Ship FrankenPHP Compose services with Reverb WebSockets for real-time transcription status, skip re-uploading identical audio via content hash, and format disk usage without requiring intl.
192 lines
5.9 KiB
PHP
192 lines
5.9 KiB
PHP
<?php
|
|
|
|
namespace App\Jobs;
|
|
|
|
use App\Models\Recording;
|
|
use App\Services\TranscriptionService;
|
|
use Illuminate\Contracts\Queue\ShouldQueue;
|
|
use Illuminate\Foundation\Queue\Queueable;
|
|
use Illuminate\Support\Facades\Log;
|
|
use Throwable;
|
|
|
|
class TranscribeRecording implements ShouldQueue
|
|
{
|
|
use Queueable;
|
|
|
|
/**
|
|
* The number of times the job may be attempted.
|
|
*/
|
|
public int $tries = 1;
|
|
|
|
/**
|
|
* The number of seconds the job can run before timing out.
|
|
*/
|
|
public int $timeout;
|
|
|
|
/**
|
|
* ISO-8601 transcription_started_at this job owns (ignored after cancel/restart).
|
|
*/
|
|
public ?string $runStartedAt = null;
|
|
|
|
/**
|
|
* Create a new job instance.
|
|
*/
|
|
public function __construct(public Recording $recording)
|
|
{
|
|
$this->timeout = max(60, (int) config('ai.transcription_timeout', 600));
|
|
$this->runStartedAt = $recording->transcription_started_at?->toIso8601String();
|
|
}
|
|
|
|
/**
|
|
* Execute the job.
|
|
*/
|
|
public function handle(TranscriptionService $transcription): void
|
|
{
|
|
$recording = Recording::query()->find($this->recording->id);
|
|
|
|
if ($recording === null) {
|
|
return;
|
|
}
|
|
|
|
$this->recording = $recording;
|
|
|
|
if ($this->runStartedAt !== null && ! $this->recording->ownsTranscriptionRun($this->runStartedAt)) {
|
|
return;
|
|
}
|
|
|
|
if ($this->runStartedAt === null && ! $this->recording->isTranscribing()) {
|
|
return;
|
|
}
|
|
|
|
$this->recording->forceFill([
|
|
'transcription_status' => 'processing',
|
|
'transcription_started_at' => $this->recording->transcription_started_at ?? now(),
|
|
'transcription_error' => null,
|
|
])->save();
|
|
|
|
$this->runStartedAt ??= $this->recording->transcription_started_at?->toIso8601String();
|
|
|
|
$this->reportIfOwned('Preparing audio file…', 15);
|
|
|
|
try {
|
|
$text = $transcription->transcribe(
|
|
$this->recording,
|
|
function (string $message, int $percent): void {
|
|
$this->reportIfOwned($message, $percent);
|
|
},
|
|
);
|
|
|
|
if (! $this->claimSuccessfulTranscript($text)) {
|
|
return;
|
|
}
|
|
} catch (Throwable $e) {
|
|
if (! $this->recording->ownsTranscriptionRun($this->runStartedAt)) {
|
|
Log::warning('Transcription exception ignored after run was superseded', [
|
|
'recording_id' => $this->recording->id,
|
|
'message' => $e->getMessage(),
|
|
]);
|
|
|
|
return;
|
|
}
|
|
|
|
Log::error('Transcription failed', [
|
|
'recording_id' => $this->recording->id,
|
|
'driver' => $this->recording->transcription_driver,
|
|
'message' => $e->getMessage(),
|
|
'exception' => $e::class,
|
|
]);
|
|
|
|
$this->recording->markTranscriptionFailed($e->getMessage());
|
|
|
|
throw $e;
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Handle a job failure (timeouts, worker kill, etc.).
|
|
*/
|
|
public function failed(?Throwable $e): void
|
|
{
|
|
$recording = Recording::query()->find($this->recording->id);
|
|
|
|
if ($recording === null || ! $recording->ownsTranscriptionRun($this->runStartedAt)) {
|
|
return;
|
|
}
|
|
|
|
$recording->markTranscriptionFailed(
|
|
$e?->getMessage() ?: 'Transcription stopped unexpectedly.',
|
|
);
|
|
}
|
|
|
|
/**
|
|
* Persist a finished transcript when this job still owns the run.
|
|
*
|
|
* Also recovers runs that were falsely marked failed by orphan detection
|
|
* while Whisper was still working.
|
|
*/
|
|
private function claimSuccessfulTranscript(string $text): bool
|
|
{
|
|
$this->recording->refresh();
|
|
|
|
if ($this->recording->ownsTranscriptionRun($this->runStartedAt)) {
|
|
$this->reportIfOwned('Saving transcript…', 90);
|
|
$this->recording->refresh();
|
|
}
|
|
|
|
if ($this->recording->ownsTranscriptionRun($this->runStartedAt)) {
|
|
$this->recording->forceFill([
|
|
'transcript' => $text,
|
|
'transcription_status' => 'done',
|
|
'transcription_progress' => 'Transcription complete',
|
|
'transcription_percent' => 100,
|
|
'transcription_error' => null,
|
|
'transcribed_at' => now(),
|
|
])->save();
|
|
|
|
$this->recording->broadcastTranscriptionUpdated();
|
|
|
|
return true;
|
|
}
|
|
|
|
// Same run was wrongly marked failed as "orphaned" while Whisper was still running.
|
|
if (
|
|
$this->recording->transcription_status === 'failed'
|
|
&& $this->recording->matchesTranscriptionRun($this->runStartedAt)
|
|
&& str_contains((string) $this->recording->transcription_error, 'worker stopped')
|
|
) {
|
|
Log::warning('Recovering transcript after false orphan failure', [
|
|
'recording_id' => $this->recording->id,
|
|
]);
|
|
|
|
$this->recording->forceFill([
|
|
'transcript' => $text,
|
|
'transcription_status' => 'done',
|
|
'transcription_progress' => 'Transcription complete',
|
|
'transcription_percent' => 100,
|
|
'transcription_error' => null,
|
|
'transcribed_at' => now(),
|
|
])->save();
|
|
|
|
$this->recording->broadcastTranscriptionUpdated();
|
|
|
|
return true;
|
|
}
|
|
|
|
Log::info('Discarding transcript because transcription run was superseded', [
|
|
'recording_id' => $this->recording->id,
|
|
'status' => $this->recording->transcription_status,
|
|
]);
|
|
|
|
return false;
|
|
}
|
|
|
|
private function reportIfOwned(string $message, int $percent): void
|
|
{
|
|
if (! $this->recording->ownsTranscriptionRun($this->runStartedAt)) {
|
|
return;
|
|
}
|
|
|
|
$this->recording->reportProgress($message, $percent);
|
|
}
|
|
}
|