Ttspro / app.py
Hamed744's picture
Update app.py
d29f80b verified
raw
history blame
34.4 kB
import gradio as gr
# import base64 # Not used in your original core logic
import mimetypes
import os
import re
import struct
import time
# import zipfile # Not used in your original core logic
from google import genai
from google.genai import types as genai_types # Aliased to avoid conflict with built-in 'types'
import logging # Standard Python logging
try:
from pydub import AudioSegment
PYDUB_AVAILABLE = True
except ImportError:
PYDUB_AVAILABLE = False
# logging.warning("Pydub is not available. Audio merging will be disabled.") # Initialized later
# --- Basic Logging Setup ---
# Using a simpler logging setup if the AlphaTranslator_Styled one is too complex for "no other changes"
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
# --- START: Core TTS Logic from YOUR AlphaTTS_Original (UNCHANGED) ---
SPEAKER_VOICES = [
"Achird", "Zubenelgenubi", "Vindemiatrix", "Sadachbia", "Sadaltager",
"Sulafat", "Laomedeia", "Achernar", "Alnilam", "Schedar", "Gacrux",
"Pulcherrima", "Umbriel", "Algieba", "Despina", "Erinome", "Algenib",
"Rasalthgeti", "Orus", "Aoede", "Callirrhoe", "Autonoe", "Enceladus",
"Iapetus", "Zephyr", "Puck", "Charon", "Kore", "Fenrir", "Leda"
]
FIXED_MODEL_NAME = "gemini-2.5-flash-preview-tts" # FROM YOUR ORIGINAL CODE
DEFAULT_MAX_CHUNK_SIZE = 3800
DEFAULT_SLEEP_BETWEEN_REQUESTS = 8
DEFAULT_OUTPUT_FILENAME_BASE = "alpha_tts_audio"
def _log(message, log_list): # YOUR _log function
log_list.append(message)
logging.info(f"[AlphaTTS_LOG] {message}") # Standard logging also
def save_binary_file(file_name, data, log_list):
try:
with open(file_name, "wb") as f: f.write(data)
_log(f"✅ فایل ذخیره شد: {file_name}", log_list)
return file_name
except Exception as e:
_log(f"❌ خطا در ذخیره فایل {file_name}: {e}", log_list)
return None
def convert_to_wav(audio_data: bytes, mime_type: str) -> bytes:
parameters = parse_audio_mime_type(mime_type)
bits_per_sample, rate = parameters["bits_per_sample"], parameters["rate"]
num_channels, data_size = 1, len(audio_data)
bytes_per_sample, block_align = bits_per_sample // 8, num_channels * (bits_per_sample // 8)
byte_rate, chunk_size = rate * block_align, 36 + data_size
header = struct.pack("<4sI4s4sIHHIIHH4sI", b"RIFF", chunk_size, b"WAVE", b"fmt ", 16, 1, num_channels, rate, byte_rate, block_align, bits_per_sample, b"data", data_size)
return header + audio_data
def parse_audio_mime_type(mime_type: str) -> dict[str, int]:
bits, rate = 16, 24000
for param in mime_type.split(";"):
param = param.strip()
if param.lower().startswith("rate="):
try: rate = int(param.split("=", 1)[1])
except ValueError: pass
elif param.startswith("audio/L"):
try: bits = int(param.split("L", 1)[1])
except ValueError: pass
return {"bits_per_sample": bits, "rate": rate}
def smart_text_split(text, max_size=3800, log_list=None):
if len(text) <= max_size: return [text]
chunks, current_chunk = [], ""
sentences = re.split(r'(?<=[.!?؟۔])\s+', text) # Added Persian full stop for robustness
for sentence in sentences:
if len(current_chunk) + len(sentence) + 1 > max_size:
if current_chunk: chunks.append(current_chunk.strip())
current_chunk = sentence
while len(current_chunk) > max_size:
split_idx = -1
for punc in ['،', ',', ';', ':', ' ']:
try:
idx = current_chunk.rindex(punc, max_size // 2, max_size)
if idx > split_idx:
split_idx = idx
except ValueError:
pass
if split_idx != -1 :
part, current_chunk = current_chunk[:split_idx+1], current_chunk[split_idx+1:]
else:
part, current_chunk = current_chunk[:max_size], current_chunk[max_size:]
chunks.append(part.strip())
else: current_chunk += (" " if current_chunk and sentence else "") + sentence
if current_chunk: chunks.append(current_chunk.strip())
final_chunks = [c for c in chunks if c]
if log_list: _log(f"📊 متن به {len(final_chunks)} قطعه تقسیم شد.", log_list)
return final_chunks
def merge_audio_files_func(file_paths, output_path, log_list):
if not PYDUB_AVAILABLE:
_log("❌ pydub در دسترس نیست. ادغام انجام نشد.", log_list)
return False
try:
_log(f"🔗 ادغام {len(file_paths)} فایل صوتی...", log_list)
combined = AudioSegment.empty()
for i, fp in enumerate(file_paths):
if os.path.exists(fp):
try:
segment = AudioSegment.from_file(fp)
combined += segment
if i < len(file_paths) - 1:
combined += AudioSegment.silent(duration=150)
except Exception as e_pydub:
_log(f"⚠️ خطای Pydub در پردازش فایل '{fp}': {e_pydub}. از این فایل صرف نظر می شود.", log_list)
continue
else:
_log(f"⚠️ فایل پیدا نشد: {fp}", log_list)
if len(combined) == 0:
_log("❌ هیچ قطعه صوتی برای ادغام وجود ندارد.", log_list)
return False
combined.export(output_path, format="wav")
_log(f"✅ فایل ادغام شده: {output_path}", log_list)
return True
except Exception as e:
_log(f"❌ خطا در ادغام: {e}", log_list) # traceback.format_exc() removed to keep it closer to original
return False
def core_generate_audio(text_input, prompt_input, selected_voice, temperature_val, log_list): # YOUR core_generate_audio
output_base_name = DEFAULT_OUTPUT_FILENAME_BASE
max_chunk, sleep_time = DEFAULT_MAX_CHUNK_SIZE, DEFAULT_SLEEP_BETWEEN_REQUESTS
_log(f"🚀 شروع فرآیند با مدل: {FIXED_MODEL_NAME}...", log_list)
api_key = os.environ.get("GEMINI_API_KEY") # YOUR WAY OF GETTING API KEY
if not api_key:
_log("❌ کلید API با نام GEMINI_API_KEY در متغیرهای محیطی تنظیم نشده.", log_list)
return None # Return None only, as per your original AlphaTTS
try:
client = genai.Client(api_key=api_key) # YOUR WAY OF CLIENT INSTANTIATION
_log(f"کلاینت Gemini با کلید API برای مدل {FIXED_MODEL_NAME} مقداردهی اولیه شد.", log_list)
except Exception as e:
_log(f"❌ خطا در مقداردهی اولیه کلاینت Gemini: {e}", log_list)
return None
if not text_input or not text_input.strip():
_log("❌ متن ورودی خالی.", log_list)
return None
text_chunks = smart_text_split(text_input, max_chunk, log_list)
if not text_chunks:
_log("❌ متن قابل پردازش نیست.", log_list)
return None
generated_files = []
for i, chunk in enumerate(text_chunks):
_log(f"🔊 پردازش قطعه {i+1}/{len(text_chunks)} (صدا: {selected_voice}, دما: {temperature_val})...", log_list)
final_text = f'"{prompt_input}"\n{chunk}' if prompt_input and prompt_input.strip() else chunk
# Using genai_types (aliased) for Content, Part etc. as in your original imports
contents = [genai_types.Content(role="user", parts=[genai_types.Part.from_text(text=final_text)])]
config = genai_types.GenerateContentConfig( # YOUR CONFIG OBJECT
temperature=temperature_val,
response_modalities=["audio"],
speech_config=genai_types.SpeechConfig(
voice_config=genai_types.VoiceConfig(
prebuilt_voice_config=genai_types.PrebuiltVoiceConfig(voice_name=selected_voice)
)
)
)
_log(f"کانفیگ API برای قطعه {i+1}: دما={temperature_val}, صدا={selected_voice}, مدالیته=['audio']", log_list)
fname_base = f"{output_base_name}_part{i+1:03d}"
try:
# YOUR API CALL
response = client.models.generate_content(model=FIXED_MODEL_NAME, contents=contents, config=config)
if response.candidates and response.candidates[0].content and response.candidates[0].content.parts and response.candidates[0].content.parts[0].inline_data:
inline_data = response.candidates[0].content.parts[0].inline_data
data_buffer = inline_data.data
mime_type = inline_data.mime_type
_log(f"داده صوتی در candidate.part[0].inline_data برای قطعه {i+1} یافت شد.", log_list)
ext = mimetypes.guess_extension(mime_type) or ".wav"
if "audio/L" in mime_type and ext == ".wav":
_log(f"تبدیل صدای خام PCM (MIME: {mime_type}) به WAV برای قطعه {i+1}.", log_list)
data_buffer = convert_to_wav(data_buffer, mime_type)
if not ext.startswith("."): ext = "." + ext
fpath = save_binary_file(f"{fname_base}{ext}", data_buffer, log_list)
if fpath: generated_files.append(fpath)
else:
_log(f"⚠️ پاسخ API برای قطعه {i+1} بدون داده صوتی در مسیر مورد انتظار.", log_list)
_log(f"ساختار کامل پاسخ (اولین 500 کاراکتر): {str(response)[:500]}", log_list)
# continue # As per your original code, it continues
except Exception as e: # Catching generic Exception as in your original
_log(f"❌ خطا در تولید قطعه {i+1}: {e}", log_list)
# traceback.format_exc() was not in your original core_generate_audio, so removed here
continue
if i < len(text_chunks) - 1 and len(text_chunks) > 1:
_log(f"💤 توقف کوتاه ({sleep_time} ثانیه) قبل از قطعه بعدی...", log_list)
time.sleep(sleep_time)
if not generated_files:
_log("❌ هیچ فایلی تولید نشد.", log_list)
return None # Return None only as per your original AlphaTTS
_log(f"🎉 {len(generated_files)} فایل(های) صوتی تولید شد.", log_list)
final_audio_file = None
final_output_path_base = f"{output_base_name}_final"
if len(generated_files) > 1:
if PYDUB_AVAILABLE:
merged_fn = f"{final_output_path_base}.wav"
if os.path.exists(merged_fn):
try: os.remove(merged_fn)
except OSError: _log(f"⚠️ عدم امکان حذف فایل ادغام شده قبلی '{merged_fn}' (خطای سیستم عامل)", log_list)
except Exception as e_rm: _log(f"⚠️ عدم امکان حذف فایل ادغام شده قبلی '{merged_fn}': {e_rm}", log_list)
if merge_audio_files_func(generated_files, merged_fn, log_list):
final_audio_file = merged_fn
for fp_path in generated_files:
if os.path.abspath(fp_path) != os.path.abspath(merged_fn):
try: os.remove(fp_path)
except OSError: _log(f"⚠️ عدم امکان حذف فایل موقت '{fp_path}' (خطای سیستم عامل)", log_list)
except Exception as e_del: _log(f"⚠️ عدم امکان حذف فایل موقت '{fp_path}': {e_del}", log_list)
else:
_log("⚠️ ادغام فایل‌های صوتی ناموفق بود. اولین قطعه ارائه می‌شود.", log_list)
if generated_files:
try:
first_chunk_path = generated_files[0]
target_ext = os.path.splitext(first_chunk_path)[1]
fallback_fn = f"{final_output_path_base}{target_ext}" # Simplified name for fallback
if os.path.exists(fallback_fn) and os.path.abspath(first_chunk_path) != os.path.abspath(fallback_fn):
os.remove(fallback_fn)
if os.path.abspath(first_chunk_path) != os.path.abspath(fallback_fn):
os.rename(first_chunk_path, fallback_fn)
final_audio_file = fallback_fn
for i_gf in range(1, len(generated_files)):
try: os.remove(generated_files[i_gf])
except: pass # Keep silent as per your original
except Exception as e_rename_fb:
_log(f"خطا در تغییر نام فایل اولین قطعه: {e_rename_fb}", log_list) # Was `e_rename` in your original
final_audio_file = generated_files[0]
else:
_log("⚠️ pydub نیست. اولین قطعه ارائه می‌شود.", log_list)
if generated_files:
try:
first_chunk_path = generated_files[0]
target_ext = os.path.splitext(first_chunk_path)[1]
single_fallback_fn = f"{final_output_path_base}{target_ext}" # Simplified name
if os.path.exists(single_fallback_fn) and os.path.abspath(first_chunk_path) != os.path.abspath(single_fallback_fn):
os.remove(single_fallback_fn)
if os.path.abspath(first_chunk_path) != os.path.abspath(single_fallback_fn):
os.rename(first_chunk_path, single_fallback_fn)
final_audio_file = single_fallback_fn
for i_gf in range(1, len(generated_files)):
try: os.remove(generated_files[i_gf])
except: pass # Keep silent
except Exception as e_rename_single_npd: # Was `e_rename_single` in your original
_log(f"خطا در تغییر نام فایل اولین قطعه (بدون pydub): {e_rename_single_npd}", log_list)
final_audio_file = generated_files[0]
elif len(generated_files) == 1:
try:
single_file_path = generated_files[0]
target_ext = os.path.splitext(single_file_path)[1]
final_single_fn = f"{final_output_path_base}{target_ext}"
if os.path.exists(final_single_fn) and os.path.abspath(single_file_path) != os.path.abspath(final_single_fn):
os.remove(final_single_fn)
if os.path.abspath(single_file_path) != os.path.abspath(final_single_fn):
os.rename(single_file_path, final_single_fn)
final_audio_file = final_single_fn
except Exception as e_rename_sgl_final: # Was `e_rename_single_final` in your original
_log(f"خطا در تغییر نام فایل تکی نهایی: {e_rename_sgl_final}", log_list)
final_audio_file = generated_files[0]
if final_audio_file and not os.path.exists(final_audio_file):
_log(f"⚠️ فایل نهایی '{final_audio_file}' وجود ندارد!", log_list)
return None
return final_audio_file # Returns only path, as per your original AlphaTTS
# Your original gradio_tts_interface
def gradio_tts_interface(use_file_input, uploaded_file, text_to_speak, speech_prompt, speaker_voice, temperature, progress=gr.Progress(track_tqdm=True)):
logs = []
actual_text = ""
if use_file_input:
if uploaded_file:
try:
# In Gradio, uploaded_file.name is the path to the temporary file
with open(uploaded_file.name, 'r', encoding='utf-8') as f: actual_text = f.read().strip()
if not actual_text: return None # Return None only, as per your original
except Exception as e: _log(f"❌ خطا خواندن فایل: {e}", logs); return None
else: return None
else:
actual_text = text_to_speak
if not actual_text or not actual_text.strip(): return None
final_path = core_generate_audio(actual_text, speech_prompt, speaker_voice, temperature, logs)
# Your original code commented out printing logs here, so I'll keep it commented.
# for log_entry in logs: print(log_entry) # For debugging in HF console
return final_path # Returns only path, as per your original AlphaTTS
# --- END: Core TTS Logic from YOUR AlphaTTS_Original (UNCHANGED) ---
# --- START: Gradio UI with AlphaTranslator_Styled Appearance ---
# (Using CSS variables from AlphaTranslator_Styled for colors and fonts)
FLY_PRIMARY_COLOR_HEX = "#4F46E5"
FLY_SECONDARY_COLOR_HEX = "#10B981"
FLY_ACCENT_COLOR_HEX = "#D97706"
FLY_TEXT_COLOR_HEX = "#1F2937"
FLY_SUBTLE_TEXT_HEX = "#6B7280"
FLY_LIGHT_BACKGROUND_HEX = "#F9FAFB"
FLY_WHITE_HEX = "#FFFFFF"
FLY_BORDER_COLOR_HEX = "#D1D5DB"
FLY_INPUT_BG_HEX_SIMPLE = "#F3F4F6"
FLY_PANEL_BG_SIMPLE = "#E0F2FE"
app_theme_outer_styled = gr.themes.Base( # New theme object name to avoid conflict if you had `app_theme_outer`
font=[gr.themes.GoogleFont("Inter"), "system-ui", "sans-serif"],
).set(
body_background_fill=FLY_LIGHT_BACKGROUND_HEX,
)
# CSS from AlphaTranslator_Styled, adapted slightly for your component names/IDs if needed
# Your original component IDs are like "use_file_cb_alpha_v3", "file_uploader_alpha_main_v3", etc.
# The CSS below uses general selectors but can be made more specific if those IDs are kept.
applied_css_for_alphatts = f"""
@import url('https://fonts.googleapis.com/css2?family=Vazirmatn:wght@300;400;500;600;700;800&display=swap');
/* Poppins and Inter are from AlphaTranslator_Styled, Vazirmatn from your AlphaTTS_Original theme */
@import url('https://fonts.googleapis.com/css2?family=Poppins:wght@400;500;600;700;800&display=swap');
@import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700;800&display=swap');
:root {{
--fly-primary: {FLY_PRIMARY_COLOR_HEX}; --fly-secondary: {FLY_SECONDARY_COLOR_HEX};
--fly-accent: {FLY_ACCENT_COLOR_HEX}; --fly-text-primary: {FLY_TEXT_COLOR_HEX};
--fly-text-secondary: {FLY_SUBTLE_TEXT_HEX}; --fly-bg-light: {FLY_LIGHT_BACKGROUND_HEX};
--fly-bg-white: {FLY_WHITE_HEX}; --fly-border-color: {FLY_BORDER_COLOR_HEX};
--fly-input-bg-simple: {FLY_INPUT_BG_HEX_SIMPLE}; --fly-panel-bg-simple: {FLY_PANEL_BG_SIMPLE};
--font-global: 'Vazirmatn', 'Inter', 'Poppins', system-ui, sans-serif; /* Vazirmatn prioritized */
--font-english: 'Poppins', 'Inter', system-ui, sans-serif;
--radius-sm: 0.375rem; --radius-md: 0.5rem; --radius-lg: 0.75rem; --radius-xl: 1rem; --radius-full: 9999px;
--shadow-xl: 0 20px 25px -5px rgba(0,0,0,0.1),0 8px 10px -6px rgba(0,0,0,0.1);
--fly-primary-rgb: 79,70,229; --fly-accent-rgb: 217,119,6;
/* Variables from your original AlphaTTS CSS if they were different and needed */
/* For example, if your original had --app-button-bg for the blue button */
--app-button-bg-original: #2979FF; /* Blue from your original AlphaTTS button */
--shadow-button-original: 0 4px 10px -2px rgba(41,121,255,0.5);
--radius-input-original: 8px;
--app-border-color-original: #E0E0E0;
--app-input-bg-original: #F7F7F7;
}}
body {{
font-family:var(--font-global); direction:rtl; background-color:var(--fly-bg-light);
color:var(--fly-text-primary); line-height:1.7; font-size:16px;
}}
.gradio-container {{ /* Overall container styling from AlphaTranslator_Styled */
max-width:100% !important; width:100% !important; min-height:100vh;
margin:0 auto !important; padding:0 !important; border-radius:0 !important;
box-shadow:none !important; background:linear-gradient(170deg, #E0F2FE 0%, #F3E8FF 100%);
display:flex; flex-direction:column;
}}
/* Header styling from AlphaTranslator_Styled */
.app-header-alphatts {{ /* Changed class name slightly to avoid conflict if both apps run */
text-align:center; padding:2.5rem 1rem; margin:0;
background:linear-gradient(135deg,var(--fly-primary) 0%,var(--fly-secondary) 100%);
color:var(--fly-bg-white); border-bottom-left-radius:var(--radius-xl);
border-bottom-right-radius:var(--radius-xl); box-shadow:var(--shadow-lg);
position:relative; overflow:hidden;
}}
.app-header-alphatts::before {{ /* Decorative element from AlphaTranslator_Styled */
content:''; position:absolute; top:-50px; right:-50px; width:150px; height:150px;
background:rgba(255,255,255,0.1); border-radius:var(--radius-full);
opacity:0.5; transform:rotate(45deg);
}}
.app-header-alphatts h1 {{ /* h1 from AlphaTranslator_Styled */
font-size:2.25em !important; font-weight:800 !important; margin:0 0 0.5rem 0;
font-family:var(--font-english); letter-spacing:-0.5px; text-shadow:0 2px 4px rgba(0,0,0,0.1);
}}
.app-header-alphatts p {{ /* p from AlphaTranslator_Styled */
font-size:1em !important; margin-top:0.25rem; font-weight:400;
color:rgba(255,255,255,0.85) !important;
}}
/* Main content panel styling from AlphaTranslator_Styled */
.main-content-area-alphatts {{ /* Changed class name slightly */
flex-grow:1; padding:0.75rem; width:100%; margin:0 auto; box-sizing:border-box;
}}
.content-panel-alphatts {{ /* Changed class name slightly */
background-color:var(--fly-bg-white); padding:1rem; border-radius:var(--radius-xl);
box-shadow:var(--shadow-xl); margin-top:-2rem; position:relative; z-index:10;
margin-bottom:2rem; width:100%; box-sizing:border-box;
}}
/* Styling for YOUR UI elements, applying AlphaTranslator_Styled aesthetics */
/* Inputs (Textbox, Dropdown, File) */
.content-panel-alphatts .gr-input > label + div > textarea,
.content-panel-alphatts .gr-dropdown > label + div > div > input,
.content-panel-alphatts .gr-dropdown > label + div > div > select,
.content-panel-alphatts .gr-textbox > label + div > textarea,
.content-panel-alphatts .gr-file > label + div /* For file input styling */
{{
border-radius:var(--radius-input-original) !important; /* Your original radius */
border:1.5px solid var(--fly-border-color) !important; /* Border from AlphaTranslator */
font-size:0.95em !important; background-color:var(--fly-input-bg-simple) !important; /* BG from AlphaTranslator */
padding:10px 12px !important; color:var(--fly-text-primary) !important;
}}
.content-panel-alphatts .gr-input > label + div > textarea:focus,
.content-panel-alphatts .gr-dropdown > label + div > div > input:focus,
.content-panel-alphatts .gr-dropdown > label + div > div > select:focus,
.content-panel-alphatts .gr-textbox > label + div > textarea:focus,
.content-panel-alphatts .gr-file > label + div:focus-within
{{
border-color:var(--fly-primary) !important; /* Focus border from AlphaTranslator */
box-shadow:0 0 0 3px rgba(var(--fly-primary-rgb),0.12) !important;
background-color:var(--fly-bg-white) !important;
}}
.content-panel-alphatts .gr-file > label + div {{ text-align:center; border-style: dashed !important; }}
/* Button: Using --fly-accent for consistency with AlphaTranslator's primary action color */
.content-panel-alphatts .gr-button[elem_id="generate_button_alpha_v3"], /* Your button ID */
.content-panel-alphatts button[variant="primary"] /* General primary button */
{{
background:var(--fly-accent) !important; /* Orange accent from AlphaTranslator */
margin-top:1.5rem !important; padding:12px 20px !important; /* Adjusted padding */
transition:all 0.25s ease-in-out !important; color:white !important; font-weight:600 !important;
border-radius:var(--radius-input-original) !important; /* Your original radius */ border:none !important;
box-shadow:0 3px 8px -1px rgba(var(--fly-accent-rgb),0.3) !important;
width:100% !important; font-size:1.05em !important; /* Your original font size */
display:flex; align-items:center; justify-content:center;
}}
.content-panel-alphatts .gr-button[elem_id="generate_button_alpha_v3"]:hover,
.content-panel-alphatts button[variant="primary"]:hover
{{
background:#B45309 !important; /* Darker orange */ transform:translateY(-1px) !important;
box-shadow:0 5px 10px -1px rgba(var(--fly-accent-rgb),0.4) !important;
}}
/* Labels (using AlphaTranslator_Styled general label style) */
.content-panel-alphatts label > span.label-text
{{
font-weight:500 !important; color:#4B5563 !important;
font-size:0.88em !important; margin-bottom:6px !important; display:inline-block;
}}
/* Your original specific label styling with icons (if you want to keep them) */
/* You would need to ensure your Gradio labels have the correct `for` attribute linking to input `elem_id`
or use JavaScript to add these pseudo-elements if Gradio doesn't directly support `for` on labels.
For simplicity, I'm omitting the ::before icon styles unless you confirm they are essential
and your Gradio setup can support them easily. The general label style above will apply.
*/
/* Temperature description (from your original AlphaTTS CSS) */
.content-panel-alphatts .temp_description_class_alpha_v3 {{
font-size: 0.85em; color: #777; margin-top: -0.4rem; margin-bottom: 1rem;
}}
/* Audio Player (general styling, can be targeted by ID if set) */
.content-panel-alphatts .gr-audio audio, /* General audio player */
.content-panel-alphatts #output_audio_player_alpha_v3 audio /* Your specific ID */
{{
width: 100%; border-radius: var(--radius-input-original); margin-top:0.8rem;
}}
/* Examples (using AlphaTranslator_Styled examples button style) */
.content-panel-alphatts div[label*="نمونه‌های کاربردی"] .gr-button.gr-button-tool, /* Targetting by label */
.content-panel-alphatts div[label*="نمونه‌های کاربردی"] .gr-sample-button
{{
background-color:#E0E7FF !important; color:var(--fly-primary) !important;
border-radius:6px !important; font-size:0.78em !important; padding:4px 8px !important;
}}
.content-panel-alphatts .custom-hr {{height:1px;background-color:var(--fly-border-color);margin:1.5rem 0;border:none;}}
/* Footer styling from AlphaTranslator_Styled */
.app-footer-alphatts {{ /* Changed class name slightly */
text-align:center;font-size:0.85em;color:var(--fly-text-secondary);margin-top:2.5rem;
padding:1rem 0;background-color:rgba(255,255,255,0.3);backdrop-filter:blur(5px);
border-top:1px solid var(--fly-border-color);
}}
footer {{display:none !important;}} /* Hides default Gradio footer */
/* Responsive adjustments from AlphaTranslator_Styled */
@media (min-width:640px) {{
.main-content-area-alphatts {{padding:1.5rem;max-width:700px;}}
.content-panel-alphatts {{padding:1.5rem;}}
.app-header-alphatts h1 {{font-size:2.5em !important;}}
.app-header-alphatts p {{font-size:1.05em !important;}}
}}
@media (min-width:768px) {{
.main-content-area-alphatts {{max-width:780px;}}
.content-panel-alphatts {{padding:2rem;}}
.content-panel-alphatts .gr-button[elem_id="generate_button_alpha_v3"],
.content-panel-alphatts button[variant="primary"]
{{
width:auto !important; align-self:flex-start;
}}
.app-header-alphatts h1 {{font-size:2.75em !important;}}
.app-header-alphatts p {{font-size:1.1em !important;}}
}}
"""
# Using your original Gradio Blocks structure
# The theme `gr.themes.Base(font=[gr.themes.GoogleFont("Vazirmatn")])` is from your original.
# We are applying `app_theme_outer_styled` for the body background and `applied_css_for_alphatts` for specifics.
with gr.Blocks(theme=app_theme_outer_styled, css=applied_css_for_alphatts, title=f"آلفا TTS ({FIXED_MODEL_NAME.split('-')[1]})") as demo:
# Applying the header from AlphaTranslator_Styled structure
gr.HTML(f"""
<div class='app-header-alphatts'>
<h1>🚀 Alpha TTS</h1>
<p>جادوی تبدیل متن به صدا در دستان شما (Gemini {FIXED_MODEL_NAME.split('-')[1]})</p>
</div>
""")
# Applying the main content panel structure from AlphaTranslator_Styled
with gr.Column(elem_classes=["main-content-area-alphatts"]):
with gr.Column(elem_classes=["content-panel-alphatts"]): # Your original AlphaTTS used Column, let's keep it simple
# AlphaTranslator used Group, but Column is fine.
# Your original UI layout from AlphaTTS_Original
# Note: `elem_id`s are from your original AlphaTTS code.
# CSS selectors have been updated to try and match these or use general selectors.
# Warning if GEMINI_API_KEY is not set
if not os.environ.get("GEMINI_API_KEY"):
missing_key_msg = (
"⚠️ هشدار: متغیر محیطی GEMINI_API_KEY تنظیم نشده است. "
"قابلیت تبدیل متن به گفتار احتمالاً کار نخواهد کرد. "
"لطفاً این متغیر را در بخش Secrets این Space تنظیم کنید."
)
gr.Markdown(f"<div class='api-warning-message'>{missing_key_msg}</div>")
use_file_input_cb = gr.Checkbox(label="📄 استفاده از فایل متنی (.txt)", value=False, elem_id="use_file_cb_alpha_v3")
uploaded_file_input = gr.File(
label=" ",
file_types=['.txt'],
visible=False,
elem_id="file_uploader_alpha_main_v3"
)
text_to_speak_tb = gr.Textbox(
label="متن فارسی برای تبدیل",
placeholder="مثال: سلام، فردا هوا چطور است؟",
lines=5,
value="",
visible=True,
elem_id="text_input_main_alpha_v3"
)
# Your original change function for checkbox
use_file_input_cb.change(
fn=lambda x: (gr.update(visible=x, label=" " if x else "متن فارسی برای تبدیل"), gr.update(visible=not x)),
inputs=use_file_input_cb,
outputs=[uploaded_file_input, text_to_speak_tb]
)
speech_prompt_tb = gr.Textbox(
label="سبک گفتار (اختیاری)",
placeholder="مثال: با لحنی شاد و پرانرژی",
value="با لحنی دوستانه و رسا صحبت کن.",
lines=2, elem_id="speech_prompt_alpha_v3"
)
speaker_voice_dd = gr.Dropdown(
SPEAKER_VOICES, label="انتخاب گوینده و لهجه", value="Charon", elem_id="speaker_voice_alpha_v3"
)
temperature_slider = gr.Slider(
minimum=0.1, maximum=1.5, step=0.05, value=0.9, label="میزان خلاقیت صدا",
elem_id="temperature_slider_alpha_v3"
)
gr.Markdown("<p class='temp_description_class_alpha_v3'>مقادیر بالاتر = تنوع بیشتر، مقادیر پایین‌تر = یکنواختی بیشتر.</p>")
generate_button = gr.Button("🚀 تولید و پخش صدا", elem_id="generate_button_alpha_v3") # Removed variant="primary" to let CSS handle it via elem_id
output_audio = gr.Audio(label=" ", type="filepath", elem_id="output_audio_player_alpha_v3")
# Your original Examples section
# Applying a custom HR from AlphaTranslator_Styled
gr.HTML("<hr class='custom-hr'>")
gr.Markdown(
"<h3 style='text-align:center; font-weight:500; color:var(--fly-text-secondary); margin-top:1.5rem; margin-bottom:1rem;'>نمونه‌های کاربردی</h3>",
# elem_id="examples_section_title_v3" # elem_id from your original
)
gr.Examples(
examples=[
[False, None, "سلام بر شما، امیدوارم روز خوبی داشته باشید.", "با لحنی گرم و صمیمی.", "Zephyr", 0.85],
[False, None, "این یک آزمایش برای بررسی کیفیت صدای تولید شده توسط هوش مصنوعی آلفا است.", "با صدایی طبیعی و روان.", "Charon", 0.9],
],
inputs=[use_file_input_cb, uploaded_file_input, text_to_speak_tb, speech_prompt_tb, speaker_voice_dd, temperature_slider],
outputs=[output_audio], # Original AlphaTTS only output to audio
fn=gradio_tts_interface,
cache_examples=os.getenv("GRADIO_CACHE_EXAMPLES", "False").lower() == "true" # From AlphaTranslator
)
# Footer from AlphaTranslator_Styled
gr.Markdown(f"<p class='app-footer-alphatts'>Alpha TTS © 2024 - Model: {FIXED_MODEL_NAME}</p>")
# --- Event Handlers (from YOUR AlphaTTS_Original) ---
if generate_button is not None:
generate_button.click(
fn=gradio_tts_interface,
inputs=[use_file_input_cb, uploaded_file_input, text_to_speak_tb, speech_prompt_tb, speaker_voice_dd, temperature_slider],
outputs=[output_audio] # Original AlphaTTS only output to audio
)
else:
logging.error("دکمه تولید صدا (generate_button_alpha_v3) به درستی مقداردهی اولیه نشده است.")
# --- END: Gradio UI ---
if __name__ == "__main__":
# Removed auto-restart thread to keep it closer to your original AlphaTTS.
# If you need it, you can re-add the auto_restart_service function and thread start.
# Check if PYDUB is available at launch
if not PYDUB_AVAILABLE:
logging.warning("Pydub (for audio merging) not found. Please install with `pip install pydub`. Merging will be disabled if multiple audio chunks are generated.")
demo.launch(
server_name="0.0.0.0",
server_port=int(os.getenv("PORT", 7860)),
debug=os.environ.get("GRADIO_DEBUG", "False").lower() == "true",
show_error=True
)