Spaces:
Sleeping
Sleeping
File size: 3,931 Bytes
e785be9 | 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 | print("INIT: Script starting...")
import os
import shutil
print("INIT: Importing datasets...")
import datasets
print("INIT: Imports done.")
from pathlib import Path
# Setup paths
PROJECT_ROOT = Path(__file__).parent.parent
OUTPUT_DIR = PROJECT_ROOT / "Convolve" / "photo" / "voxceleb_data"
OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
def extract_indian_celebs():
print("🚀 Starting extraction of Indian Celebrities from VoxCeleb1...")
# Load the custom builder script
# We use "audio1" config because that's where nationality is available in metadata
ds = datasets.load_dataset(
"scripts/custom_voxceleb.py",
"audio1",
trust_remote_code=True,
split="train" # Use 'train' (dev) partition
)
# LOCAL IMAGE SOURCE
LOCAL_IMG_ROOT = PROJECT_ROOT / "HQ-VoxCeleb" / "HQ-VoxCeleb" / "vox1" / "origin_faces"
# Get list of local celebrities
if not LOCAL_IMG_ROOT.exists():
print(f"❌ Error: Local image directory not found at {LOCAL_IMG_ROOT}")
return
local_celebs = {p.name for p in LOCAL_IMG_ROOT.iterdir() if p.is_dir()}
# 🧪 TEST MODE: Take only 1 person (preferably starting with 'A' to match partaa)
sorted_celebs = sorted(list(local_celebs))
if sorted_celebs:
target = sorted_celebs[0] # e.g. "A.J._Buckley"
print(f"🧪 TEST MODE: Restricting to single celebrity: {target}")
local_celebs = {target}
print(f"📂 Found {len(local_celebs)} local celebrity folders (Filtered to 1).")
# Filter for celebrities that exist locally
# We match 'speaker_name' (e.g. 'A.J._Buckley') with the folder name
filtered_celebs = ds.filter(lambda x: x.get("speaker_name", "") in local_celebs)
print(f"📊 Found {len(filtered_celebs)} clips matching local folders.")
# Group by speaker
speakers = {}
for item in filtered_celebs:
s_id = item['speaker_id']
s_name = item['speaker_name']
if s_id not in speakers:
speakers[s_id] = {
"name": s_name,
"gender": item['speaker_gender'],
"clips": []
}
# Keep track of clips (limit to 5 per person)
if len(speakers[s_id]["clips"]) < 5:
speakers[s_id]["clips"].append(item['audio']['path'])
print(f"👥 Found {len(speakers)} matched speakers.")
# Save data
for s_id, data in speakers.items():
person_dir = OUTPUT_DIR / f"{data['name'].replace(' ', '_')}_{s_id}"
person_dir.mkdir(exist_ok=True)
# Save Metadata
with open(person_dir / "metadata.json", "w") as f:
f.write(str(data))
# 1. Copy Audio Files (from Dataset Script)
for i, clip_path in enumerate(data["clips"]):
src = Path(clip_path)
dest = person_dir / f"voice_sample_{i}.wav"
try:
shutil.copy2(src, dest)
# print(f" Saved audio {dest.name}")
except Exception as e:
print(f" Error copying audio {src}: {e}")
# 2. Copy Local Images (from HQ-VoxCeleb)
# Structure is likely: origin_faces/id10001/1.jpg ...
local_face_dir = LOCAL_IMG_ROOT / s_id
if local_face_dir.exists():
print(f" 📸 Found local images for {data['name']} ({s_id})")
# Copy first 5 images
found_imgs = list(local_face_dir.glob("*.jpg")) + list(local_face_dir.glob("*.png"))
for i, img_path in enumerate(found_imgs[:5]):
dest = person_dir / f"face_{i}{img_path.suffix}"
shutil.copy2(img_path, dest)
else:
print(f" ⚠️ No local images found for {data['name']} ({s_id}) at {local_face_dir}")
print(f"✅ Extraction Complete. Data saved to {OUTPUT_DIR}")
if __name__ == "__main__":
extract_indian_celebs()
|