]> git.immae.eu Git - github/Chocobozzz/PeerTube.git/blob - server/tools/peertube-import-videos.ts
Fix missing python dependency in Dockerfile.buster (#2495)
[github/Chocobozzz/PeerTube.git] / 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 const options = [ '-j', '--flat-playlist', '--playlist-reverse', ...command.args ]
76
77 youtubeDL.getInfo(program['targetUrl'], options, processOptions, async (err, info) => {
78 if (err) {
79 exitError(err.stderr + ' ' + err.message)
80 }
81
82 let infoArray: any[]
83
84 // Normalize utf8 fields
85 infoArray = [].concat(info)
86 if (program['first']) {
87 infoArray = infoArray.slice(0, program['first'])
88 } else if (program['last']) {
89 infoArray = infoArray.slice(-program['last'])
90 }
91 infoArray = infoArray.map(i => normalizeObject(i))
92
93 log.info('Will download and upload %d videos.\n', infoArray.length)
94
95 for (const info of infoArray) {
96 await processVideo({
97 cwd: program['tmpdir'],
98 url,
99 user,
100 youtubeInfo: info
101 })
102 }
103
104 log.info('Video/s for user %s imported: %s', user.username, program['targetUrl'])
105 process.exit(0)
106 })
107 }
108
109 function processVideo (parameters: {
110 cwd: string
111 url: string
112 user: { username: string, password: string }
113 youtubeInfo: any
114 }) {
115 const { youtubeInfo, cwd, url, user } = parameters
116
117 return new Promise(async res => {
118 log.debug('Fetching object.', youtubeInfo)
119
120 const videoInfo = await fetchObject(youtubeInfo)
121 log.debug('Fetched object.', videoInfo)
122
123 if (program['since']) {
124 if (buildOriginallyPublishedAt(videoInfo).getTime() < program['since'].getTime()) {
125 log.info('Video "%s" has been published before "%s", don\'t upload it.\n',
126 videoInfo.title, formatDate(program['since']))
127 return res()
128 }
129 }
130 if (program['until']) {
131 if (buildOriginallyPublishedAt(videoInfo).getTime() > program['until'].getTime()) {
132 log.info('Video "%s" has been published after "%s", don\'t upload it.\n',
133 videoInfo.title, formatDate(program['until']))
134 return res()
135 }
136 }
137
138 const result = await searchVideoWithSort(url, videoInfo.title, '-match')
139
140 log.info('############################################################\n')
141
142 if (result.body.data.find(v => v.name === videoInfo.title)) {
143 log.info('Video "%s" already exists, don\'t reupload it.\n', videoInfo.title)
144 return res()
145 }
146
147 const path = join(cwd, sha256(videoInfo.url) + '.mp4')
148
149 log.info('Downloading video "%s"...', videoInfo.title)
150
151 const options = [ '-f', 'bestvideo[ext=mp4]+bestaudio[ext=m4a]/best', ...command.args, '-o', path ]
152 try {
153 const youtubeDL = await safeGetYoutubeDL()
154 youtubeDL.exec(videoInfo.url, options, processOptions, async (err, output) => {
155 if (err) {
156 log.error(err)
157 return res()
158 }
159
160 log.info(output.join('\n'))
161 await uploadVideoOnPeerTube({
162 cwd,
163 url,
164 user,
165 videoInfo: normalizeObject(videoInfo),
166 videoPath: path
167 })
168 return res()
169 })
170 } catch (err) {
171 log.error(err.message)
172 return res()
173 }
174 })
175 }
176
177 async function uploadVideoOnPeerTube (parameters: {
178 videoInfo: any
179 videoPath: string
180 cwd: string
181 url: string
182 user: { username: string, password: string }
183 }) {
184 const { videoInfo, videoPath, cwd, url, user } = parameters
185
186 const category = await getCategory(videoInfo.categories, url)
187 const licence = getLicence(videoInfo.license)
188 let tags = []
189 if (Array.isArray(videoInfo.tags)) {
190 tags = videoInfo.tags
191 .filter(t => t.length < CONSTRAINTS_FIELDS.VIDEOS.TAG.max && t.length > CONSTRAINTS_FIELDS.VIDEOS.TAG.min)
192 .map(t => t.normalize())
193 .slice(0, 5)
194 }
195
196 let thumbnailfile
197 if (videoInfo.thumbnail) {
198 thumbnailfile = join(cwd, sha256(videoInfo.thumbnail) + '.jpg')
199
200 await doRequestAndSaveToFile({
201 method: 'GET',
202 uri: videoInfo.thumbnail
203 }, thumbnailfile)
204 }
205
206 const originallyPublishedAt = buildOriginallyPublishedAt(videoInfo)
207
208 const defaultAttributes = {
209 name: truncate(videoInfo.title, {
210 length: CONSTRAINTS_FIELDS.VIDEOS.NAME.max,
211 separator: /,? +/,
212 omission: ' […]'
213 }),
214 category,
215 licence,
216 nsfw: isNSFW(videoInfo),
217 description: videoInfo.description,
218 tags
219 }
220
221 const videoAttributes = await buildVideoAttributesFromCommander(url, program, defaultAttributes)
222
223 Object.assign(videoAttributes, {
224 originallyPublishedAt: originallyPublishedAt ? originallyPublishedAt.toISOString() : null,
225 thumbnailfile,
226 previewfile: thumbnailfile,
227 fixture: videoPath
228 })
229
230 log.info('\nUploading on PeerTube video "%s".', videoAttributes.name)
231
232 let accessToken = await getAccessTokenOrDie(url, user)
233
234 try {
235 await uploadVideo(url, accessToken, videoAttributes)
236 } catch (err) {
237 if (err.message.indexOf('401') !== -1) {
238 log.info('Got 401 Unauthorized, token may have expired, renewing token and retry.')
239
240 accessToken = await getAccessTokenOrDie(url, user)
241
242 await uploadVideo(url, accessToken, videoAttributes)
243 } else {
244 exitError(err.message)
245 }
246 }
247
248 await remove(videoPath)
249 if (thumbnailfile) await remove(thumbnailfile)
250
251 log.warn('Uploaded video "%s"!\n', videoAttributes.name)
252 }
253
254 /* ---------------------------------------------------------- */
255
256 async function getCategory (categories: string[], url: string) {
257 if (!categories) return undefined
258
259 const categoryString = categories[0]
260
261 if (categoryString === 'News & Politics') return 11
262
263 const res = await getVideoCategories(url)
264 const categoriesServer = res.body
265
266 for (const key of Object.keys(categoriesServer)) {
267 const categoryServer = categoriesServer[key]
268 if (categoryString.toLowerCase() === categoryServer.toLowerCase()) return parseInt(key, 10)
269 }
270
271 return undefined
272 }
273
274 function getLicence (licence: string) {
275 if (!licence) return undefined
276
277 if (licence.indexOf('Creative Commons Attribution licence') !== -1) return 1
278
279 return undefined
280 }
281
282 function normalizeObject (obj: any) {
283 const newObj: any = {}
284
285 for (const key of Object.keys(obj)) {
286 // Deprecated key
287 if (key === 'resolution') continue
288
289 const value = obj[key]
290
291 if (typeof value === 'string') {
292 newObj[key] = value.normalize()
293 } else {
294 newObj[key] = value
295 }
296 }
297
298 return newObj
299 }
300
301 function fetchObject (info: any) {
302 const url = buildUrl(info)
303
304 return new Promise<any>(async (res, rej) => {
305 const youtubeDL = await safeGetYoutubeDL()
306 youtubeDL.getInfo(url, undefined, processOptions, (err, videoInfo) => {
307 if (err) return rej(err)
308
309 const videoInfoWithUrl = Object.assign(videoInfo, { url })
310 return res(normalizeObject(videoInfoWithUrl))
311 })
312 })
313 }
314
315 function buildUrl (info: any) {
316 const webpageUrl = info.webpage_url as string
317 if (webpageUrl?.match(/^https?:\/\//)) return webpageUrl
318
319 const url = info.url as string
320 if (url?.match(/^https?:\/\//)) return url
321
322 // It seems youtube-dl does not return the video url
323 return 'https://www.youtube.com/watch?v=' + info.id
324 }
325
326 function isNSFW (info: any) {
327 return info.age_limit && info.age_limit >= 16
328 }
329
330 function normalizeTargetUrl (url: string) {
331 let normalizedUrl = url.replace(/\/+$/, '')
332
333 if (!normalizedUrl.startsWith('http://') && !normalizedUrl.startsWith('https://')) {
334 normalizedUrl = 'https://' + normalizedUrl
335 }
336
337 return normalizedUrl
338 }
339
340 async function promptPassword () {
341 return new Promise<string>((res, rej) => {
342 prompt.start()
343 const schema = {
344 properties: {
345 password: {
346 hidden: true,
347 required: true
348 }
349 }
350 }
351 prompt.get(schema, function (err, result) {
352 if (err) {
353 return rej(err)
354 }
355 return res(result.password)
356 })
357 })
358 }
359
360 async function getAccessTokenOrDie (url: string, user: UserInfo) {
361 const resClient = await getClient(url)
362 const client = {
363 id: resClient.body.client_id,
364 secret: resClient.body.client_secret
365 }
366
367 try {
368 const res = await login(url, client, user)
369 return res.body.access_token
370 } catch (err) {
371 exitError('Cannot authenticate. Please check your username/password.')
372 }
373 }
374
375 function parseDate (dateAsStr: string): Date {
376 if (!/\d{4}-\d{2}-\d{2}/.test(dateAsStr)) {
377 exitError(`Invalid date passed: ${dateAsStr}. Expected format: YYYY-MM-DD. See help for usage.`)
378 }
379 const date = new Date(dateAsStr)
380 date.setHours(0, 0, 0)
381 if (isNaN(date.getTime())) {
382 exitError(`Invalid date passed: ${dateAsStr}. See help for usage.`)
383 }
384 return date
385 }
386
387 function formatDate (date: Date): string {
388 return date.toISOString().split('T')[0]
389 }
390
391 function exitError (message: string, ...meta: any[]) {
392 // use console.error instead of log.error here
393 console.error(message, ...meta)
394 process.exit(-1)
395 }