| |
| |
| |
|
|
| import os, sys, time, json, uuid, shutil |
| import uvicorn |
| from fastapi import FastAPI, Form, File, UploadFile, HTTPException |
| from fastapi.responses import HTMLResponse, FileResponse, JSONResponse |
| from fastapi.middleware.cors import CORSMiddleware |
|
|
| import torch |
| import torchaudio |
|
|
| |
| import pyarabic.araby as araby |
|
|
| |
| from TTS.api import TTS |
|
|
| |
| from resemble_enhance.enhancer.inference import enhance |
|
|
| |
| os.environ["COQUI_TOS_AGREED"] = "1" |
| os.environ.setdefault("HF_HOME", "/data/huggingface") |
|
|
| VOICE_LIB = "/data/voice_library" |
| OUTPUT_DIR = "/data/outputs" |
| HISTORY_FILE = "/data/history.json" |
|
|
| for d in [VOICE_LIB, OUTPUT_DIR]: |
| os.makedirs(d, exist_ok=True) |
|
|
| |
| device = "cuda" if torch.cuda.is_available() else "cpu" |
| print(f"[*] Pipeline running on {device.upper()} โฆ") |
|
|
| |
| print("[*] Loading Model 2: XTTS v2 โฆ") |
| xtts_engine = TTS("tts_models/multilingual/multi-dataset/xtts_v2").to(device) |
| print("[โ] XTTS Model ready.") |
|
|
| |
| def normalize_arabic_text(text: str) -> str: |
| """Pre-process text to guide the TTS model better.""" |
| text = araby.strip_tashkeel(text) |
| text = araby.normalize_ligature(text) |
| |
| text = text.replace("ุ", "ุ ").replace(".", ". ") |
| return " ".join(text.split()) |
|
|
| |
| def enhance_audio_quality(input_path: str, output_path: str): |
| """Pass XTTS output through Resemble Enhance for ElevenLabs-like quality.""" |
| print("[*] Running Model 3: Resemble Enhance (Super Resolution)...") |
| dwav, sr = torchaudio.load(input_path) |
| |
| |
| hwav, new_sr = enhance(dwav=dwav, sr=sr, device=device, nfe=32) |
| |
| |
| torchaudio.save(output_path, hwav, new_sr) |
| print("[โ] Enhancement complete.") |
|
|
| |
| def load_history(): |
| if os.path.exists(HISTORY_FILE): |
| try: return json.load(open(HISTORY_FILE)) |
| except: pass |
| return [] |
|
|
| def save_history(h): |
| json.dump(h, open(HISTORY_FILE, "w"), ensure_ascii=False, indent=2) |
|
|
| |
| app = FastAPI(title="Pro Voice Studio") |
| app.add_middleware(CORSMiddleware, allow_origins=["*"], allow_methods=["*"], allow_headers=["*"]) |
|
|
| LANGUAGES = { |
| "ar": "ุงูุนุฑุจูุฉ", "en": "English", "es": "Espaรฑol", "fr": "Franรงais", |
| "de": "Deutsch", "it": "Italiano", "pt": "Portuguรชs","ru": "ะ ัััะบะธะน", |
| "zh-cn": "ไธญๆ", "ja": "ๆฅๆฌ่ช" |
| } |
|
|
| |
| |
| |
| HTML = r"""<!DOCTYPE html> |
| <html lang="ar" dir="rtl"> |
| <head> |
| <meta charset="UTF-8"> |
| <meta name="viewport" content="width=device-width, initial-scale=1.0"> |
| <title>Pro Voice Studio | Multi-Model</title> |
| <link href="https://fonts.googleapis.com/css2?family=IBM+Plex+Mono:wght@400;600&family=IBM+Plex+Sans+Arabic:wght@300;400;600;700&display=swap" rel="stylesheet"> |
| <script src="https://cdn.tailwindcss.com"></script> |
| <script src="https://unpkg.com/react@18/umd/react.production.min.js"></script> |
| <script src="https://unpkg.com/react-dom@18/umd/react-dom.production.min.js"></script> |
| <script src="https://unpkg.com/@babel/standalone/babel.min.js"></script> |
| <style> |
| :root { --bg: #0b0c0f; --surface: #13151a; --border: #1f2330; --amber: #f5a623; --amber-dim:#a06a10; --green: #3ddc84; --red: #ff5252; --text: #e8eaf0; --muted: #6b7280; --mono: 'IBM Plex Mono', monospace; --sans: 'IBM Plex Sans Arabic', sans-serif; } |
| * { box-sizing: border-box; } body { margin: 0; background: var(--bg); color: var(--text); font-family: var(--sans); min-height: 100vh; } |
| ::-webkit-scrollbar { width: 5px; } ::-webkit-scrollbar-track { background: var(--surface); } ::-webkit-scrollbar-thumb { background: var(--border); border-radius: 4px; } |
| .card { background: var(--surface); border: 1px solid var(--border); border-radius: 12px; } .amber { color: var(--amber); } .tag { font-family: var(--mono); font-size: 10px; letter-spacing: .12em; text-transform: uppercase; color: var(--muted); } |
| input[type=range] { -webkit-appearance: none; width: 100%; height: 3px; background: var(--border); border-radius: 2px; outline: none; } input[type=range]::-webkit-slider-thumb { -webkit-appearance: none; width: 14px; height: 14px; background: var(--amber); border-radius: 50%; cursor: pointer; transition: transform .15s; } |
| select, textarea { background: var(--bg); border: 1px solid var(--border); color: var(--text); border-radius: 6px; padding: 6px 10px; font-family: var(--sans); outline: none; width: 100%;} textarea { resize: vertical; padding: 12px; } |
| .file-drop { border: 2px dashed var(--border); border-radius: 8px; padding: 16px; text-align: center; cursor: pointer; position: relative; } .file-drop:hover { border-color: var(--amber); background: rgba(245,166,35,.05); } .file-drop input { position: absolute; inset: 0; opacity: 0; cursor: pointer; } |
| .btn-primary { background: var(--amber); color: #000; font-weight: 700; border: none; border-radius: 8px; padding: 14px 24px; cursor: pointer; font-family: var(--sans); font-size: 15px; width: 100%; transition: opacity .2s; } .btn-primary:disabled { opacity: .4; cursor: not-allowed; } |
| .btn-ghost { background: transparent; border: 1px solid var(--border); color: var(--muted); border-radius: 6px; padding: 6px 12px; cursor: pointer; font-size: 12px; } .btn-ghost:hover { color: var(--text); border-color: var(--muted); } |
| .badge { display: inline-flex; align-items: center; gap: 4px; background: rgba(245,166,35,.12); border: 1px solid rgba(245,166,35,.3); color: var(--amber); border-radius: 20px; padding: 2px 10px; font-size: 11px; font-family: var(--mono); } .badge.green { background: rgba(61,220,132,.1); border-color: rgba(61,220,132,.3); color: var(--green); } .badge.purple { background: rgba(168, 85, 247, .1); border-color: rgba(168, 85, 247, .3); color: #a855f7; } |
| .waveform { display: flex; align-items: center; gap: 3px; height: 28px; } .waveform span { flex: 1; background: var(--amber); border-radius: 2px; animation: wave 1s ease-in-out infinite; opacity: .7; } @keyframes wave { 0%,100%{height:4px} 50%{height:24px} } .waveform span:nth-child(2){animation-delay:.1s} .waveform span:nth-child(3){animation-delay:.2s} .waveform span:nth-child(4){animation-delay:.3s} |
| .tab { cursor:pointer; padding:8px 16px; border-radius:6px; font-size:13px; color:var(--muted); transition:all .2s; } .tab.active { background:rgba(245,166,35,.15); color:var(--amber); } |
| audio { width:100%; accent-color:var(--amber); } |
| .param-row { display:grid; grid-template-columns:140px 1fr 48px; align-items:center; gap:12px; } .param-label { font-size:12px; color:var(--muted); font-family:var(--mono); } .param-val { font-size:13px; color:var(--amber); font-family:var(--mono); text-align:right; } |
| </style> |
| </head> |
| <body> |
| <div id="root"></div> |
| <script type="text/babel"> |
| const { useState, useEffect } = React; |
| const device = "DEVICE_PLACEHOLDER"; |
| const fmt = v => parseFloat(v).toFixed(2); |
| const apiPost = (url, body) => fetch(url, { method:"POST", body }).then(r => { if (!r.ok) return r.json().then(e => { throw new Error(e.detail || "Server error"); }); return r.json(); }); |
| |
| function Slider({ label, min, max, step, value, onChange }) { |
| return ( <div className="param-row"><span className="param-label">{label}</span><input type="range" min={min} max={max} step={step} value={value} onChange={e => onChange(parseFloat(e.target.value))} /><span className="param-val">{fmt(value)}</span></div> ); |
| } |
| |
| function FileZone({ label, file, onFile }) { |
| return ( <div><div className="tag mb-1">{label}</div><div className="file-drop"> <input type="file" accept="audio/*" onChange={e=>onFile(e.target.files[0])} /> {file ? <span style={{color:"var(--green)",fontSize:12}}>โ {file.name}</span> : <span style={{color:"var(--muted)",fontSize:12}}>ุงุณุญุจ ู
ููุงู ุฃู ุงููุฑ</span>} </div></div> ); |
| } |
| |
| function WaveAnim() { return <div className="waveform">{[1,2,3,4,5,6,7].map(i=><span key={i}/>)}</div>; } |
| |
| function App() { |
| const [tab, setTab] = useState("generate"); |
| const [text, setText] = useState(""); const [lang, setLang] = useState("ar"); const [file1, setFile1] = useState(null); |
| const [temperature, setTemp] = useState(0.75); const [speed, setSpeed] = useState(1.0); |
| const [useEnhancer, setUseEnhancer] = useState(false); |
| const [status, setStatus] = useState("idle"); const [statusMsg, setStatusMsg] = useState(""); const [audioUrl, setAudioUrl] = useState(null); |
| const [history, setHistory] = useState([]); const [voices, setVoices] = useState([]); const [selVoice, setSelVoice] = useState(null); |
| |
| useEffect(() => { fetch("/history").then(r=>r.json()).then(setHistory).catch(()=>{}); fetch("/voices").then(r=>r.json()).then(setVoices).catch(()=>{}); }, []); |
| |
| const generate = async () => { |
| if (!text.trim()) return setStatusMsg("ุฃุฏุฎู ุงููุต."); |
| if (!file1 && !selVoice) return setStatusMsg("ูุฌุจ ุชุญุฏูุฏ ุตูุช."); |
| setStatus("running"); setStatusMsg(useEnhancer ? "ุฌุงุฑู ุงูุชูููุฏ... (ุงูุชูุงู
ู ู
ูุนูุ ูุฏ ูุณุชุบุฑู ููุชุงู ุทูููุงู)" : ""); setAudioUrl(null); |
| const fd = new FormData(); |
| fd.append("text", text); fd.append("language", lang); fd.append("temperature", temperature); fd.append("speed", speed); |
| fd.append("enhance_audio", useEnhancer); |
| if (file1) fd.append("files", file1); if (selVoice) fd.append("voice_name", selVoice); |
| try { const data = await apiPost("/generate", fd); setAudioUrl(`/audio/${data.filename}`); setStatus("done"); } |
| catch(e) { setStatus("error"); setStatusMsg(e.message); } |
| }; |
| |
| const isRTL = ["ar"].includes(lang); |
| |
| return ( |
| <div style={{maxWidth:720,margin:"0 auto",padding:"24px 16px"}}> |
| <div style={{marginBottom:28,textAlign:"center"}}> |
| <div className="tag" style={{marginBottom:6}}>Multi-Model Pipeline</div> |
| <h1 style={{margin:0,fontSize:26,fontWeight:700,letterSpacing:"-.02em"}}><span className="amber">Pro Voice</span> Studio</h1> |
| <div style={{marginTop:8,display:"flex",justifyContent:"center",gap:6}}> |
| <span className={`badge ${device==="cuda"?"green":"red"}`}>โ {device.toUpperCase()}</span> |
| <span className="badge purple">ElevenLabs-Like Pipeline</span> |
| </div> |
| </div> |
| |
| {tab==="generate" && ( |
| <div style={{display:"flex",flexDirection:"column",gap:14}}> |
| <div className="card" style={{padding:16}}><div style={{display:"flex",justifyContent:"space-between",alignItems:"center",marginBottom:10}}><div className="tag">ุงููุต (ูุชู
ู
ุนุงูุฌุชู ุจู NLP)</div><select value={lang} onChange={e=>setLang(e.target.value)} style={{width:'auto'}}>{Object.entries(LANGUAGES_JSON).map(([k,v])=><option key={k} value={k}>{v}</option>)}</select></div><textarea dir={isRTL?"rtl":"ltr"} rows={4} value={text} onChange={e=>setText(e.target.value)} /></div> |
| <div className="card" style={{padding:16}}><div className="tag" style={{marginBottom:10}}>ุจุตู
ุฉ ุงูุตูุช</div><FileZone label="ุนููุฉ" file={file1} onFile={setFile1} /> |
| {voices.length>0 && <div style={{marginTop:12,display:"flex",gap:8}}>{voices.map(v=>( <div key={v} style={{padding:"6px 14px",borderRadius:20,cursor:"pointer",fontSize:12,border:"1px solid",borderColor:selVoice===v?"var(--amber)":"var(--border)",color:selVoice===v?"var(--amber)":"var(--muted)"}} onClick={()=>setSelVoice(selVoice===v?null:v)}>{v}</div> ))}</div>} |
| </div> |
| <div className="card" style={{padding:16, border:"1px solid rgba(168, 85, 247, 0.4)", background:"rgba(168, 85, 247, 0.05)"}}> |
| <label style={{display:"flex", alignItems:"center", gap:10, cursor:"pointer"}}> |
| <input type="checkbox" checked={useEnhancer} onChange={e=>setUseEnhancer(e.target.checked)} style={{width:18, height:18, accentColor:"#a855f7"}}/> |
| <div> |
| <div style={{fontWeight:600, color:"#a855f7"}}>ุชูุนูู ุฌูุฏุฉ ุงูุงุณุชูุฏูู (Super-Resolution)</div> |
| <div style={{fontSize:11, color:"var(--muted)", marginTop:4}}>ูุณุชุฎุฏู
ูู
ูุฐุฌ Resemble Enhance ูุชูููุฉ ุงูุตูุช ูุฑูุน ุฏูุชู ูู 44.1kHz (ูุชุทูุจ GPU).</div> |
| </div> |
| </label> |
| </div> |
| {statusMsg && <div style={{padding:"10px 14px",borderRadius:8,fontSize:13,background:"rgba(255,82,82,.08)",color:"var(--red)"}}>{statusMsg}</div>} |
| <button className="btn-primary" onClick={generate} disabled={status==="running"}>{status==="running" ? <span style={{display:"flex",justifyContent:"center",gap:10}}><WaveAnim/> ุฌุงุฑู ุงูู
ุนุงูุฌุฉ ุงูู
ุนู
ุงุฑูุฉโฆ</span> : "โก ุชูููุฏ ุงูุตูุช"}</button> |
| |
| {audioUrl && ( |
| <div className="card" style={{padding:16}}> |
| <audio src={audioUrl} controls autoPlay /> |
| <div style={{marginTop:12}}><a href={audioUrl} download><button className="btn-ghost" style={{width:'100%'}}>โฌ ุชุญู
ูู ุงูู
ูู ุงูููุงุฆู</button></a></div> |
| </div> |
| )} |
| </div> |
| )} |
| </div> |
| ); |
| } |
| const LANGUAGES_JSON = {"ar": "ุงูุนุฑุจูุฉ", "en": "English"}; |
| ReactDOM.createRoot(document.getElementById("root")).render(<App/>); |
| </script> |
| </body> |
| </html> |
| """ |
|
|
| |
| |
| |
|
|
| @app.get("/", response_class=HTMLResponse) |
| async def ui(): |
| return HTML.replace("DEVICE_PLACEHOLDER", device) |
|
|
| @app.post("/generate") |
| async def generate( |
| text: str = Form(...), |
| language: str = Form("ar"), |
| temperature: float = Form(0.75), |
| speed: float = Form(1.0), |
| enhance_audio: str = Form("false"), |
| voice_name: str = Form(None), |
| files: list[UploadFile] = File(default=[]), |
| ): |
| if not text.strip(): raise HTTPException(400, "ุงููุต ูุงุฑุบ.") |
| |
| |
| processed_text = normalize_arabic_text(text) if language == "ar" else text |
|
|
| |
| ref_paths, tmp_files = [], [] |
| for f in files: |
| path = f"/tmp/ref_{uuid.uuid4().hex}_{f.filename}" |
| with open(path, "wb") as buf: shutil.copyfileobj(f.file, buf) |
| ref_paths.append(path) |
| tmp_files.append(path) |
|
|
| if voice_name: |
| lib_dir = os.path.join(VOICE_LIB, voice_name) |
| if os.path.isdir(lib_dir): |
| ref_paths += [os.path.join(lib_dir, fn) for fn in os.listdir(lib_dir) if fn.endswith((".wav", ".mp3"))] |
|
|
| if not ref_paths: raise HTTPException(400, "ูุฌุจ ุชุญุฏูุฏ ุนููุฉ ุตูุชูุฉ.") |
|
|
| |
| base_out_name = f"gen_base_{uuid.uuid4().hex[:8]}.wav" |
| base_out_path = os.path.join(OUTPUT_DIR, base_out_name) |
| final_out_name = base_out_name |
| |
| try: |
| xtts_engine.tts_to_file( |
| text=processed_text, speaker_wav=ref_paths, language=language, |
| file_path=base_out_path, temperature=float(temperature), speed=float(speed) |
| ) |
| |
| |
| if enhance_audio.lower() == "true": |
| enhanced_name = f"gen_pro_{uuid.uuid4().hex[:8]}.wav" |
| enhanced_path = os.path.join(OUTPUT_DIR, enhanced_name) |
| enhance_audio_quality(base_out_path, enhanced_path) |
| final_out_name = enhanced_name |
|
|
| finally: |
| for p in tmp_files: |
| try: os.remove(p) |
| except: pass |
|
|
| hist = load_history() |
| hist.append({"filename": final_out_name, "text": processed_text[:100], "ts": int(time.time())}) |
| save_history(hist) |
| return {"filename": final_out_name} |
|
|
| @app.get("/audio/{filename}") |
| def get_audio(filename: str): |
| path = os.path.join(OUTPUT_DIR, filename) |
| if not os.path.exists(path): raise HTTPException(404, "Not found") |
| return FileResponse(path, media_type="audio/wav") |
|
|
| @app.get("/history") |
| def get_history(): return JSONResponse(load_history()) |
|
|
| @app.get("/voices") |
| def list_voices(): return [d for d in os.listdir(VOICE_LIB) if os.path.isdir(os.path.join(VOICE_LIB, d))] |
|
|
| if __name__ == "__main__": |
| uvicorn.run(app, host="0.0.0.0", port=7860, log_level="info") |
