Update app.py

# -*- coding: utf-8 -*-
import os, re, time, random, zipfile
import requests
import gradio as gr
from pydub import AudioSegment
import natsort

# ================= CONFIG =================
VOICE_DIR = "voices"
os.makedirs(VOICE_DIR, exist_ok=True)

# ================= CORE FUNCTIONS =================
def check_api_key(api_key):
try:
r = requests.get(
"https://api.elevenlabs.io/v1/user",
headers={"xi-api-key": api_key},
timeout=10
)
if r.status_code == 200:
sub = r.json().get("subscription", {})
return {
"valid": True,
"remaining": sub.get("character_limit", 0) - sub.get("character_count", 0)
}
return {"valid": False}
except:
return {"valid": False}


def parse_text_blocks(text, max_len=200):
blocks, cur = [], ""
for s in re.split(r'(?<=[.!?])\s+', text):
if len(cur) + len(s) <= max_len:
cur += " " + s
else:
blocks.append(cur.strip())
cur = s
if cur:
blocks.append(cur.strip())
return blocks


def estimate_credit(text):
return len(text) + 50


def generate_voice(text, api_key, voice_id, model_id,
stability, similarity, style, speed, boost):
time.sleep(random.uniform(1, 2))
url = f"https://api.elevenlabs.io/v1/text-to-speech/{voice_id}"
headers = {"xi-api-key": api_key, "Content-Type": "application/json"}

payload = {
"text": text,
"model_id": model_id,
"voice_settings": {
"stability": stability,
"similarity_boost": similarity,
"style": style,
"speed": speed,
"use_speaker_boost": boost
}
}

r = requests.post(url, headers=headers, json=payload, timeout=30)
if r.status_code == 200:
return r.content
return None


def merge_audio_files(fmt):
files = natsort.natsorted(
[f for f in os.listdir(VOICE_DIR) if f.endswith(fmt.lower())]
)
if not files:
return None

combined = AudioSegment.from_file(os.path.join(VOICE_DIR, files[0]))
for f in files[1:]:
combined += AudioSegment.silent(500)
combined += AudioSegment.from_file(os.path.join(VOICE_DIR, f))

out = f"output_full.{fmt.lower()}"
combined.export(out, format=fmt.lower())
return out


# ================= MAIN RUN =================
def run_tts(
api_keys_text,
voice_id,
text,
model_id,
fmt,
stability,
similarity,
style,
speed,
boost
):
for f in os.listdir(VOICE_DIR):
os.remove(os.path.join(VOICE_DIR, f))

api_keys = [k.strip() for k in api_keys_text.splitlines() if k.strip()]
valid_keys = []

for k in api_keys:
info = check_api_key(k)
if info.get("valid") and info["remaining"] > 500:
valid_keys.append([k, info["remaining"]])

if not valid_keys:
return "❌ Không có API key hợp lệ", None

texts = parse_text_blocks(text)
key_index = 0

for i, block in enumerate(texts):
success = False
while valid_keys:
key, remain = valid_keys[key_index]
need = estimate_credit(block)

if remain < need:
valid_keys.pop(key_index)
continue

audio = generate_voice(
block, key, voice_id, model_id,
stability, similarity, style, speed, boost
)

if audio:
path = f"{VOICE_DIR}/voice_{i+1:03d}.{fmt.lower()}"
with open(path, "wb") as f:
f.write(audio)
valid_keys[key_index][1] -= need
success = True
break
else:
valid_keys.pop(key_index)

if not success:
return "❌ Hết API key khi đang chạy", None

merged = merge_audio_files(fmt)
return "✅ Hoàn tất", merged


# ================= GRADIO UI =================
with gr.Blocks(title="ElevenLabs TTS PRO") as demo:
gr.Markdown("## 🔊 ElevenLabs TTS – Hugging Face Edition")

api_keys = gr.Textbox(label="API Keys (mỗi dòng 1 key)", lines=4)
voice_id = gr.Textbox(label="Voice ID")
text = gr.Textbox(label="Text", lines=6)

model = gr.Dropdown(
choices=[
("Turbo v2.5", "eleven_turbo_v2_5"),
("Flash v2.5", "eleven_flash_v2_5"),
("Multilingual v2", "eleven_multilingual_v2"),
],
value="eleven_multilingual_v2",
label="Model"
)

fmt = gr.Dropdown(["MP3", "WAV"], value="MP3", label="Format")

stability = gr.Slider(0, 1, 0.9, label="Stability")
similarity = gr.Slider(0, 1, 0.5, label="Similarity")
style = gr.Slider(0, 1, 0.4, label="Style")
speed = gr.Slider(0.7, 1.2, 0.81, label="Speed")
boost = gr.Checkbox(True, label="Speaker Boost")

run_btn = gr.Button("🎧 Tạo giọng nói")
status = gr.Textbox(label="Status")
output_audio = gr.Audio(label="Output", type="filepath")

run_btn.click(
run_tts,
inputs=[
api_keys, voice_id, text, model, fmt,
stability, similarity, style, speed, boost
],
outputs=[status, output_audio]
)

demo.launch()

app.py

@@ -1,161 +0,0 @@
-# -*- coding: utf-8 -*-
-import os, re, time, random, zipfile, requests, natsort
-import gradio as gr
-from pydub import AudioSegment
-
-# ================= API CHECK =================
-def check_api_key(api_key):
-    try:
-        r = requests.get(
-            "https://api.elevenlabs.io/v1/user",
-            headers={"xi-api-key": api_key},
-            timeout=10
-        )
-        if r.status_code != 200:
-            return None
-        sub = r.json().get("subscription", {})
-        return sub.get("character_limit", 0) - sub.get("character_count", 0)
-    except:
-        return None
-
-
-# ================= TEXT =================
-def split_text(text, max_len=200):
-    out, cur = [], ""
-    for s in re.split(r'(?<=[.!?])\s+', text):
-        if len(cur) + len(s) <= max_len:
-            cur += " " + s
-        else:
-            out.append(cur.strip())
-            cur = s
-    if cur:
-        out.append(cur.strip())
-    return out
-
-
-# ================= TTS (FIX QUAN TRỌNG) =================
-def tts(text, api_key, voice_id, model):
-    time.sleep(random.uniform(0.8, 1.5))
-    r = requests.post(
-        f"https://api.elevenlabs.io/v1/text-to-speech/{voice_id}",
-        headers={
-            "xi-api-key": api_key,
-            "Content-Type": "application/json",
-            "Accept": "audio/mpeg"   # 🔥 BẮT BUỘC
-        },
-        json={
-            "text": text,
-            "model_id": model,
-            "voice_settings": {
-                "stability": 0.9,
-                "similarity_boost": 0.5,
-                "style": 0.4,
-                "speed": 0.81,
-                "use_speaker_boost": True
-            }
-        },
-        timeout=60
-    )
-    if r.status_code == 200 and len(r.content) > 1000:
-        return r.content
-    return None
-
-
-# ================= AUDIO + SRT =================
-def merge_audio(folder, fmt):
-    files = natsort.natsorted(f for f in os.listdir(folder) if f.endswith(fmt))
-    audio = AudioSegment.from_file(os.path.join(folder, files[0]))
-    for f in files[1:]:
-        audio += AudioSegment.silent(500)
-        audio += AudioSegment.from_file(os.path.join(folder, f))
-    out = f"output_full.{fmt}"
-    audio.export(out, format=fmt)
-    return out
-
-
-def create_srt(folder, texts):
-    t, lines = 0, []
-    files = natsort.natsorted(f for f in os.listdir(folder) if f.startswith("voice_"))
-    for i, (f, txt) in enumerate(zip(files, texts), 1):
-        a = AudioSegment.from_file(os.path.join(folder, f))
-        lines += [
-            str(i),
-            f"00:00:{t//1000:02},{t%1000:03} --> 00:00:{(t+len(a))//1000:02},{(t+len(a))%1000:03}",
-            txt, ""
-        ]
-        t += len(a) + 500
-    with open(os.path.join(folder, "output_full.srt"), "w", encoding="utf-8") as f:
-        f.write("\n".join(lines))
-
-
-# ================= RUN =================
-def run(api_text, api_file, voice_id, text, model, fmt):
-    keys = api_file.decode().splitlines() if api_file else api_text.splitlines()
-    keys = [k.strip() for k in keys if k.strip()]
-
-    table = "| API KEY | CREDIT |\n|---|---|\n"
-    valid = []
-
-    for k in keys:
-        rem = check_api_key(k)
-        show = f"{k[:6]}...{k[-4:]}"
-        if rem and rem > 600:
-            valid.append(k)
-            table += f"| {show} | {rem} |\n"
-        else:
-            table += f"| {show} | ❌ |\n"
-
-    if not valid:
-        return "❌ Không có API key >600", None, None, table
-
-    texts = split_text(text)
-    os.makedirs("voices", exist_ok=True)
-    for f in os.listdir("voices"):
-        os.remove(os.path.join("voices", f))
-
-    for i, t in enumerate(texts):
-        success = False
-        for key in valid:
-            audio = tts(t, key, voice_id, model)
-            if audio:
-                with open(f"voices/voice_{i+1:03d}.{fmt}", "wb") as f:
-                    f.write(audio)
-                success = True
-                break
-        if not success:
-            return "❌ Không API key nào tạo được audio", None, None, table
-
-    merged = merge_audio("voices", fmt)
-    create_srt("voices", texts)
-
-    zipf = "output.zip"
-    with zipfile.ZipFile(zipf, "w") as z:
-        for f in os.listdir("voices"):
-            z.write(os.path.join("voices", f), f)
-        z.write(merged)
-
-    return "✅ HOÀN TẤT", merged, zipf, table
-
-
-# ================= UI =================
-with gr.Blocks() as app:
-    gr.Markdown("## 🔊 ElevenLabs TTS – FIX CHUẨN HF")
-
-    api_text = gr.Textbox(lines=4, label="API key")
-    api_file = gr.File(type="binary", label="Upload API file")
-    voice_id = gr.Textbox(label="Voice ID")
-    text = gr.Textbox(lines=6, label="Text")
-
-    model = gr.Dropdown(["eleven_multilingual_v2"], value="eleven_multilingual_v2")
-    fmt = gr.Dropdown(["mp3", "wav"], value="mp3")
-
-    btn = gr.Button("🎧 TẠO GIỌNG")
-    status = gr.Textbox()
-    audio = gr.Audio(type="filepath")
-    zipf = gr.File()
-    table = gr.Markdown()
-
-    btn.click(run, [api_text, api_file, voice_id, text, model, fmt],
-              [status, audio, zipf, table])
-
-app.launch()