Intel update

drb-c2-core/app/internal/transcription.py

@@ -28,38 +28,40 @@ async def transcribe_call(
     call_id: str,
     gcs_uri: str,
     talkgroup_name: Optional[str] = None,
-) -> Optional[str]:
+) -> tuple[Optional[str], list[dict]]:
     """
     Transcribe audio at the given GCS URI and store the result in Firestore.
 
-    Args:
-        call_id:        Firestore document ID in the 'calls' collection.
-        gcs_uri:        GCS URI of the audio file, e.g. gs://bucket/calls/xyz.mp3
-        talkgroup_name: Passed through to the intelligence layer; unused here.
-
     Returns:
-        The raw Whisper transcript string, or None if transcription failed.
+        (transcript, segments) — segments is a list of {start, end, text} dicts,
+        one per detected transmission. Empty list if transcription failed.
     """
     if not gcs_uri or not gcs_uri.startswith("gs://"):
-        return None
+        return None, []
 
     try:
-        transcript = await asyncio.to_thread(_sync_transcribe, gcs_uri)
+        transcript, segments = await asyncio.to_thread(_sync_transcribe, gcs_uri)
     except Exception as e:
         logger.warning(f"Transcription failed for call {call_id}: {e}")
-        return None
+        return None, []
 
     if transcript:
+        updates: dict = {"transcript": transcript}
+        if segments:
+            updates["segments"] = segments
         try:
-            await fstore.doc_set("calls", call_id, {"transcript": transcript})
-            logger.info(f"Transcript saved for call {call_id} ({len(transcript)} chars)")
+            await fstore.doc_set("calls", call_id, updates)
+            logger.info(
+                f"Transcript saved for call {call_id} "
+                f"({len(transcript)} chars, {len(segments)} segment(s))"
+            )
         except Exception as e:
             logger.warning(f"Could not save transcript for {call_id}: {e}")
 
-    return transcript
+    return transcript, segments
 
 
-def _sync_transcribe(gcs_uri: str) -> Optional[str]:
+def _sync_transcribe(gcs_uri: str) -> tuple[Optional[str], list[dict]]:
     """Download audio from GCS and transcribe with OpenAI Whisper."""
     from google.cloud import storage as gcs
     from google.oauth2 import service_account
@@ -99,8 +101,15 @@ def _sync_transcribe(gcs_uri: str) -> Optional[str]:
                 file=f,
                 language="en",
                 prompt=_WHISPER_PROMPT,
+                response_format="verbose_json",
             )
-        return response.text.strip() or None
+        text = response.text.strip() or None
+        segments = [
+            {"start": round(s.start, 2), "end": round(s.end, 2), "text": s.text.strip()}
+            for s in (response.segments or [])
+            if s.text.strip()
+        ]
+        return text, segments
     finally:
         try:
             os.unlink(tmp_path)