chore: add performance markers

packages/tests/src/transcription/transcribers.spec.ts

@@ -4,7 +4,8 @@ import { expect } from 'chai'
 import { existsSync } from 'node:fs'
 import { rm, mkdir, readFile } from 'node:fs/promises'
 import { buildAbsoluteFixturePath, root } from '@peertube/peertube-node-utils'
-import { transcriberFactory } from '@peertube/peertube-transcription'
+import { toHumanReadable, transcriberFactory } from '@peertube/peertube-transcription'
+import { performance, PerformanceObserver } from 'node:perf_hooks'
 
 describe('Transcribers', function () {
   const transcriptDirectory = join(root(), 'test-transcript')
@@ -17,6 +18,13 @@ describe('Transcribers', function () {
 
   before(async function () {
     await mkdir(transcriptDirectory, { recursive: true })
+
+    const performanceObserver = new PerformanceObserver((items) => {
+      items
+        .getEntries()
+        .forEach((entry) => console.log(`Transcription ${entry.name} took ${toHumanReadable(entry.duration)}`))
+    })
+    performanceObserver.observe({ type: 'measure' })
   })
 
   transcribers.forEach(function (transcriberName) {
@@ -55,5 +63,6 @@ describe('Transcribers', function () {
 
   after(async function () {
     await rm(transcriptDirectory, { recursive: true, force: true })
+    performance.clearMarks()
   })
 })

packages/transcription/src/abstract-transcriber.ts

@@ -5,6 +5,9 @@ import { TranscriptionEngine } from './transcription-engine.js'
 import { TranscriptionModel } from './transcription-model.js'
 import { Transcript, TranscriptFormat } from './transcript.js'
 import { existsSync } from 'fs'
+import { PerformanceObserver } from 'node:perf_hooks'
+import short from 'short-uuid'
+import assert from 'node:assert'
 
 export abstract class AbstractTranscriber {
   public static DEFAULT_TRANSCRIPT_DIRECTORY = join(root(), 'dist', 'transcripts')
@@ -12,15 +15,19 @@ export abstract class AbstractTranscriber {
   engine: TranscriptionEngine
   logger: Logger
   transcriptDirectory: string
+  performanceObserver?: PerformanceObserver
+  runId?: string
 
   constructor (
     engine: TranscriptionEngine,
     logger: Logger,
-    transcriptDirectory: string = AbstractTranscriber.DEFAULT_TRANSCRIPT_DIRECTORY
+    transcriptDirectory: string = AbstractTranscriber.DEFAULT_TRANSCRIPT_DIRECTORY,
+    performanceObserver?: PerformanceObserver
   ) {
     this.engine = engine
     this.logger = logger
     this.transcriptDirectory = transcriptDirectory
+    this.performanceObserver = performanceObserver
   }
 
   detectLanguage () {
@@ -35,5 +42,44 @@ export abstract class AbstractTranscriber {
     return model.format === 'PyTorch'
   }
 
+  createPerformanceMark () {
+    this.runId = `${short.uuid()}-${this.engine.name}`
+    performance.mark(this.getStartPerformanceMarkName())
+  }
+
+  measurePerformanceMark () {
+    try {
+      performance.mark(this.getEndPerformanceMarkName())
+      performance.measure(
+        this.runId,
+        this.getStartPerformanceMarkName(),
+        this.getEndPerformanceMarkName()
+      )
+    } catch (e) {
+      this.logger.log({ level: 'error', message: e })
+    }
+  }
+
+  getStartPerformanceMarkName () {
+    assert(!!this.runId, 'Each transcription run should have an id.')
+
+    return `${this.runId}-started`
+  }
+
+  getEndPerformanceMarkName () {
+    assert(!!this.runId, 'Each transcription run should have an id.')
+
+    return `${this.runId}-ended`
+  }
+
+  perf () {
+    // const transcriptionPerformanceObserver = new PerformanceObserver((items) => {
+    //   items
+    //     .getEntries()
+    //     .forEach((entry) => logger.debug(`Transcription n°${entry.name} took ${toHumanReadable(entry.duration)}`, entry))
+    //   performance.clearMarks()
+    // })
+  }
+
   abstract transcribe (mediaFilePath: string, model: TranscriptionModel, language: string, format: TranscriptFormat): Promise<Transcript>
 }

packages/transcription/src/duration.spec.ts

@@ -0,0 +1,22 @@
+import { toHumanReadable, toTimecode } from './duration';
+
+describe('duration', () => {
+  test('toHumanReadable', async () => {
+    const ONE_MINUTE = 60000;
+    let humanDuration = toHumanReadable(ONE_MINUTE);
+    expect(humanDuration).toEqual('1m');
+
+    humanDuration = toHumanReadable(ONE_MINUTE * 60 + ONE_MINUTE);
+    expect(humanDuration).toEqual('1h 1m');
+  });
+
+  test('toTimecode', async () => {
+    const MORE_OR_LESS_ONE_MINUTE = '60.41545';
+    let timecode = toTimecode(MORE_OR_LESS_ONE_MINUTE);
+    expect(timecode).toEqual('00:01:00');
+
+    const ONE_HOUR = '3600';
+    timecode = toTimecode(ONE_HOUR);
+    expect(timecode).toEqual('01:00:00');
+  });
+});

packages/transcription/src/duration.ts

@@ -0,0 +1,35 @@
+export interface DurationDescriptor {
+  duration: number
+  unit: string
+}
+
+export function toHumanReadable (ms: number) {
+  const date = new Date(ms)
+
+  const durationDescriptors: DurationDescriptor[] = [
+    { duration: date.getUTCHours(), unit: 'h' },
+    { duration: date.getUTCMinutes(), unit: 'm' },
+    { duration: date.getUTCSeconds(), unit: 's' }
+  ]
+
+  return durationDescriptors
+    .map(toWords)
+    .filter((words) => words)
+    .join(' ')
+}
+
+export function toWords ({ duration, unit }: DurationDescriptor) {
+  return duration > 0 ? `${duration}${unit}` : ''
+}
+
+export function toTimecode (s: number | string) {
+  const date = new Date(0, 0, 0, 0, 0, parseFloat(s.toString()), 0)
+  const hours = date.getHours()
+  const minutes = date.getMinutes()
+  const seconds = date.getSeconds()
+  return `${padLeft(hours)}:${padLeft(minutes)}:${padLeft(seconds)}`
+}
+
+function padLeft (value: number, length = 2): string {
+  return value.toString().padStart(length, '0')
+}

packages/transcription/src/index.ts

@@ -1,6 +1,8 @@
 import { TranscriberFactory } from './transcriber-factory.js'
 import { engines } from './whisper/index.js'
 
+export * from './duration.js'
+
 export * from './transcription-engine.js'
 export * from './transcription-model.js'
 export * from './transcript.js'

packages/transcription/src/whisper/transcriber/ctranslate2-transcriber.ts

@@ -12,6 +12,7 @@ export class Ctranslate2Transcriber extends AbstractTranscriber {
     language: string,
     format: TranscriptFormat = 'vtt'
   ): Promise<Transcript> {
+    this.createPerformanceMark()
     const $$ = $({ verbose: true })
     const { baseName } = getFileInfo(mediaFilePath)
 
@@ -25,7 +26,7 @@ export class Ctranslate2Transcriber extends AbstractTranscriber {
       this.transcriptDirectory
     ]}`
 
-    await $$`ls ${this.transcriptDirectory}`
+    this.measurePerformanceMark()
 
     return {
       language,

packages/transcription/src/whisper/transcriber/openai-transcriber.ts

@@ -12,12 +12,13 @@ export class OpenaiTranscriber extends AbstractTranscriber {
     language: string,
     format: TranscriptFormat = 'vtt'
   ): Promise<Transcript> {
+    this.createPerformanceMark()
     // Shall we run the command with `{ shell: true }` to get the same error as in sh ?
     // ex: ENOENT => Command not found
     const $$ = $({ verbose: true })
     const { baseName } = getFileInfo(mediaFilePath)
 
-    const { stdout } = await $$`whisper ${[
+    await $$`whisper ${[
       mediaFilePath,
       '--model',
       model.name,
@@ -26,10 +27,8 @@ export class OpenaiTranscriber extends AbstractTranscriber {
       '--output_dir',
       this.transcriptDirectory
     ]}`
-    console.log(stdout)
 
-    const { stdout: lsStdout } = await $$`ls ${this.transcriptDirectory}`
-    console.log(lsStdout)
+    this.measurePerformanceMark()
 
     return {
       language,

packages/transcription/src/whisper/transcriber/timestamped-transcriber.ts

@@ -15,6 +15,8 @@ export class WhisperTimestampedTranscriber extends AbstractTranscriber {
     language: string,
     format: TranscriptFormat = 'vtt'
   ): Promise<Transcript> {
+    this.createPerformanceMark()
+
     const $$ = $({ verbose: true })
     const { baseName, name } = getFileInfo(mediaFilePath)
     await $$`whisper_timestamped ${[
@@ -30,9 +32,9 @@ export class WhisperTimestampedTranscriber extends AbstractTranscriber {
     const internalTranscriptPath = join(this.transcriptDirectory, `${name}.${format}`)
     const transcriptPath = join(this.transcriptDirectory, `${baseName}.${format}`)
     assert(existsSync(internalTranscriptPath), '')
-
     await rename(internalTranscriptPath, transcriptPath)
-    await $$`ls ${this.transcriptDirectory}`
+
+    this.measurePerformanceMark()
 
     return {
       language,