Create app.py

app.py

@@ -0,0 +1,136 @@
+import os
+import uuid
+import shutil
+import secrets
+from fastapi import FastAPI, Depends, HTTPException, status, File, UploadFile, Form
+from fastapi.security import HTTPBasic, HTTPBasicCredentials
+from fastapi.responses import FileResponse, JSONResponse
+from TTS.api import TTS
+
+# --- KONFIGURASI FOLDER TEMP ---
+TEMP_DIR = "/tmp/coqui_data"
+AUDIO_DIR = "/tmp/audio_output"
+os.makedirs(TEMP_DIR, exist_ok=True)
+os.makedirs(AUDIO_DIR, exist_ok=True)
+
+# Paksa environment variable untuk simpan modul bahasa ke /tmp
+os.environ["TTS_HOME"] = TEMP_DIR
+
+app = FastAPI(
+    title="Coqui TTS API Interaktif",
+    description="API lengkap untuk Coqui TTS dengan otentikasi. Semua cache dan unduhan diarahkan ke /tmp."
+)
+security = HTTPBasic()
+
+# --- OTENTIKASI ---
+def verify_auth(credentials: HTTPBasicCredentials = Depends(security)):
+    # Username default: admin
+    correct_username = secrets.compare_digest(credentials.username, "admin")
+    correct_password = secrets.compare_digest(credentials.password, "Rahasia1234")
+    if not (correct_username and correct_password):
+        raise HTTPException(
+            status_code=status.HTTP_401_UNAUTHORIZED,
+            detail="Autentikasi gagal. Cek username dan password.",
+            headers={"WWW-Authenticate": "Basic"},
+        )
+    return credentials.username
+
+# --- HELPER FUNCTION ---
+def get_tts_instance(model_name: str) -> TTS:
+    """Load model secara dinamis (akan download ke /tmp jika belum ada)"""
+    try:
+        # Gunakan CPU jika Space gratis (tidak ada GPU)
+        return TTS(model_name=model_name, progress_bar=False, gpu=False)
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Gagal memuat model: {str(e)}")
+
+# --- ENDPOINTS ---
+
+@app.get("/")
+def root():
+    return {"message": "Coqui TTS API Aktif. Akses /docs untuk interaksi UI API."}
+
+@app.get("/models", tags=["Info"])
+def list_models(username: str = Depends(verify_auth)):
+    """Mengambil daftar semua model TTS dan Voice Conversion yang tersedia."""
+    try:
+        models = TTS().list_models()
+        return {"models": models}
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+@app.post("/tts", tags=["Generation"])
+def generate_tts(
+    text: str = Form(..., description="Teks yang akan diubah menjadi suara"),
+    model_name: str = Form("tts_models/en/ljspeech/vits", description="Nama model dari endpoint /models"),
+    speaker: str = Form(None, description="Nama speaker (jika model multi-speaker)"),
+    language: str = Form(None, description="Kode bahasa (jika model multi-lingual)"),
+    username: str = Depends(verify_auth)
+):
+    """Sintesis teks ke suara dasar."""
+    tts = get_tts_instance(model_name)
+    output_path = os.path.join(AUDIO_DIR, f"tts_{uuid.uuid4().hex}.wav")
+    
+    try:
+        tts.tts_to_file(text=text, speaker=speaker, language=language, file_path=output_path)
+        return FileResponse(output_path, media_type="audio/wav", filename="output.wav")
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Error sintesis TTS: {str(e)}")
+
+@app.post("/tts_voice_clone", tags=["Generation"])
+def generate_tts_voice_clone(
+    text: str = Form(...),
+    model_name: str = Form("tts_models/multilingual/multi-dataset/xtts_v2"),
+    language: str = Form("id", description="Kode bahasa output (misal: 'id', 'en')"),
+    reference_audio: UploadFile = File(..., description="File audio WAV singkat untuk di-clone suaranya"),
+    username: str = Depends(verify_auth)
+):
+    """Sintesis teks dengan meniru suara dari file audio referensi (Zero-shot Voice Cloning)."""
+    tts = get_tts_instance(model_name)
+    
+    # Simpan file audio referensi ke /tmp
+    ref_path = os.path.join(AUDIO_DIR, f"ref_{uuid.uuid4().hex}.wav")
+    with open(ref_path, "wb") as buffer:
+        shutil.copyfileobj(reference_audio.file, buffer)
+        
+    output_path = os.path.join(AUDIO_DIR, f"clone_{uuid.uuid4().hex}.wav")
+    
+    try:
+        tts.tts_to_file(
+            text=text, 
+            language=language, 
+            speaker_wav=ref_path, 
+            file_path=output_path
+        )
+        return FileResponse(output_path, media_type="audio/wav", filename="cloned_output.wav")
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Error Voice Cloning: {str(e)}")
+
+@app.post("/voice_conversion", tags=["Generation"])
+def voice_conversion(
+    source_audio: UploadFile = File(..., description="Audio sumber yang ingin diubah suaranya"),
+    reference_audio: UploadFile = File(..., description="Audio target/referensi (suara yang akan ditiru)"),
+    model_name: str = Form("voice_conversion_models/multilingual/vctk/freevc24"),
+    username: str = Depends(verify_auth)
+):
+    """Mengubah suara dari satu file audio menjadi suara di file audio referensi (Audio-to-Audio)."""
+    tts = get_tts_instance(model_name)
+    
+    source_path = os.path.join(AUDIO_DIR, f"src_{uuid.uuid4().hex}.wav")
+    ref_path = os.path.join(AUDIO_DIR, f"ref_vc_{uuid.uuid4().hex}.wav")
+    output_path = os.path.join(AUDIO_DIR, f"vc_{uuid.uuid4().hex}.wav")
+    
+    with open(source_path, "wb") as buffer:
+        shutil.copyfileobj(source_audio.file, buffer)
+    with open(ref_path, "wb") as buffer:
+        shutil.copyfileobj(reference_audio.file, buffer)
+        
+    try:
+        tts.voice_conversion_to_file(
+            source_wav=source_path, 
+            target_wav=ref_path, 
+            file_path=output_path
+        )
+        return FileResponse(output_path, media_type="audio/wav", filename="converted_output.wav")
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Error Voice Conversion: {str(e)}")