Atualizar app/app.py

app/app.py

@@ -341,8 +341,63 @@ def search():
     db.close()
     return jsonify(results)
 
+from pywhispercpp.model import Model
+@app.route("/api/audio/transcribe/<cid>", methods=["POST"])
+@login_required
+@require_capability("can_use_audio")
+def transcribe(cid):
+    # Initialize the model (automatically downloads 'base.en' if not present)
+    model = Model('base.en', print_realtime=False, print_progress=False)
+    
+   
+    file = request.files.get("file")
+    if not file:
+        return jsonify({"error": "No audio file"}), 400
 
+    path = f"/data/audio-{uuid.uuid4()}.wav"
+    file.save(path)
 
+    # Example using whisper (pseudo-code)
+    # result = whisper.transcribe(path)
+    # text = result["text"]
+
+    # Transcribe your audio file (must be 16kHz WAV format)
+    segments = model.transcribe(path)
+    
+    # Print the text results
+    for segment in segments:
+        print(f"[{segment.t0} -> {segment.t1}]: {segment.text}")
+
+    text = "Transcribed text placeholder"
+
+    db = SessionLocal()
+    msg = Message(
+        id=str(uuid.uuid4()),
+        conversation_id=cid,
+        role="user",
+        content=text,
+    )
+    db.add(msg)
+    db.commit()
+    db.close()
+
+    return jsonify({"status": "ok", "text": text})
+
+from gtts import gTTS
+
+@app.route("/api/audio/synthesize/<cid>", methods=["POST"])
+@login_required
+@require_capability("can_use_audio")
+def synthesize(cid):
+    text = request.json.get("text", "")
+    if not text:
+        return jsonify({"error": "No text"}), 400
+
+    tts = gTTS(text=text, lang="en")
+    path = f"/data/tts-{uuid.uuid4()}.mp3"
+    tts.save(path)
+
+    return send_file(path, mimetype="audio/mpeg", as_attachment=False)
 
 
 # ***