PeerTube/server/lib/job-queue/handlers/video-import.ts

275 lines
10 KiB
TypeScript
Raw Normal View History

2021-08-27 07:32:44 -05:00
import { Job } from 'bull'
import { move, remove, stat } from 'fs-extra'
import { getLowercaseExtension } from '@server/helpers/core-utils'
2021-02-09 04:22:42 -06:00
import { retryTransactionWrapper } from '@server/helpers/database-utils'
import { YoutubeDL } from '@server/helpers/youtube-dl'
import { isPostImportVideoAccepted } from '@server/lib/moderation'
Add support for saving video files to object storage (#4290) * Add support for saving video files to object storage * Add support for custom url generation on s3 stored files Uses two config keys to support url generation that doesn't directly go to (compatible s3). Can be used to generate urls to any cache server or CDN. * Upload files to s3 concurrently and delete originals afterwards * Only publish after move to object storage is complete * Use base url instead of url template * Fix mistyped config field * Add rudenmentary way to download before transcode * Implement Chocobozzz suggestions https://github.com/Chocobozzz/PeerTube/pull/4290#issuecomment-891670478 The remarks in question: Try to use objectStorage prefix instead of s3 prefix for your function/variables/config names Prefer to use a tree for the config: s3.streaming_playlists_bucket -> object_storage.streaming_playlists.bucket Use uppercase for config: S3.STREAMING_PLAYLISTS_BUCKETINFO.bucket -> OBJECT_STORAGE.STREAMING_PLAYLISTS.BUCKET (maybe BUCKET_NAME instead of BUCKET) I suggest to rename moveJobsRunning to pendingMovingJobs (or better, create a dedicated videoJobInfo table with a pendingMove & videoId columns so we could also use this table to track pending transcoding jobs) https://github.com/Chocobozzz/PeerTube/pull/4290/files#diff-3e26d41ca4bda1de8e1747af70ca2af642abcc1e9e0bfb94239ff2165acfbde5R19 uses a string instead of an integer I think we should store the origin object storage URL in fileUrl, without base_url injection. Instead, inject the base_url at "runtime" so admins can easily change this configuration without running a script to update DB URLs * Import correct function * Support multipart upload * Remove import of node 15.0 module stream/promises * Extend maximum upload job length Using the same value as for redundancy downloading seems logical * Use dynamic part size for really large uploads Also adds very small part size for local testing * Fix decreasePendingMove query * Resolve various PR comments * Move to object storage after optimize * Make upload size configurable and increase default * Prune webtorrent files that are stored in object storage * Move files after transcoding jobs * Fix federation * Add video path manager * Support move to external storage job in client * Fix live object storage tests Co-authored-by: Chocobozzz <me@florianbigard.com>
2021-08-17 01:26:20 -05:00
import { generateWebTorrentVideoFilename } from '@server/lib/paths'
import { Hooks } from '@server/lib/plugins/hooks'
import { ServerConfigManager } from '@server/lib/server-config-manager'
2020-09-25 09:19:35 -05:00
import { isAbleToUploadVideo } from '@server/lib/user'
Add support for saving video files to object storage (#4290) * Add support for saving video files to object storage * Add support for custom url generation on s3 stored files Uses two config keys to support url generation that doesn't directly go to (compatible s3). Can be used to generate urls to any cache server or CDN. * Upload files to s3 concurrently and delete originals afterwards * Only publish after move to object storage is complete * Use base url instead of url template * Fix mistyped config field * Add rudenmentary way to download before transcode * Implement Chocobozzz suggestions https://github.com/Chocobozzz/PeerTube/pull/4290#issuecomment-891670478 The remarks in question: Try to use objectStorage prefix instead of s3 prefix for your function/variables/config names Prefer to use a tree for the config: s3.streaming_playlists_bucket -> object_storage.streaming_playlists.bucket Use uppercase for config: S3.STREAMING_PLAYLISTS_BUCKETINFO.bucket -> OBJECT_STORAGE.STREAMING_PLAYLISTS.BUCKET (maybe BUCKET_NAME instead of BUCKET) I suggest to rename moveJobsRunning to pendingMovingJobs (or better, create a dedicated videoJobInfo table with a pendingMove & videoId columns so we could also use this table to track pending transcoding jobs) https://github.com/Chocobozzz/PeerTube/pull/4290/files#diff-3e26d41ca4bda1de8e1747af70ca2af642abcc1e9e0bfb94239ff2165acfbde5R19 uses a string instead of an integer I think we should store the origin object storage URL in fileUrl, without base_url injection. Instead, inject the base_url at "runtime" so admins can easily change this configuration without running a script to update DB URLs * Import correct function * Support multipart upload * Remove import of node 15.0 module stream/promises * Extend maximum upload job length Using the same value as for redundancy downloading seems logical * Use dynamic part size for really large uploads Also adds very small part size for local testing * Fix decreasePendingMove query * Resolve various PR comments * Move to object storage after optimize * Make upload size configurable and increase default * Prune webtorrent files that are stored in object storage * Move files after transcoding jobs * Fix federation * Add video path manager * Support move to external storage job in client * Fix live object storage tests Co-authored-by: Chocobozzz <me@florianbigard.com>
2021-08-17 01:26:20 -05:00
import { addMoveToObjectStorageJob, addOptimizeOrMergeAudioJob } from '@server/lib/video'
import { VideoPathManager } from '@server/lib/video-path-manager'
import { buildNextVideoState } from '@server/lib/video-state'
2021-02-09 04:22:42 -06:00
import { ThumbnailModel } from '@server/models/video/thumbnail'
2020-06-18 03:45:25 -05:00
import { MVideoImportDefault, MVideoImportDefaultFiles, MVideoImportVideo } from '@server/types/models/video/video-import'
import {
VideoImportPayload,
VideoImportTorrentPayload,
VideoImportTorrentPayloadType,
VideoImportYoutubeDLPayload,
VideoImportYoutubeDLPayloadType,
VideoState
} from '../../../../shared'
import { VideoImportState } from '../../../../shared/models/videos'
import { ThumbnailType } from '../../../../shared/models/videos/thumbnail.type'
2020-11-20 10:16:55 -06:00
import { getDurationFromVideoFile, getVideoFileFPS, getVideoFileResolution } from '../../../helpers/ffprobe-utils'
import { logger } from '../../../helpers/logger'
2018-08-07 02:54:36 -05:00
import { getSecureTorrentName } from '../../../helpers/utils'
import { createTorrentAndSetInfoHash, downloadWebTorrentVideo } from '../../../helpers/webtorrent'
import { VIDEO_IMPORT_TIMEOUT } from '../../../initializers/constants'
import { sequelizeTypescript } from '../../../initializers/database'
import { VideoModel } from '../../../models/video/video'
import { VideoFileModel } from '../../../models/video/video-file'
import { VideoImportModel } from '../../../models/video/video-import'
2020-06-18 03:45:25 -05:00
import { MThumbnail } from '../../../types/models/video/thumbnail'
import { federateVideoIfNeeded } from '../../activitypub/videos'
import { Notifier } from '../../notifier'
import { generateVideoMiniature } from '../../thumbnail'
2018-08-06 10:13:39 -05:00
2021-08-27 07:32:44 -05:00
async function processVideoImport (job: Job) {
const payload = job.data as VideoImportPayload
2018-08-06 10:13:39 -05:00
if (payload.type === 'youtube-dl') return processYoutubeDLImport(job, payload)
2018-08-07 02:54:36 -05:00
if (payload.type === 'magnet-uri' || payload.type === 'torrent-file') return processTorrentImport(job, payload)
2018-08-06 10:13:39 -05:00
}
// ---------------------------------------------------------------------------
export {
processVideoImport
}
// ---------------------------------------------------------------------------
2021-08-27 07:32:44 -05:00
async function processTorrentImport (job: Job, payload: VideoImportTorrentPayload) {
2018-08-06 10:13:39 -05:00
logger.info('Processing torrent video import in job %d.', job.id)
const videoImport = await getVideoImportOrDie(payload.videoImportId)
2018-08-07 02:54:36 -05:00
2018-08-06 10:13:39 -05:00
const options = {
type: payload.type,
videoImportId: payload.videoImportId
2018-08-06 10:13:39 -05:00
}
2018-08-07 02:54:36 -05:00
const target = {
torrentName: videoImport.torrentName ? getSecureTorrentName(videoImport.torrentName) : undefined,
magnetUri: videoImport.magnetUri
}
return processFile(() => downloadWebTorrentVideo(target, VIDEO_IMPORT_TIMEOUT), videoImport, options)
2018-08-06 10:13:39 -05:00
}
2021-08-27 07:32:44 -05:00
async function processYoutubeDLImport (job: Job, payload: VideoImportYoutubeDLPayload) {
2018-08-06 10:13:39 -05:00
logger.info('Processing youtubeDL video import in job %d.', job.id)
const videoImport = await getVideoImportOrDie(payload.videoImportId)
const options = {
type: payload.type,
videoImportId: videoImport.id
2018-08-06 10:13:39 -05:00
}
const youtubeDL = new YoutubeDL(videoImport.targetUrl, ServerConfigManager.Instance.getEnabledResolutions('vod'))
2021-05-11 03:54:05 -05:00
return processFile(
2021-05-11 03:54:05 -05:00
() => youtubeDL.downloadYoutubeDLVideo(payload.fileExt, VIDEO_IMPORT_TIMEOUT),
videoImport,
options
)
2018-08-06 10:13:39 -05:00
}
async function getVideoImportOrDie (videoImportId: number) {
const videoImport = await VideoImportModel.loadAndPopulateVideo(videoImportId)
if (!videoImport || !videoImport.Video) {
throw new Error('Cannot import video %s: the video import or video linked to this import does not exist anymore.')
}
2018-08-06 10:13:39 -05:00
return videoImport
}
type ProcessFileOptions = {
type: VideoImportYoutubeDLPayloadType | VideoImportTorrentPayloadType
2018-08-06 10:13:39 -05:00
videoImportId: number
}
2019-08-20 06:52:49 -05:00
async function processFile (downloader: () => Promise<string>, videoImport: MVideoImportDefault, options: ProcessFileOptions) {
let tempVideoPath: string
let videoFile: VideoFileModel
2018-12-04 09:02:49 -06:00
try {
// Download video from youtubeDL
2018-08-06 10:13:39 -05:00
tempVideoPath = await downloader()
// Get information about this video
2018-08-27 09:23:34 -05:00
const stats = await stat(tempVideoPath)
2020-09-25 09:19:35 -05:00
const isAble = await isAbleToUploadVideo(videoImport.User.id, stats.size)
2018-08-07 03:07:53 -05:00
if (isAble === false) {
throw new Error('The user video quota is exceeded with this video to import.')
}
2021-08-06 06:35:25 -05:00
const { resolution } = await getVideoFileResolution(tempVideoPath)
2018-08-03 02:27:30 -05:00
const fps = await getVideoFileFPS(tempVideoPath)
const duration = await getDurationFromVideoFile(tempVideoPath)
// Prepare video file object for creation in database
const fileExt = getLowercaseExtension(tempVideoPath)
const videoFileData = {
extname: fileExt,
2021-08-06 06:35:25 -05:00
resolution,
2018-08-07 08:17:17 -05:00
size: stats.size,
2021-08-06 06:35:25 -05:00
filename: generateWebTorrentVideoFilename(resolution, fileExt),
fps,
videoId: videoImport.videoId
}
videoFile = new VideoFileModel(videoFileData)
2019-08-20 06:52:49 -05:00
const hookName = options.type === 'youtube-dl'
? 'filter:api.video.post-import-url.accept.result'
: 'filter:api.video.post-import-torrent.accept.result'
// Check we accept this video
const acceptParameters = {
videoImport,
video: videoImport.Video,
videoFilePath: tempVideoPath,
videoFile,
user: videoImport.User
}
const acceptedResult = await Hooks.wrapFun(isPostImportVideoAccepted, acceptParameters, hookName)
if (acceptedResult.accepted !== true) {
logger.info('Refused imported video.', { acceptedResult, acceptParameters })
videoImport.state = VideoImportState.REJECTED
await videoImport.save()
throw new Error(acceptedResult.errorMessage)
}
// Video is accepted, resuming preparation
const videoWithFiles = Object.assign(videoImport.Video, { VideoFiles: [ videoFile ], VideoStreamingPlaylists: [] })
2018-11-16 09:48:17 -06:00
// To clean files if the import fails
2019-08-20 06:52:49 -05:00
const videoImportWithFiles: MVideoImportDefaultFiles = Object.assign(videoImport, { Video: videoWithFiles })
// Move file
Add support for saving video files to object storage (#4290) * Add support for saving video files to object storage * Add support for custom url generation on s3 stored files Uses two config keys to support url generation that doesn't directly go to (compatible s3). Can be used to generate urls to any cache server or CDN. * Upload files to s3 concurrently and delete originals afterwards * Only publish after move to object storage is complete * Use base url instead of url template * Fix mistyped config field * Add rudenmentary way to download before transcode * Implement Chocobozzz suggestions https://github.com/Chocobozzz/PeerTube/pull/4290#issuecomment-891670478 The remarks in question: Try to use objectStorage prefix instead of s3 prefix for your function/variables/config names Prefer to use a tree for the config: s3.streaming_playlists_bucket -> object_storage.streaming_playlists.bucket Use uppercase for config: S3.STREAMING_PLAYLISTS_BUCKETINFO.bucket -> OBJECT_STORAGE.STREAMING_PLAYLISTS.BUCKET (maybe BUCKET_NAME instead of BUCKET) I suggest to rename moveJobsRunning to pendingMovingJobs (or better, create a dedicated videoJobInfo table with a pendingMove & videoId columns so we could also use this table to track pending transcoding jobs) https://github.com/Chocobozzz/PeerTube/pull/4290/files#diff-3e26d41ca4bda1de8e1747af70ca2af642abcc1e9e0bfb94239ff2165acfbde5R19 uses a string instead of an integer I think we should store the origin object storage URL in fileUrl, without base_url injection. Instead, inject the base_url at "runtime" so admins can easily change this configuration without running a script to update DB URLs * Import correct function * Support multipart upload * Remove import of node 15.0 module stream/promises * Extend maximum upload job length Using the same value as for redundancy downloading seems logical * Use dynamic part size for really large uploads Also adds very small part size for local testing * Fix decreasePendingMove query * Resolve various PR comments * Move to object storage after optimize * Make upload size configurable and increase default * Prune webtorrent files that are stored in object storage * Move files after transcoding jobs * Fix federation * Add video path manager * Support move to external storage job in client * Fix live object storage tests Co-authored-by: Chocobozzz <me@florianbigard.com>
2021-08-17 01:26:20 -05:00
const videoDestFile = VideoPathManager.Instance.getFSVideoFileOutputPath(videoImportWithFiles.Video, videoFile)
await move(tempVideoPath, videoDestFile)
tempVideoPath = null // This path is not used anymore
// Generate miniature if the import did not created it
2019-08-15 04:53:26 -05:00
let thumbnailModel: MThumbnail
2021-02-09 04:22:42 -06:00
let thumbnailSave: object
if (!videoImportWithFiles.Video.getMiniature()) {
2021-02-16 01:50:40 -06:00
thumbnailModel = await generateVideoMiniature({
video: videoImportWithFiles.Video,
videoFile,
type: ThumbnailType.MINIATURE
})
2021-02-09 04:22:42 -06:00
thumbnailSave = thumbnailModel.toJSON()
}
// Generate preview if the import did not created it
2019-08-15 04:53:26 -05:00
let previewModel: MThumbnail
2021-02-09 04:22:42 -06:00
let previewSave: object
if (!videoImportWithFiles.Video.getPreview()) {
2021-02-16 01:50:40 -06:00
previewModel = await generateVideoMiniature({
video: videoImportWithFiles.Video,
videoFile,
type: ThumbnailType.PREVIEW
})
2021-02-09 04:22:42 -06:00
previewSave = previewModel.toJSON()
}
// Create torrent
2021-02-18 04:28:00 -06:00
await createTorrentAndSetInfoHash(videoImportWithFiles.Video, videoFile)
2021-02-09 04:22:42 -06:00
const videoFileSave = videoFile.toJSON()
2019-08-15 04:53:26 -05:00
2021-02-09 04:22:42 -06:00
const { videoImportUpdated, video } = await retryTransactionWrapper(() => {
return sequelizeTypescript.transaction(async t => {
const videoImportToUpdate = videoImportWithFiles as MVideoImportVideo
2021-02-09 04:22:42 -06:00
// Refresh video
const video = await VideoModel.load(videoImportToUpdate.videoId, t)
if (!video) throw new Error('Video linked to import ' + videoImportToUpdate.videoId + ' does not exist anymore.')
2021-02-09 04:22:42 -06:00
const videoFileCreated = await videoFile.save({ transaction: t })
2021-02-09 04:22:42 -06:00
// Update video DB object
video.duration = duration
Add support for saving video files to object storage (#4290) * Add support for saving video files to object storage * Add support for custom url generation on s3 stored files Uses two config keys to support url generation that doesn't directly go to (compatible s3). Can be used to generate urls to any cache server or CDN. * Upload files to s3 concurrently and delete originals afterwards * Only publish after move to object storage is complete * Use base url instead of url template * Fix mistyped config field * Add rudenmentary way to download before transcode * Implement Chocobozzz suggestions https://github.com/Chocobozzz/PeerTube/pull/4290#issuecomment-891670478 The remarks in question: Try to use objectStorage prefix instead of s3 prefix for your function/variables/config names Prefer to use a tree for the config: s3.streaming_playlists_bucket -> object_storage.streaming_playlists.bucket Use uppercase for config: S3.STREAMING_PLAYLISTS_BUCKETINFO.bucket -> OBJECT_STORAGE.STREAMING_PLAYLISTS.BUCKET (maybe BUCKET_NAME instead of BUCKET) I suggest to rename moveJobsRunning to pendingMovingJobs (or better, create a dedicated videoJobInfo table with a pendingMove & videoId columns so we could also use this table to track pending transcoding jobs) https://github.com/Chocobozzz/PeerTube/pull/4290/files#diff-3e26d41ca4bda1de8e1747af70ca2af642abcc1e9e0bfb94239ff2165acfbde5R19 uses a string instead of an integer I think we should store the origin object storage URL in fileUrl, without base_url injection. Instead, inject the base_url at "runtime" so admins can easily change this configuration without running a script to update DB URLs * Import correct function * Support multipart upload * Remove import of node 15.0 module stream/promises * Extend maximum upload job length Using the same value as for redundancy downloading seems logical * Use dynamic part size for really large uploads Also adds very small part size for local testing * Fix decreasePendingMove query * Resolve various PR comments * Move to object storage after optimize * Make upload size configurable and increase default * Prune webtorrent files that are stored in object storage * Move files after transcoding jobs * Fix federation * Add video path manager * Support move to external storage job in client * Fix live object storage tests Co-authored-by: Chocobozzz <me@florianbigard.com>
2021-08-17 01:26:20 -05:00
video.state = buildNextVideoState(video.state)
2021-02-09 04:22:42 -06:00
await video.save({ transaction: t })
2021-02-09 04:22:42 -06:00
if (thumbnailModel) await video.addAndSaveThumbnail(thumbnailModel, t)
if (previewModel) await video.addAndSaveThumbnail(previewModel, t)
2021-02-09 04:22:42 -06:00
// Now we can federate the video (reload from database, we need more attributes)
const videoForFederation = await VideoModel.loadAndPopulateAccountAndServerAndTags(video.uuid, t)
await federateVideoIfNeeded(videoForFederation, true, t)
2021-02-09 04:22:42 -06:00
// Update video import object
videoImportToUpdate.state = VideoImportState.SUCCESS
const videoImportUpdated = await videoImportToUpdate.save({ transaction: t }) as MVideoImportVideo
videoImportUpdated.Video = video
2021-02-09 04:22:42 -06:00
videoImportToUpdate.Video = Object.assign(video, { VideoFiles: [ videoFileCreated ] })
logger.info('Video %s imported.', video.uuid)
return { videoImportUpdated, video: videoForFederation }
}).catch(err => {
// Reset fields
if (thumbnailModel) thumbnailModel = new ThumbnailModel(thumbnailSave)
if (previewModel) previewModel = new ThumbnailModel(previewSave)
videoFile = new VideoFileModel(videoFileSave)
throw err
})
})
2021-07-30 09:51:27 -05:00
Notifier.Instance.notifyOnFinishedVideoImport({ videoImport: videoImportUpdated, success: true })
2018-12-28 06:47:17 -06:00
2019-08-15 04:53:26 -05:00
if (video.isBlacklisted()) {
const videoBlacklist = Object.assign(video.VideoBlacklist, { Video: video })
Notifier.Instance.notifyOnVideoAutoBlacklist(videoBlacklist)
} else {
2019-08-15 04:53:26 -05:00
Notifier.Instance.notifyOnNewVideoIfNeeded(video)
}
Add support for saving video files to object storage (#4290) * Add support for saving video files to object storage * Add support for custom url generation on s3 stored files Uses two config keys to support url generation that doesn't directly go to (compatible s3). Can be used to generate urls to any cache server or CDN. * Upload files to s3 concurrently and delete originals afterwards * Only publish after move to object storage is complete * Use base url instead of url template * Fix mistyped config field * Add rudenmentary way to download before transcode * Implement Chocobozzz suggestions https://github.com/Chocobozzz/PeerTube/pull/4290#issuecomment-891670478 The remarks in question: Try to use objectStorage prefix instead of s3 prefix for your function/variables/config names Prefer to use a tree for the config: s3.streaming_playlists_bucket -> object_storage.streaming_playlists.bucket Use uppercase for config: S3.STREAMING_PLAYLISTS_BUCKETINFO.bucket -> OBJECT_STORAGE.STREAMING_PLAYLISTS.BUCKET (maybe BUCKET_NAME instead of BUCKET) I suggest to rename moveJobsRunning to pendingMovingJobs (or better, create a dedicated videoJobInfo table with a pendingMove & videoId columns so we could also use this table to track pending transcoding jobs) https://github.com/Chocobozzz/PeerTube/pull/4290/files#diff-3e26d41ca4bda1de8e1747af70ca2af642abcc1e9e0bfb94239ff2165acfbde5R19 uses a string instead of an integer I think we should store the origin object storage URL in fileUrl, without base_url injection. Instead, inject the base_url at "runtime" so admins can easily change this configuration without running a script to update DB URLs * Import correct function * Support multipart upload * Remove import of node 15.0 module stream/promises * Extend maximum upload job length Using the same value as for redundancy downloading seems logical * Use dynamic part size for really large uploads Also adds very small part size for local testing * Fix decreasePendingMove query * Resolve various PR comments * Move to object storage after optimize * Make upload size configurable and increase default * Prune webtorrent files that are stored in object storage * Move files after transcoding jobs * Fix federation * Add video path manager * Support move to external storage job in client * Fix live object storage tests Co-authored-by: Chocobozzz <me@florianbigard.com>
2021-08-17 01:26:20 -05:00
if (video.state === VideoState.TO_MOVE_TO_EXTERNAL_STORAGE) {
return addMoveToObjectStorageJob(videoImportUpdated.Video)
}
// Create transcoding jobs?
2019-08-15 04:53:26 -05:00
if (video.state === VideoState.TO_TRANSCODE) {
await addOptimizeOrMergeAudioJob(videoImportUpdated.Video, videoFile, videoImport.User)
}
} catch (err) {
try {
if (tempVideoPath) await remove(tempVideoPath)
} catch (errUnlink) {
logger.warn('Cannot cleanup files after a video import error.', { err: errUnlink })
}
2018-08-03 02:27:30 -05:00
videoImport.error = err.message
if (videoImport.state !== VideoImportState.REJECTED) {
videoImport.state = VideoImportState.FAILED
}
await videoImport.save()
2021-07-30 09:51:27 -05:00
Notifier.Instance.notifyOnFinishedVideoImport({ videoImport, success: false })
throw err
}
}