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/Events/RecordingTranscriptionUpdated.php

@@ -14,9 +14,21 @@ class RecordingTranscriptionUpdated implements ShouldBroadcastNow
     use Dispatchable, InteractsWithSockets, SerializesModels;
 
     /**
-     * Create a new event instance.
+     * Reverb's default max payload is 10KB. Keep streamed text well under that.
      */
-    public function __construct(public Recording $recording) {}
+    public const MAX_DELTA_BYTES = 8_000;
+
+    /**
+     * Create a new event instance.
+     *
+     * @param  array<string, mixed>|null  $whisperDelta
+     */
+    public function __construct(
+        public Recording $recording,
+        public ?string $transcriptDelta = null,
+        public bool $transcriptReplace = false,
+        public ?array $whisperDelta = null,
+    ) {}
 
     /**
      * Get the channels the event should broadcast on.
@@ -41,11 +53,89 @@ class RecordingTranscriptionUpdated implements ShouldBroadcastNow
      */
     public function broadcastWith(): array
     {
-        return array_merge(
+        $payload = array_merge(
             $this->recording->transcriptionStatusPayload(),
             [
                 'word_count' => $this->recording->word_count,
             ],
         );
+
+        unset($payload['transcript'], $payload['whisper']);
+
+        $delta = $this->transcriptDelta;
+
+        if ($delta !== null && $delta !== '' && strlen($delta) <= self::MAX_DELTA_BYTES) {
+            $payload['transcript_delta'] = $delta;
+            $payload['transcript_replace'] = $this->transcriptReplace;
+        }
+
+        $whisper = $this->compactWhisperDelta($this->whisperDelta);
+
+        if ($whisper !== null) {
+            $payload['whisper_delta'] = $whisper;
+        }
+
+        return $payload;
+    }
+
+    /**
+     * @param  array<string, mixed>|null  $whisper
+     * @return array<string, mixed>|null
+     */
+    private function compactWhisperDelta(?array $whisper): ?array
+    {
+        if ($whisper === null || $whisper === []) {
+            return null;
+        }
+
+        // Accumulated segment lists belong in HTTP hydrate, not on Reverb.
+        unset($whisper['segments']);
+
+        if ($whisper === []) {
+            return null;
+        }
+
+        foreach (['tokens', 'logprobs', 'words'] as $drop) {
+            $encoded = json_encode($whisper);
+
+            if (! is_string($encoded) || strlen($encoded) <= self::MAX_DELTA_BYTES) {
+                return $whisper;
+            }
+
+            $whisper = $this->dropWhisperField($whisper, $drop);
+        }
+
+        $encoded = json_encode($whisper);
+
+        if (! is_string($encoded) || strlen($encoded) > self::MAX_DELTA_BYTES) {
+            return null;
+        }
+
+        return $whisper;
+    }
+
+    /**
+     * @param  array<string, mixed>  $whisper
+     * @return array<string, mixed>
+     */
+    private function dropWhisperField(array $whisper, string $field): array
+    {
+        unset($whisper[$field]);
+
+        if (isset($whisper['segment']) && is_array($whisper['segment'])) {
+            unset($whisper['segment'][$field]);
+        }
+
+        if (isset($whisper['segments']) && is_array($whisper['segments'])) {
+            $whisper['segments'] = array_map(function (mixed $segment) use ($field): mixed {
+                if (is_array($segment)) {
+                    unset($segment[$field]);
+                }
+
+                return $segment;
+            }, $whisper['segments']);
+        }
+
+        return $whisper;
     }
 }