add url audio

app/api/transcribe.py

@@ -71,7 +71,7 @@ def _enqueue_async_job(audio_url: str, note_id: str, user_id: str | None = None)
 
 
 
-async def _run_sync_pipeline(tmp_wav: str, note_id: str):
+async def _run_sync_pipeline(tmp_wav: str, note_id: str, audio_url: str | None = None):
     """
     Sync ASR → update existing note
     """
@@ -106,48 +106,54 @@ async def _run_sync_pipeline(tmp_wav: str, note_id: str):
 
     # 🔥 UPDATE — KHÔNG CREATE
     # Use internal update endpoint to match create_audio_note (internal API)
-    await note_service.update_note_internal(
-        note_id,
-        {
-            "status": status,
-            "raw_text": text,
-            "metadata": {
-                "audio": {
-                    "duration": info.get("duration"),
-                    "sample_rate": info.get("samplerate"),
-                    "chunks": chunks,
-                    "asr_model": "PhoWhisper-base",
-                }
-            },
-            "generate": ["normalize", "keywords", "summary", "mindmap"],
+    payload = {
+        "status": status,
+        "raw_text": text,
+        "metadata": {
+            "audio": {
+                "duration": info.get("duration"),
+                "sample_rate": info.get("samplerate"),
+                "chunks": chunks,
+                "asr_model": "PhoWhisper-base",
+            }
         },
-    )
+        "generate": ["normalize", "keywords", "summary", "mindmap"],
+    }
+
+    if audio_url:
+        payload["metadata"]["audio"]["url"] = audio_url
+
+    await note_service.update_note_internal(note_id, payload)
 
     return {
         "note_id": note_id,
         "status": status,
         "duration": info.get("duration"),
+        "audio_url": audio_url,
     }
 
 
-async def _create_placeholder_note(note_id: str, duration: float):
+async def _create_placeholder_note(note_id: str, duration: float, audio_url: str | None = None):
     client = NoteServiceClient()
-    res = await client.create_audio_note(
-        {
-            "note_id": note_id,
-            "type": "audio",
-            "status": "processing",
-            "raw_text": "",
-            "metadata": {
-                "audio": {
-                    "duration": duration,
-                    "chunks": [],
-                    "asr_model": "PhoWhisper-base",
-                }
-            },
-            # ❌ KHÔNG generate ở đây
-        }
-    )
+    payload = {
+        "note_id": note_id,
+        "type": "audio",
+        "status": "processing",
+        "raw_text": "",
+        "metadata": {
+            "audio": {
+                "duration": duration,
+                "chunks": [],
+                "asr_model": "PhoWhisper-base",
+            }
+        },
+        # ❌ KHÔNG generate ở đây
+    }
+
+    if audio_url:
+        payload["metadata"]["audio"]["url"] = audio_url
+
+    res = await client.create_audio_note(payload)
 
     if res is None:
         # 🔥 FAIL FAST
@@ -183,7 +189,7 @@ async def transcribe(file: UploadFile = File(...)):
             if duration > ASYNC_THRESHOLD:
                 audio_url = await asyncio.to_thread(upload_temp_audio, tmp_wav)
 
-                await _create_placeholder_note(note_id, duration)
+                await _create_placeholder_note(note_id, duration, audio_url)
                 job = _enqueue_async_job(audio_url, note_id)
 
                 REQUEST_COUNT.labels(endpoint, "queued").inc()
@@ -194,12 +200,15 @@ async def transcribe(file: UploadFile = File(...)):
                         "job_id": job.id,
                         "status": "queued",
                         "duration": duration,
+                        "audio_url": audio_url,
                     },
                 )
 
             # ---------- SYNC ----------
-            await _create_placeholder_note(note_id, duration)
-            result = await _run_sync_pipeline(tmp_wav, note_id)
+            # Upload small audio so we can return a stable URL and persist it
+            audio_url = await asyncio.to_thread(upload_temp_audio, tmp_wav)
+            await _create_placeholder_note(note_id, duration, audio_url)
+            result = await _run_sync_pipeline(tmp_wav, note_id, audio_url)
 
             REQUEST_COUNT.labels(endpoint, "success").inc()
             return result
@@ -242,7 +251,7 @@ async def transcribe_url(payload: dict):
 
             # ---------- ASYNC ----------
             if duration > ASYNC_THRESHOLD:
-                await _create_placeholder_note(note_id, duration)
+                await _create_placeholder_note(note_id, duration, audio_url)
                 job = _enqueue_async_job(audio_url, note_id, user_id)
 
                 REQUEST_COUNT.labels(endpoint, "queued").inc()
@@ -253,12 +262,13 @@ async def transcribe_url(payload: dict):
                         "job_id": job.id,
                         "status": "queued",
                         "duration": duration,
+                        "audio_url": audio_url,
                     },
                 )
 
             # ---------- SYNC ----------
-            await _create_placeholder_note(note_id, duration)
-            result = await _run_sync_pipeline(tmp_wav, note_id)
+            await _create_placeholder_note(note_id, duration, audio_url)
+            result = await _run_sync_pipeline(tmp_wav, note_id, audio_url)
 
             REQUEST_COUNT.labels(endpoint, "success").inc()
             return result

app/schemas/transcribe.py

@@ -5,3 +5,4 @@ class TranscribeResponse(BaseModel):
     note_id: str
     status: str
     duration: Optional[float] = None
+    audio_url: Optional[str] = None