#!/usr/bin/env python3 """ FastAPI Server untuk TTS Bahasa Indonesia Berdasarkan tts_standalone.py API ini menyediakan endpoint untuk menghasilkan audio dari teks Bahasa Indonesia dengan menggunakan model TTS yang sudah dilatih khusus untuk bahasa Indonesia. Fitur: - Limit maksimal 10 file di folder outputs - Auto cleanup file lama ketika melebihi limit - Kompatibel dengan HuggingFace Spaces Endpoints: - POST /api/tts: Generate audio dari teks - GET /api/health: Health check endpoint - GET /api/speakers: Daftar speaker tersedia - GET /api/download/{file_path}: Download file audio """ from fastapi import FastAPI, HTTPException, File, UploadFile from fastapi.responses import FileResponse, JSONResponse from fastapi.middleware.cors import CORSMiddleware from pydantic import BaseModel, Field from typing import Optional import uuid from pathlib import Path import os from datetime import datetime from tts_standalone import generate_tts, params app = FastAPI( title="TTS Bahasa Indonesia API", description="API untuk Text-to-Speech Bahasa Indonesia", version="1.0.0" ) # Enable CORS app.add_middleware( CORSMiddleware, allow_origins=["*"], allow_credentials=True, allow_methods=["*"], allow_headers=["*"], ) # Direktori untuk menyimpan file audio hasil TTS OUTPUT_DIR = Path("outputs") OUTPUT_DIR.mkdir(parents=True, exist_ok=True) # Limit maksimal file yang disimpan MAX_FILES = 10 # Pydantic models untuk request/response class TTSRequest(BaseModel): text: str = Field(..., description="Teks yang akan diubah menjadi suara") speaker: Optional[str] = Field(default=None, description="Nama speaker (default: dari config)") file_path: Optional[str] = Field(default=None, description="Custom file path (optional)") class TTSResponse(BaseModel): success: bool message: str file_path: Optional[str] = None download_url: Optional[str] = None text: Optional[str] = None speaker: Optional[str] = None class HealthResponse(BaseModel): status: str message: str def cleanup_old_files(max_files: int = MAX_FILES): """ Menghapus file lama jika jumlah file melebihi max_files. File yang paling lama akan dihapus terlebih dahulu. Args: max_files: Jumlah maksimal file yang diizinkan """ try: # Get semua file .wav di OUTPUT_DIR wav_files = list(OUTPUT_DIR.glob("*.wav")) if len(wav_files) <= max_files: return # Sort berdasarkan waktu modifikasi (file lama di depan) wav_files.sort(key=lambda x: x.stat().st_mtime) # Hapus file lama sampai jumlah file <= max_files files_to_delete = wav_files[:len(wav_files) - max_files + 1] for file_path in files_to_delete: try: file_path.unlink() print(f"Deleted old file: {file_path}") except Exception as e: print(f"Error deleting file {file_path}: {e}") except Exception as e: print(f"Error in cleanup_old_files: {e}") @app.get("/api/health", response_model=HealthResponse) async def health_check(): """ Health check endpoint """ return HealthResponse( status="healthy", message="TTS API is running" ) @app.post("/api/tts", response_model=TTSResponse) async def tts_endpoint(request: TTSRequest): """ Endpoint untuk generate TTS audio dari teks Request Body (JSON): { "text": "Teks yang akan diubah menjadi suara", "speaker": "gadis", // Optional, default: dari config "file_path": "custom_filename.wav" // Optional } """ try: # Validasi text if not request.text or not request.text.strip(): raise HTTPException( status_code=400, detail="Parameter 'text' wajib diisi dan tidak boleh kosong" ) # Get speaker (gunakan default dari config jika tidak ada) speaker = request.speaker or params.get("speaker", "gadis") # Generate file path if request.file_path: # Jika user memberikan custom file path if not request.file_path.endswith(".wav"): request.file_path += ".wav" file_path = OUTPUT_DIR / request.file_path else: # Generate UUID untuk file name short_uuid = str(uuid.uuid4())[:8] file_path = OUTPUT_DIR / f"{speaker}-{short_uuid}.wav" # Cleanup file lama sebelum generate file baru cleanup_old_files(MAX_FILES) # Generate TTS success = generate_tts(request.text, str(file_path), speaker) if success: # Cleanup lagi setelah generate (untuk memastikan) cleanup_old_files(MAX_FILES) return TTSResponse( success=True, message="Audio berhasil dibuat", file_path=str(file_path), download_url=f"/api/download/{file_path}", text=request.text, speaker=speaker ) else: raise HTTPException( status_code=500, detail="Gagal membuat audio. Periksa log untuk detail." ) except HTTPException: raise except Exception as e: raise HTTPException( status_code=500, detail=f"Internal server error: {str(e)}" ) @app.get("/api/download/{file_path:path}") async def download_file(file_path: str): """ Endpoint untuk download file audio Args: file_path: Path relatif ke file audio (contoh: outputs/gadis-abc12345.wav) """ try: file_path_obj = Path(file_path) # Security check: pastikan file ada di dalam OUTPUT_DIR try: file_path_obj.resolve().relative_to(OUTPUT_DIR.resolve()) except ValueError: raise HTTPException( status_code=400, detail="Invalid file path" ) if not file_path_obj.exists(): raise HTTPException( status_code=404, detail="File tidak ditemukan" ) return FileResponse( path=str(file_path_obj), filename=file_path_obj.name, media_type="audio/wav" ) except HTTPException: raise except Exception as e: raise HTTPException( status_code=500, detail=f"Error: {str(e)}" ) @app.get("/api/speakers") async def get_speakers(): """ Endpoint untuk mendapatkan daftar speaker yang tersedia """ speakers = [ {"id": "wibowo", "name": "Wibowo - Suara jantan berwibawa"}, {"id": "ardi", "name": "Ardi - Suara lembut dan hangat"}, {"id": "gadis", "name": "Gadis - Suara perempuan yang merdu"}, {"id": "JV-00264", "name": "Juminten - Suara perempuan jawa (bahasa jawa)"}, {"id": "SU-00060", "name": "Asep - Suara lelaki sunda (bahasa sunda)"} ] return { "success": True, "speakers": speakers, "default": params.get("speaker", "gadis") } @app.get("/") async def root(): """ Root endpoint dengan informasi API """ return { "message": "TTS Bahasa Indonesia API", "version": "1.0.0", "docs": "/docs", "endpoints": { "POST /api/tts": "Generate audio dari teks", "GET /api/health": "Health check", "GET /api/speakers": "Daftar speaker tersedia", "GET /api/download/{file_path}": "Download file audio" } } if __name__ == "__main__": import uvicorn # Get port from environment variable (untuk HuggingFace Spaces) port = int(os.environ.get("PORT", 7860)) print("=" * 60) print("TTS API Server - Bahasa Indonesia (FastAPI)") print("=" * 60) print("API Endpoints:") print(" POST /api/tts - Generate audio dari teks") print(" GET /api/health - Health check") print(" GET /api/speakers - Daftar speaker tersedia") print(" GET /api/download/{file_path} - Download file audio") print(" GET /docs - Swagger UI documentation") print("=" * 60) print(f"\nServer starting on http://0.0.0.0:{port}") print(f"Max files in outputs folder: {MAX_FILES}") print("Press CTRL+C to stop\n") uvicorn.run(app, host="0.0.0.0", port=port)