Stream Whisper verbose metadata live without overflowing Reverb.

Persist accumulated verbose_json on the recording and broadcast only transcript/whisper deltas so the show page can render language, segments, word timestamps, and confidence while a run is in progress.

app/Models/Recording.php

@@ -12,6 +12,7 @@ use Illuminate\Database\Eloquent\Model;
 use Illuminate\Database\Eloquent\Relations\BelongsTo;
 use Illuminate\Support\Facades\Cache;
 use Illuminate\Support\Facades\DB;
+use Illuminate\Support\Facades\Log;
 use Illuminate\Support\Facades\Storage;
 use Illuminate\Support\Str;
 use Throwable;
@@ -33,6 +34,7 @@ class Recording extends Model
         'file_size_bytes',
         'content_hash',
         'transcript',
+        'transcription_verbose',
         'transcription_status',
         'transcription_progress',
         'transcription_percent',
@@ -57,6 +59,7 @@ class Recording extends Model
             'transcription_duration_seconds' => 'integer',
             'file_size_bytes' => 'integer',
             'transcription_percent' => 'integer',
+            'transcription_verbose' => 'array',
             'user_id' => 'integer',
         ];
     }
@@ -146,7 +149,7 @@ class Recording extends Model
         ]);
 
         $recording = $this->fresh();
-        RecordingTranscriptionUpdated::dispatch($recording);
+        $this->broadcastTranscriptionUpdated();
         TranscribeRecording::dispatch($recording);
     }
 
@@ -424,7 +427,7 @@ class Recording extends Model
             'transcription_error' => 'Stopped by user',
         ])->save();
 
-        RecordingTranscriptionUpdated::dispatch($this->fresh());
+        $this->broadcastTranscriptionUpdated();
     }
 
     /**
@@ -458,7 +461,7 @@ class Recording extends Model
             'transcription_duration_seconds' => $durationSeconds,
         ])->save();
 
-        RecordingTranscriptionUpdated::dispatch($this->fresh());
+        $this->broadcastTranscriptionUpdated();
     }
 
     /**
@@ -477,7 +480,7 @@ class Recording extends Model
             'transcription_error' => $message,
         ])->save();
 
-        RecordingTranscriptionUpdated::dispatch($this->fresh());
+        $this->broadcastTranscriptionUpdated();
     }
 
     /**
@@ -501,9 +504,13 @@ class Recording extends Model
 
     /**
      * Update the live progress fields shown in the UI.
+     *
+     * @param  array<string, mixed>|null  $whisperDelta
      */
-    public function reportProgress(string $message, int $percent, string $status = 'processing', ?string $partialTranscript = null): void
+    public function reportProgress(string $message, int $percent, string $status = 'processing', ?string $partialTranscript = null, ?array $whisperDelta = null): void
     {
+        $diff = $this->transcriptBroadcastDiff($partialTranscript);
+
         $attributes = [
             'transcription_status' => $status,
             'transcription_progress' => $message,
@@ -515,17 +522,158 @@ class Recording extends Model
             $attributes['transcript'] = $partialTranscript;
         }
 
+        if ($whisperDelta !== null) {
+            $attributes['transcription_verbose'] = $this->mergeWhisperVerbose($whisperDelta);
+        }
+
         $this->forceFill($attributes)->save();
 
-        RecordingTranscriptionUpdated::dispatch($this->fresh());
+        $this->broadcastTranscriptionUpdated($diff['delta'], $diff['replace'], $whisperDelta);
     }
 
     /**
      * Broadcast the current transcription status to connected browsers.
+     *
+     * @param  array<string, mixed>|null  $whisperDelta
      */
-    public function broadcastTranscriptionUpdated(): void
+    public function broadcastTranscriptionUpdated(?string $transcriptDelta = null, bool $transcriptReplace = false, ?array $whisperDelta = null): void
     {
-        RecordingTranscriptionUpdated::dispatch($this->fresh() ?? $this);
+        try {
+            RecordingTranscriptionUpdated::dispatch(
+                $this->fresh() ?? $this,
+                $transcriptDelta,
+                $transcriptReplace,
+                $whisperDelta,
+            );
+        } catch (Throwable $e) {
+            Log::warning('Failed to broadcast transcription status', [
+                'recording_id' => $this->id,
+                'message' => $e->getMessage(),
+            ]);
+        }
+    }
+
+    /**
+     * Incremental text to push over Reverb instead of the full transcript.
+     *
+     * @return array{delta: ?string, replace: bool}
+     */
+    public function transcriptBroadcastDiff(?string $next): array
+    {
+        if ($next === null) {
+            return ['delta' => null, 'replace' => false];
+        }
+
+        $previous = (string) $this->transcript;
+
+        if ($previous === '' || ! str_starts_with($next, $previous)) {
+            return ['delta' => $next, 'replace' => true];
+        }
+
+        $delta = substr($next, strlen($previous));
+
+        return ['delta' => $delta === '' ? null : $delta, 'replace' => false];
+    }
+
+    /**
+     * Fold a streamed Whisper chunk into the stored verbose_json snapshot.
+     *
+     * @param  array<string, mixed>  $delta
+     * @return array<string, mixed>
+     */
+    public function mergeWhisperVerbose(array $delta): array
+    {
+        $verbose = is_array($this->transcription_verbose) ? $this->transcription_verbose : [];
+
+        if (isset($delta['language']) && is_string($delta['language'])) {
+            $verbose['language'] = $delta['language'];
+        }
+
+        if (isset($delta['duration']) && is_numeric($delta['duration'])) {
+            $verbose['duration'] = (float) $delta['duration'];
+        }
+
+        $segments = is_array($verbose['segments'] ?? null) ? $verbose['segments'] : [];
+
+        if (isset($delta['segments']) && is_array($delta['segments'])) {
+            $incoming = [];
+
+            foreach ($delta['segments'] as $segment) {
+                if (is_array($segment)) {
+                    $incoming[] = $segment;
+                }
+            }
+
+            if ($incoming !== []) {
+                $segments = $incoming;
+            }
+        } elseif (isset($delta['segment']) && is_array($delta['segment'])) {
+            $segments = $this->appendVerboseSegment($segments, $delta['segment']);
+        }
+
+        if (isset($delta['words']) && is_array($delta['words']) && $delta['words'] !== []) {
+            $last = $segments === [] ? null : count($segments) - 1;
+
+            if ($last === null) {
+                $segments[] = ['words' => $delta['words']];
+            } else {
+                $existing = is_array($segments[$last]['words'] ?? null) ? $segments[$last]['words'] : [];
+                $segments[$last]['words'] = array_merge($existing, $delta['words']);
+            }
+        }
+
+        $logprobs = is_array($verbose['logprobs'] ?? null) ? $verbose['logprobs'] : [];
+
+        if (isset($delta['logprobs']) && is_array($delta['logprobs'])) {
+            foreach ($delta['logprobs'] as $row) {
+                if (is_array($row)) {
+                    $logprobs[] = $row;
+                }
+            }
+        }
+
+        if ($segments !== []) {
+            $verbose['segments'] = $segments;
+        }
+
+        if ($logprobs !== []) {
+            $verbose['logprobs'] = $logprobs;
+        }
+
+        return $verbose;
+    }
+
+    /**
+     * @param  list<array<string, mixed>>  $segments
+     * @param  array<string, mixed>  $segment
+     * @return list<array<string, mixed>>
+     */
+    private function appendVerboseSegment(array $segments, array $segment): array
+    {
+        $key = $this->verboseSegmentKey($segment);
+
+        foreach ($segments as $existing) {
+            if ($this->verboseSegmentKey($existing) === $key) {
+                return $segments;
+            }
+        }
+
+        $segments[] = $segment;
+
+        return $segments;
+    }
+
+    /**
+     * @param  array<string, mixed>  $segment
+     */
+    private function verboseSegmentKey(array $segment): string
+    {
+        return implode('|', [
+            $segment['id'] ?? '',
+            $segment['start'] ?? '',
+            $segment['end'] ?? '',
+            $segment['text'] ?? '',
+        ]);
     }
 
     /**
@@ -596,6 +744,7 @@ class Recording extends Model
             'is_active' => $this->isTranscribing(),
             'has_transcript' => filled($this->transcript),
             'transcript' => $this->transcript,
+            'whisper' => $this->transcription_verbose,
             'transcribed_at' => $this->transcribed_at?->toIso8601String(),
         ];
     }