server/tools/peertube-import-videos.ts

   1 import { registerTSPaths } from '../helpers/register-ts-paths'
   2 registerTSPaths()
   3
   4 import * as program from 'commander'
   5 import { join } from 'path'
   6 import { doRequestAndSaveToFile } from '../helpers/requests'
   7 import { CONSTRAINTS_FIELDS } from '../initializers/constants'
   8 import { getClient, getVideoCategories, login, searchVideoWithSort, uploadVideo } from '../../shared/extra-utils/index'
   9 import { truncate } from 'lodash'
  10 import * as prompt from 'prompt'
  11 import { accessSync, constants } from 'fs'
  12 import { remove } from 'fs-extra'
  13 import { sha256 } from '../helpers/core-utils'
  14 import { buildOriginallyPublishedAt, safeGetYoutubeDL } from '../helpers/youtube-dl'
  15 import { buildCommonVideoOptions, buildVideoAttributesFromCommander, getLogger, getServerCredentials } from './cli'
  16
  17 type UserInfo = {
  18   username: string
  19   password: string
  20 }
  21
  22 const processOptions = {
  23   maxBuffer: Infinity
  24 }
  25
  26 let command = program
  27   .name('import-videos')
  28
  29 command = buildCommonVideoOptions(command)
  30
  31 command
  32   .option('-u, --url <url>', 'Server url')
  33   .option('-U, --username <username>', 'Username')
  34   .option('-p, --password <token>', 'Password')
  35   .option('--target-url <targetUrl>', 'Video target URL')
  36   .option('--since <since>', 'Publication date (inclusive) since which the videos can be imported (YYYY-MM-DD)', parseDate)
  37   .option('--until <until>', 'Publication date (inclusive) until which the videos can be imported (YYYY-MM-DD)', parseDate)
  38   .option('--first <first>', 'Process first n elements of returned playlist')
  39   .option('--last <last>', 'Process last n elements of returned playlist')
  40   .option('-T, --tmpdir <tmpdir>', 'Working directory', __dirname)
  41   .usage("[global options] [ -- youtube-dl options]")
  42   .parse(process.argv)
  43
  44 const log = getLogger(program['verbose'])
  45
  46 getServerCredentials(command)
  47   .then(({ url, username, password }) => {
  48     if (!program['targetUrl']) {
  49       exitError('--target-url field is required.')
  50     }
  51
  52     try {
  53       accessSync(program['tmpdir'], constants.R_OK | constants.W_OK)
  54     } catch (e) {
  55       exitError('--tmpdir %s: directory does not exist or is not accessible', program['tmpdir'])
  56     }
  57
  58     url = normalizeTargetUrl(url)
  59     program['targetUrl'] = normalizeTargetUrl(program['targetUrl'])
  60
  61     const user = { username, password }
  62
  63     run(url, user)
  64       .catch(err => exitError(err))
  65   })
  66   .catch(err => console.error(err))
  67
  68 async function run (url: string, user: UserInfo) {
  69   if (!user.password) {
  70     user.password = await promptPassword()
  71   }
  72
  73   const youtubeDL = await safeGetYoutubeDL()
  74
  75   let info = await getYoutubeDLInfo(youtubeDL, program['targetUrl'], command.args)
  76
  77   if (info?.title === 'Uploads') {
  78     console.log('Fixing URL to %s.', info.url)
  79
  80     info = await getYoutubeDLInfo(youtubeDL, info.url, command.args)
  81   }
  82
  83   let infoArray: any[]
  84
  85   // Normalize utf8 fields
  86   infoArray = [].concat(info)
  87   if (program['first']) {
  88     infoArray = infoArray.slice(0, program['first'])
  89   } else if (program['last']) {
  90     infoArray = infoArray.slice(-program['last'])
  91   }
  92   infoArray = infoArray.map(i => normalizeObject(i))
  93
  94   log.info('Will download and upload %d videos.\n', infoArray.length)
  95
  96   for (const info of infoArray) {
  97     try {
  98       await processVideo({
  99         cwd: program['tmpdir'],
 100         url,
 101         user,
 102         youtubeInfo: info
 103       })
 104     } catch (err) {
 105       console.error('Cannot process video.', { info, url })
 106     }
 107   }
 108
 109   log.info('Video/s for user %s imported: %s', user.username, program['targetUrl'])
 110   process.exit(0)
 111 }
 112
 113 function processVideo (parameters: {
 114   cwd: string
 115   url: string
 116   user: { username: string, password: string }
 117   youtubeInfo: any
 118 }) {
 119   const { youtubeInfo, cwd, url, user } = parameters
 120
 121   return new Promise(async res => {
 122     log.debug('Fetching object.', youtubeInfo)
 123
 124     const videoInfo = await fetchObject(youtubeInfo)
 125     log.debug('Fetched object.', videoInfo)
 126
 127     if (program['since']) {
 128       if (buildOriginallyPublishedAt(videoInfo).getTime() < program['since'].getTime()) {
 129         log.info('Video "%s" has been published before "%s", don\'t upload it.\n',
 130           videoInfo.title, formatDate(program['since']))
 131         return res()
 132       }
 133     }
 134     if (program['until']) {
 135       if (buildOriginallyPublishedAt(videoInfo).getTime() > program['until'].getTime()) {
 136         log.info('Video "%s" has been published after "%s", don\'t upload it.\n',
 137           videoInfo.title, formatDate(program['until']))
 138         return res()
 139       }
 140     }
 141
 142     const result = await searchVideoWithSort(url, videoInfo.title, '-match')
 143
 144     log.info('############################################################\n')
 145
 146     if (result.body.data.find(v => v.name === videoInfo.title)) {
 147       log.info('Video "%s" already exists, don\'t reupload it.\n', videoInfo.title)
 148       return res()
 149     }
 150
 151     const path = join(cwd, sha256(videoInfo.url) + '.mp4')
 152
 153     log.info('Downloading video "%s"...', videoInfo.title)
 154
 155     const options = [ '-f', 'bestvideo[ext=mp4]+bestaudio[ext=m4a]/best', ...command.args, '-o', path ]
 156     try {
 157       const youtubeDL = await safeGetYoutubeDL()
 158       youtubeDL.exec(videoInfo.url, options, processOptions, async (err, output) => {
 159         if (err) {
 160           log.error(err)
 161           return res()
 162         }
 163
 164         log.info(output.join('\n'))
 165         await uploadVideoOnPeerTube({
 166           cwd,
 167           url,
 168           user,
 169           videoInfo: normalizeObject(videoInfo),
 170           videoPath: path
 171         })
 172         return res()
 173       })
 174     } catch (err) {
 175       log.error(err.message)
 176       return res()
 177     }
 178   })
 179 }
 180
 181 async function uploadVideoOnPeerTube (parameters: {
 182   videoInfo: any
 183   videoPath: string
 184   cwd: string
 185   url: string
 186   user: { username: string, password: string }
 187 }) {
 188   const { videoInfo, videoPath, cwd, url, user } = parameters
 189
 190   const category = await getCategory(videoInfo.categories, url)
 191   const licence = getLicence(videoInfo.license)
 192   let tags = []
 193   if (Array.isArray(videoInfo.tags)) {
 194     tags = videoInfo.tags
 195                     .filter(t => t.length < CONSTRAINTS_FIELDS.VIDEOS.TAG.max && t.length > CONSTRAINTS_FIELDS.VIDEOS.TAG.min)
 196                     .map(t => t.normalize())
 197                     .slice(0, 5)
 198   }
 199
 200   let thumbnailfile
 201   if (videoInfo.thumbnail) {
 202     thumbnailfile = join(cwd, sha256(videoInfo.thumbnail) + '.jpg')
 203
 204     await doRequestAndSaveToFile({
 205       method: 'GET',
 206       uri: videoInfo.thumbnail
 207     }, thumbnailfile)
 208   }
 209
 210   const originallyPublishedAt = buildOriginallyPublishedAt(videoInfo)
 211
 212   const defaultAttributes = {
 213     name: truncate(videoInfo.title, {
 214       length: CONSTRAINTS_FIELDS.VIDEOS.NAME.max,
 215       separator: /,? +/,
 216       omission: ' […]'
 217     }),
 218     category,
 219     licence,
 220     nsfw: isNSFW(videoInfo),
 221     description: videoInfo.description,
 222     tags
 223   }
 224
 225   const videoAttributes = await buildVideoAttributesFromCommander(url, program, defaultAttributes)
 226
 227   Object.assign(videoAttributes, {
 228     originallyPublishedAt: originallyPublishedAt ? originallyPublishedAt.toISOString() : null,
 229     thumbnailfile,
 230     previewfile: thumbnailfile,
 231     fixture: videoPath
 232   })
 233
 234   log.info('\nUploading on PeerTube video "%s".', videoAttributes.name)
 235
 236   let accessToken = await getAccessTokenOrDie(url, user)
 237
 238   try {
 239     await uploadVideo(url, accessToken, videoAttributes)
 240   } catch (err) {
 241     if (err.message.indexOf('401') !== -1) {
 242       log.info('Got 401 Unauthorized, token may have expired, renewing token and retry.')
 243
 244       accessToken = await getAccessTokenOrDie(url, user)
 245
 246       await uploadVideo(url, accessToken, videoAttributes)
 247     } else {
 248       exitError(err.message)
 249     }
 250   }
 251
 252   await remove(videoPath)
 253   if (thumbnailfile) await remove(thumbnailfile)
 254
 255   log.warn('Uploaded video "%s"!\n', videoAttributes.name)
 256 }
 257
 258 /* ---------------------------------------------------------- */
 259
 260 async function getCategory (categories: string[], url: string) {
 261   if (!categories) return undefined
 262
 263   const categoryString = categories[0]
 264
 265   if (categoryString === 'News & Politics') return 11
 266
 267   const res = await getVideoCategories(url)
 268   const categoriesServer = res.body
 269
 270   for (const key of Object.keys(categoriesServer)) {
 271     const categoryServer = categoriesServer[key]
 272     if (categoryString.toLowerCase() === categoryServer.toLowerCase()) return parseInt(key, 10)
 273   }
 274
 275   return undefined
 276 }
 277
 278 function getLicence (licence: string) {
 279   if (!licence) return undefined
 280
 281   if (licence.includes('Creative Commons Attribution licence')) return 1
 282
 283   return undefined
 284 }
 285
 286 function normalizeObject (obj: any) {
 287   const newObj: any = {}
 288
 289   for (const key of Object.keys(obj)) {
 290     // Deprecated key
 291     if (key === 'resolution') continue
 292
 293     const value = obj[key]
 294
 295     if (typeof value === 'string') {
 296       newObj[key] = value.normalize()
 297     } else {
 298       newObj[key] = value
 299     }
 300   }
 301
 302   return newObj
 303 }
 304
 305 function fetchObject (info: any) {
 306   const url = buildUrl(info)
 307
 308   return new Promise<any>(async (res, rej) => {
 309     const youtubeDL = await safeGetYoutubeDL()
 310     youtubeDL.getInfo(url, undefined, processOptions, (err, videoInfo) => {
 311       if (err) return rej(err)
 312
 313       const videoInfoWithUrl = Object.assign(videoInfo, { url })
 314       return res(normalizeObject(videoInfoWithUrl))
 315     })
 316   })
 317 }
 318
 319 function buildUrl (info: any) {
 320   const webpageUrl = info.webpage_url as string
 321   if (webpageUrl?.match(/^https?:\/\//)) return webpageUrl
 322
 323   const url = info.url as string
 324   if (url?.match(/^https?:\/\//)) return url
 325
 326   // It seems youtube-dl does not return the video url
 327   return 'https://www.youtube.com/watch?v=' + info.id
 328 }
 329
 330 function isNSFW (info: any) {
 331   return info.age_limit && info.age_limit >= 16
 332 }
 333
 334 function normalizeTargetUrl (url: string) {
 335   let normalizedUrl = url.replace(/\/+$/, '')
 336
 337   if (!normalizedUrl.startsWith('http://') && !normalizedUrl.startsWith('https://')) {
 338     normalizedUrl = 'https://' + normalizedUrl
 339   }
 340
 341   return normalizedUrl
 342 }
 343
 344 async function promptPassword () {
 345   return new Promise<string>((res, rej) => {
 346     prompt.start()
 347     const schema = {
 348       properties: {
 349         password: {
 350           hidden: true,
 351           required: true
 352         }
 353       }
 354     }
 355     prompt.get(schema, function (err, result) {
 356       if (err) {
 357         return rej(err)
 358       }
 359       return res(result.password)
 360     })
 361   })
 362 }
 363
 364 async function getAccessTokenOrDie (url: string, user: UserInfo) {
 365   const resClient = await getClient(url)
 366   const client = {
 367     id: resClient.body.client_id,
 368     secret: resClient.body.client_secret
 369   }
 370
 371   try {
 372     const res = await login(url, client, user)
 373     return res.body.access_token
 374   } catch (err) {
 375     exitError('Cannot authenticate. Please check your username/password.')
 376   }
 377 }
 378
 379 function parseDate (dateAsStr: string): Date {
 380   if (!/\d{4}-\d{2}-\d{2}/.test(dateAsStr)) {
 381     exitError(`Invalid date passed: ${dateAsStr}. Expected format: YYYY-MM-DD. See help for usage.`)
 382   }
 383   const date = new Date(dateAsStr)
 384   date.setHours(0, 0, 0)
 385   if (isNaN(date.getTime())) {
 386     exitError(`Invalid date passed: ${dateAsStr}. See help for usage.`)
 387   }
 388   return date
 389 }
 390
 391 function formatDate (date: Date): string {
 392   return date.toISOString().split('T')[0]
 393 }
 394
 395 function exitError (message: string, ...meta: any[]) {
 396   // use console.error instead of log.error here
 397   console.error(message, ...meta)
 398   process.exit(-1)
 399 }
 400
 401 function getYoutubeDLInfo (youtubeDL: any, url: string, args: string[]) {
 402   return new Promise<any>((res, rej) => {
 403     const options = [ '-j', '--flat-playlist', '--playlist-reverse', ...args ]
 404
 405     youtubeDL.getInfo(url, options, processOptions, async (err, info) => {
 406       if (err) return rej(err)
 407
 408       return res(info)
 409     })
 410   })
 411 }