) + if (obj.name && obj.name.length < CONSTRAINTS_FIELDS.VIDEOS.NAME.min) obj.name += ' video' - return res(obj) - }) + return res(obj) + }) + }) + .catch(err => rej(err)) + }) +} + +function getYoutubeDLSubs (url: string, opts?: object): Promise { + return new Promise((res, rej) => { + const cwd = CONFIG.STORAGE.TMP_DIR + const options = opts || { all: true, format: 'vtt', cwd } + + safeGetYoutubeDL() + .then(youtubeDL => { + youtubeDL.getSubs(url, options, (err, files) => { + if (err) return rej(err) + if (!files) return [] + + logger.debug('Get subtitles from youtube dl.', { url, files }) + + const subtitles = files.reduce((acc, filename) => { + const matched = filename.match(/\.([a-z]{2})\.(vtt|ttml)/i) + if (!matched || !matched[1]) return acc + + return [ + ...acc, + { + language: matched[1], + path: join(cwd, filename), + filename + } + ] + }, []) + + return res(subtitles) + }) + }) + .catch(err => rej(err)) }) } -function downloadYoutubeDLVideo (url: string, timeout: number) { - const path = generateVideoImportTmpPath(url) +function downloadYoutubeDLVideo (url: string, extension: string, timeout: number) { + const path = generateVideoImportTmpPath(url, extension) let timer - logger.info('Importing youtubeDL video %s', url) + logger.info('Importing youtubeDL video %s to %s', url, path) - const options = [ '-f', 'bestvideo[ext=mp4]+bestaudio[ext=m4a]/best', '-o', path ] + let options = [ '-f', 'bestvideo[ext=mp4]+bestaudio[ext=m4a]/best', '-o', path ] + options = wrapWithProxyOptions(options) - return new Promise(async (res, rej) => { - const youtubeDL = await safeGetYoutubeDL() - youtubeDL.exec(url, options, processOptions, err => { - clearTimeout(timer) + if (process.env.FFMPEG_PATH) { + options = options.concat([ '--ffmpeg-location', process.env.FFMPEG_PATH ]) + } - if (err) { - remove(path) - .catch(err => logger.error('Cannot delete path on YoutubeDL error.', { err })) + return new Promise((res, rej) => { + safeGetYoutubeDL() + .then(youtubeDL => { + youtubeDL.exec(url, options, processOptions, err => { + clearTimeout(timer) - return rej(err) - } + if (err) { + remove(path) + .catch(err => logger.error('Cannot delete path on YoutubeDL error.', { err })) - return res(path) - }) + return rej(err) + } + + return res(path) + }) - timer = setTimeout(async () => { - await remove(path) + timer = setTimeout(() => { + const err = new Error('YoutubeDL download timeout.') - return rej(new Error('YoutubeDL download timeout.')) - }, timeout) + remove(path) + .finally(() => rej(err)) + .catch(err => { + logger.error('Cannot remove %s in youtubeDL timeout.', path, { err }) + return rej(err) + }) + }, timeout) + }) + .catch(err => rej(err)) }) } @@ -79,7 +138,7 @@ async function updateYoutubeDLBinary () { const binDirectory = join(root(), 'node_modules', 'youtube-dl', 'bin') const bin = join(binDirectory, 'youtube-dl') const detailsPath = join(binDirectory, 'details') - const url = 'https://yt-dl.org/downloads/latest/youtube-dl' + const url = process.env.YOUTUBE_DL_DOWNLOAD_HOST || 'https://yt-dl.org/downloads/latest/youtube-dl' await ensureDir(binDirectory) @@ -97,7 +156,7 @@ async function updateYoutubeDLBinary () { const url = result.headers.location const downloadFile = request.get(url) - const newVersion = /yt-dl\.org\/downloads\/(\d{4}\.\d\d\.\d\d(\.\d)?)\/youtube-dl/.exec(url)[ 1 ] + const newVersion = /yt-dl\.org\/downloads\/(\d{4}\.\d\d\.\d\d(\.\d)?)\/youtube-dl/.exec(url)[1] downloadFile.on('response', result => { if (result.statusCode !== 200) { @@ -143,13 +202,34 @@ async function safeGetYoutubeDL () { return youtubeDL } +function buildOriginallyPublishedAt (obj: any) { + let originallyPublishedAt: Date = null + + const uploadDateMatcher = /^(\d{4})(\d{2})(\d{2})$/.exec(obj.upload_date) + if (uploadDateMatcher) { + originallyPublishedAt = new Date() + originallyPublishedAt.setHours(0, 0, 0, 0) + + const year = parseInt(uploadDateMatcher[1], 10) + // Month starts from 0 + const month = parseInt(uploadDateMatcher[2], 10) - 1 + const day = parseInt(uploadDateMatcher[3], 10) + + originallyPublishedAt.setFullYear(year, month, day) + } + + return originallyPublishedAt +} + // --------------------------------------------------------------------------- export { updateYoutubeDLBinary, downloadYoutubeDLVideo, + getYoutubeDLSubs, getYoutubeDLInfo, - safeGetYoutubeDL + safeGetYoutubeDL, + buildOriginallyPublishedAt } // --------------------------------------------------------------------------- @@ -173,37 +253,36 @@ function normalizeObject (obj: any) { return newObj } -function buildVideoInfo (obj: any) { - - const date = obj.upload_date.slice(0,4) + ',' + obj.upload_date.slice(4,6) + ',' + obj.upload_date.slice(6,8) - +function buildVideoInfo (obj: any): YoutubeDLInfo { return { name: titleTruncation(obj.title), description: descriptionTruncation(obj.description), category: getCategory(obj.categories), licence: getLicence(obj.license), + language: getLanguage(obj.language), nsfw: isNSFW(obj), tags: getTags(obj.tags), thumbnailUrl: obj.thumbnail || undefined, - originallyPublishedAt: new Date(date).toISOString() + originallyPublishedAt: buildOriginallyPublishedAt(obj), + fileExt: obj.ext } } function titleTruncation (title: string) { - return truncate(title, { - 'length': CONSTRAINTS_FIELDS.VIDEOS.NAME.max, - 'separator': /,? +/, - 'omission': ' […]' + return peertubeTruncate(title, { + length: CONSTRAINTS_FIELDS.VIDEOS.NAME.max, + separator: /,? +/, + omission: ' […]' }) } function descriptionTruncation (description: string) { if (!description || description.length < CONSTRAINTS_FIELDS.VIDEOS.DESCRIPTION.min) return undefined - return truncate(description, { - 'length': CONSTRAINTS_FIELDS.VIDEOS.DESCRIPTION.max, - 'separator': /,? +/, - 'omission': ' […]' + return peertubeTruncate(description, { + length: CONSTRAINTS_FIELDS.VIDEOS.DESCRIPTION.max, + separator: /,? +/, + omission: ' […]' }) } @@ -223,7 +302,12 @@ function getTags (tags: any) { function getLicence (licence: string) { if (!licence) return undefined - if (licence.indexOf('Creative Commons Attribution') !== -1) return 1 + if (licence.includes('Creative Commons Attribution')) return 1 + + for (const key of Object.keys(VIDEO_LICENCES)) { + const peertubeLicence = VIDEO_LICENCES[key] + if (peertubeLicence.toLowerCase() === licence.toLowerCase()) return parseInt(key, 10) + } return undefined } @@ -243,3 +327,17 @@ function getCategory (categories: string[]) { return undefined } + +function getLanguage (language: string) { + return VIDEO_LANGUAGES[language] ? language : undefined +} + +function wrapWithProxyOptions (options: string[]) { + if (CONFIG.IMPORT.VIDEOS.HTTP.PROXY.ENABLED) { + logger.debug('Using proxy for YoutubeDL') + + return [ '--proxy', CONFIG.IMPORT.VIDEOS.HTTP.PROXY.URL ].concat(options) + } + + return options +}