File size: 3,601 Bytes
39da493 | 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 | #!/usr/bin/env python3
"""
Extract audio from Nigerian Common Voice parquet files and save as WAV.
"""
import os
from pathlib import Path
import pandas as pd
import json
BASE_DIR = Path.home() / "voice-training"
DATASETS_DIR = BASE_DIR / "datasets" / "nigerian_cv"
OUTPUT_DIR = BASE_DIR / "prepared_data"
LANGUAGES = ["yoruba", "hausa", "igbo"] # Skip English for Nigerian TTS
def extract_language(lang: str, max_samples: int = 500):
"""Extract audio files for a language."""
lang_dir = DATASETS_DIR / lang
output_dir = OUTPUT_DIR / lang / "wavs"
output_dir.mkdir(parents=True, exist_ok=True)
manifest = []
total_extracted = 0
print(f"\n=== Extracting {lang.upper()} ===")
for split in ["train", "validation"]:
parquet_file = lang_dir / f"{split}-00000-of-00001.parquet"
if not parquet_file.exists():
continue
print(f" Processing {split}...")
df = pd.read_parquet(parquet_file)
for idx, row in df.iterrows():
if total_extracted >= max_samples:
break
audio_data = row.get('audio')
sentence = row.get('sentence', '')
if audio_data is None or not sentence:
continue
# Audio data structure: {'array': [...], 'sampling_rate': 16000}
if isinstance(audio_data, dict):
array = audio_data.get('array')
sr = audio_data.get('sampling_rate', 16000)
if array is not None:
import numpy as np
import soundfile as sf
audio_array = np.array(array).astype(np.float32)
# Save audio
filename = f"{lang}_{split}_{idx:05d}.wav"
filepath = output_dir / filename
sf.write(str(filepath), audio_array, sr)
# Add to manifest
manifest.append({
"audio_file": str(filepath),
"text": sentence.strip(),
"language": lang,
"speaker": row.get('client_id', 'unknown')[:8]
})
total_extracted += 1
if total_extracted % 100 == 0:
print(f" Extracted {total_extracted} samples...")
if total_extracted >= max_samples:
break
# Save manifest
manifest_file = OUTPUT_DIR / lang / "manifest.json"
with open(manifest_file, 'w', encoding='utf-8') as f:
json.dump(manifest, f, indent=2, ensure_ascii=False)
print(f" Total extracted: {total_extracted} samples")
print(f" Manifest saved to: {manifest_file}")
return total_extracted
def main():
print("=== Extracting Nigerian Audio for TTS Training ===")
# Install soundfile if needed
try:
import soundfile
except ImportError:
os.system("pip install soundfile")
import soundfile
total = 0
for lang in LANGUAGES:
if (DATASETS_DIR / lang).exists():
count = extract_language(lang, max_samples=500) # 500 samples per lang for quick training
total += count
print(f"\n=== Extraction Complete ===")
print(f"Total samples extracted: {total}")
print(f"Output directory: {OUTPUT_DIR}")
if __name__ == "__main__":
main()
|