Files
usenet-indexer/commands/workers/Body.ts
T

98 lines
3.6 KiB
TypeScript

import { BaseCommand } from '@adonisjs/core/ace'
import { CommandOptions } from '@adonisjs/core/types/ace'
import { Worker } from 'bullmq'
import fs from 'node:fs/promises'
import path from 'node:path'
import queueConfig from '#config/queue'
import Shutdown from '#services/Shutdown'
import QueueService from '#services/QueueService'
import env from '#start/env'
import NntpService from '#services/NntpService'
import IngestService, { segmentFromYenc } from '#services/IngestService'
import { parseYencMeta, YencError } from '#lib/yenc/index'
import { firstChunkMd5 } from '#services/ContentService'
export default class BodyWorker extends BaseCommand {
public static commandName = 'worker:body'
public static description = 'Starts a worker to process article bodies for yEnc metadata.'
public static options: CommandOptions = {
startApp: true,
}
public async run() {
this.logger.info('Starting body worker...')
const pool = NntpService
const worker = new Worker('body-queue', async (job) => {
const { header, group, articleNumber } = job.data
const messageId = header['message-id']
this.logger.debug(`Processing header with unparsable subject: ${header.subject}`)
let conn
let failed = false
try {
conn = await pool.acquire()
const bodyBuffer = await pool.fetchBody(conn, messageId)
let meta
try {
meta = parseYencMeta(bodyBuffer)
} catch (parseError: any) {
if (parseError instanceof YencError && bodyBuffer.indexOf('=ybegin ') === -1) {
this.logger.debug(`Not a yEnc post, skipping: ${messageId}`)
return
}
this.logger.error(`Failed to parse yEnc data for message ID ${messageId}. Dumping buffer.`)
const debugDir = path.join(this.app.appRoot.pathname, 'debug')
await fs.mkdir(debugDir, { recursive: true })
const timestamp = new Date().toISOString().replace(/:/g, '-')
const dumpFile = path.join(debugDir, `body-error-${timestamp}-${messageId.replace(/[<>]/g, '')}.bin`)
await fs.writeFile(dumpFile, bodyBuffer)
this.logger.error(`Problematic body buffer saved to: ${dumpFile}`)
throw parseError
}
pool.release(conn)
conn = undefined
const segment = segmentFromYenc(header, group, meta, articleNumber)
if (segment?.number === 1) segment.md5_16k = firstChunkMd5(bodyBuffer, meta.begin.size)
if (!segment) {
this.logger.warning(`Inconsistent yEnc part numbering in ${messageId}, skipping.`)
return
}
const completed = await IngestService.ingest([segment])
this.logger.debug(
`Stored part ${segment.number}/${segment.totalParts} of "${segment.name}"` +
(completed.length > 0 ? ' (file complete)' : '')
)
} catch (error: any) {
failed = true
this.logger.error(`Error in body worker for message ID ${messageId}: ${error.message}`)
throw error
} finally {
if (conn) {
pool.release(conn, failed)
}
}
}, {
connection: queueConfig.connection,
// Each job holds one NNTP connection while downloading, so keep this <= the pool size.
concurrency: env.get('BODY_CONCURRENCY', 1),
})
worker.on('failed', (job, err) => {
this.logger.error(`Body job ${job?.id} failed: ${err.message}`)
})
this.logger.info('Body worker started and listening for jobs.')
await Shutdown.wait()
this.logger.info('Shutting down: finishing active jobs...')
await worker.close()
await NntpService.shutdown()
await QueueService.closeAll()
}
}