Spaces:
Sleeping
Sleeping
File size: 3,701 Bytes
efbc3d2 941449c | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 | import os
import base64
import json
import streamlit as st
from sarvamai import SarvamAI
from typing import List, Dict, Optional
client = None
try:
api_key = st.secrets.get("SARVAM_API_KEY") or os.getenv("SARVAM_API_KEY")
if api_key:
client = SarvamAI(api_subscription_key=api_key)
print("✅ Sarvam AI client for TTS (Bulbul) initialized successfully.")
else:
print("⚠️ Warning: SARVAM_API_KEY not found.")
except Exception as e:
print(f"❌ Error initializing Sarvam AI client: {e}")
# --- Language Mapping ---
LANGUAGE_CODE_MAP = {
"hindi": "hi-IN", "bengali": "bn-IN", "tamil": "ta-IN", "telugu": "te-IN",
"gujarati": "gu-IN", "kannada": "kn-IN", "malayalam": "ml-IN", "marathi": "mr-IN",
"punjabi": "pa-IN", "odia": "od-IN", "english": "en-IN",
}
def get_language_code(language_name: str) -> Optional[str]:
return LANGUAGE_CODE_MAP.get(language_name.lower())
def generate_audio_from_text(
text: str,
language_name: str,
gender: str,
output_file_path: str
) -> bool:
"""
Generates an audio file from a text string using the Sarvam "Bulbul" TTS API
with a specified gender and pace for the voice.
"""
if not client: return False
lang_code = get_language_code(language_name)
if not lang_code:
print(f"❌ Language '{language_name}' is not supported. Skipping.")
return False
if gender.lower() == "male":
speaker_name = "abhilash"
pace_value = 1.0
else:
speaker_name = "anushka"
pace_value = 0.9
print(f"--- 🎤 Generating audio for chunk: '{text[:50]}...' in {language_name} (Voice: {speaker_name}, Pace: {pace_value}) ---")
try:
response = client.text_to_speech.convert(
text=text,
model="bulbul:v2",
target_language_code=lang_code,
speaker=speaker_name,
pace=pace_value, # Use the selected pace
speech_sample_rate=22050,
enable_preprocessing=True
)
combined_audio_b64 = "".join(response.audios)
audio_data = base64.b64decode(combined_audio_b64)
with open(output_file_path, "wb") as f:
f.write(audio_data)
print(f"✅ Audio saved to {output_file_path}")
return True
except Exception as e:
print(f"❌ An error occurred during the Sarvam TTS API call: {e}")
return False
def generate_all_audio_from_file(
json_path: str,
target_language: str,
gender: str,
output_dir: str = "generated_audio",
output_json_path: str = "multimedia_data_final.json"
) -> List[Dict[str, str]]:
"""
Reads data from a JSON, generates audio with a specific gender, and saves a final JSON.
"""
try:
with open(json_path, 'r', encoding='utf-8') as f:
multimedia_data = json.load(f)
except (FileNotFoundError, json.JSONDecodeError) as e:
print(f"❌ Error reading or parsing {json_path}: {e}")
return []
os.makedirs(output_dir, exist_ok=True)
for i, item in enumerate(multimedia_data):
audio_text = item.get("audio_text")
if not audio_text:
item["audio_path"] = None
continue
file_path = os.path.join(output_dir, f"audio_{i:03d}.mp3")
success = generate_audio_from_text(audio_text, target_language, gender, file_path)
item["audio_path"] = file_path if success else None
with open(output_json_path, 'w', encoding='utf-8') as f:
json.dump(multimedia_data, f, indent=2, ensure_ascii=False)
print(f"\n--- ✅ Audio generation finished. Final data saved to {output_json_path}. ---")
return multimedia_data |