face_matching_mvp / scripts /prepare_database.py
pray's picture
support gpu and parallel process
738d527
"""
Data Preparation Pipeline
Downloads images, detects faces, extracts embeddings, and populates Milvus database
"""
import os
import csv
import requests
import cv2
import numpy as np
import argparse
import logging
from typing import List, Tuple
from tqdm import tqdm
from concurrent.futures import ThreadPoolExecutor, as_completed
import sys
from pathlib import Path
# Add src directory to Python path
sys.path.insert(0, str(Path(__file__).parent.parent / "src"))
from face_matcher.config import config, Config
from face_matcher.core.detection import FaceDetectorFactory
from face_matcher.core.recognition import FaceEmbeddingExtractor
from face_matcher.core.database import VectorDatabase
# Setup logging
logging.basicConfig(
level=logging.INFO,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)
def download_image(url: str, save_path: str, timeout: int = 10) -> bool:
"""
Download image from URL
Args:
url: Image URL
save_path: Path to save the image
timeout: Request timeout in seconds
Returns:
True if successful, False otherwise
"""
try:
response = requests.get(url, timeout=timeout, stream=True)
response.raise_for_status()
with open(save_path, 'wb') as f:
for chunk in response.iter_content(chunk_size=8192):
f.write(chunk)
return True
except Exception as e:
logger.debug(f"Failed to download {url}: {e}")
return False
def download_images_parallel(download_tasks: List[Tuple[str, str]], max_workers: int = 10) -> List[bool]:
"""
Download multiple images in parallel
Args:
download_tasks: List of (url, save_path) tuples
max_workers: Maximum number of parallel downloads
Returns:
List of success status for each download
"""
results = [False] * len(download_tasks)
with ThreadPoolExecutor(max_workers=max_workers) as executor:
# Submit all download tasks
future_to_idx = {
executor.submit(download_image, url, save_path): idx
for idx, (url, save_path) in enumerate(download_tasks)
}
# Collect results as they complete
for future in as_completed(future_to_idx):
idx = future_to_idx[future]
try:
results[idx] = future.result()
except Exception as e:
logger.debug(f"Download task {idx} failed: {e}")
results[idx] = False
return results
def process_dataset(
csv_path: str,
output_dir: str,
detector_type: str = "retinaface",
max_images: int = None,
skip_existing: bool = True,
parallel: bool = True,
max_workers: int = 20
) -> Tuple[List[str], List[str], List[str], List[np.ndarray]]:
"""
Process dataset: download images and detect faces
Args:
csv_path: Path to metadata CSV file
output_dir: Directory to save aligned face crops
detector_type: "retinaface" or "haarcascade"
max_images: Maximum number of images to process
skip_existing: Skip if aligned face already exists
parallel: Use parallel processing for downloads (default: True)
max_workers: Number of parallel workers for downloads (default: 20)
Returns:
Tuple of (names, image_paths, original_paths, aligned_faces)
"""
# Create output directories
download_dir = os.path.join(output_dir, "downloads")
aligned_dir = os.path.join(output_dir, "aligned_faces")
cropped_dir = os.path.join(output_dir, "cropped_faces") # Original face crops
os.makedirs(download_dir, exist_ok=True)
os.makedirs(aligned_dir, exist_ok=True)
os.makedirs(cropped_dir, exist_ok=True)
# Initialize face detector
logger.info(f"Initializing {detector_type} detector...")
detector = FaceDetectorFactory.create_detector(detector_type)
# Read CSV
with open(csv_path, 'r') as f:
reader = csv.DictReader(f)
rows = list(reader)
if max_images:
rows = rows[:max_images]
logger.info(f"Processing {len(rows)} images with {detector_type} detector...")
# Step 1: Prepare file paths and identify images to download
download_tasks = []
row_info = [] # Store (row, download_path, aligned_path, cropped_path)
for row in rows:
name = row['name']
image_id = row['image_id']
url = row['url']
# Create filenames
download_filename = f"{name}_{image_id}.jpg".replace(" ", "_")
download_path = os.path.join(download_dir, download_filename)
aligned_filename = f"{name}_{image_id}_aligned.jpg".replace(" ", "_")
aligned_path = os.path.join(aligned_dir, aligned_filename)
cropped_filename = f"{name}_{image_id}_crop.jpg".replace(" ", "_")
cropped_path = os.path.join(cropped_dir, cropped_filename)
row_info.append((row, download_path, aligned_path, cropped_path))
# Add to download queue if image doesn't exist
if not os.path.exists(download_path):
download_tasks.append((url, download_path))
# Step 2: Download missing images (parallel or sequential)
if download_tasks:
if parallel:
logger.info(f"Downloading {len(download_tasks)} images in parallel (workers={max_workers})...")
download_images_parallel(download_tasks, max_workers=max_workers)
else:
logger.info(f"Downloading {len(download_tasks)} images sequentially...")
for url, save_path in tqdm(download_tasks, desc="Downloading images"):
download_image(url, save_path)
# Step 3: Process faces (detection and alignment)
names = []
image_paths = []
original_paths = []
aligned_faces = []
for row, download_path, aligned_path, cropped_path in tqdm(row_info, desc="Detecting and aligning faces"):
name = row['name']
# Skip if both aligned and cropped faces already exist
if skip_existing and os.path.exists(aligned_path) and os.path.exists(cropped_path):
aligned_face = cv2.imread(aligned_path)
if aligned_face is not None:
names.append(name)
image_paths.append(cropped_path)
original_paths.append(download_path)
aligned_faces.append(aligned_face)
continue
# Load image
if not os.path.exists(download_path):
continue
image = cv2.imread(download_path)
if image is None:
continue
# Detect and get both aligned and original crop
aligned_face, original_crop = detector.detect_and_crop(image)
if aligned_face is None or original_crop is None:
continue
# Save both versions
cv2.imwrite(aligned_path, aligned_face)
cv2.imwrite(cropped_path, original_crop)
# Store results
names.append(name)
image_paths.append(cropped_path)
original_paths.append(download_path)
aligned_faces.append(aligned_face)
logger.info(f"Successfully processed {len(names)} faces")
return names, image_paths, original_paths, aligned_faces
def extract_embeddings(
aligned_faces: List[np.ndarray],
model_path: str,
device: str = "cuda",
batch_size: int = 32,
use_batch: bool = True
) -> np.ndarray:
"""
Extract face embeddings using MobileFaceNet
Args:
aligned_faces: List of aligned face images
model_path: Path to ONNX model
device: Device to use ("cuda" or "cpu")
batch_size: Batch size for batch processing (default: 32)
use_batch: Use batch processing for faster inference (default: True)
Returns:
Numpy array of embeddings (N x embedding_dim)
"""
# Initialize embedding extractor
extractor = FaceEmbeddingExtractor(model_path, device=device)
if use_batch and batch_size > 1:
logger.info(f"Extracting embeddings with batch processing (batch_size={batch_size}, device={device})...")
embeddings = []
num_batches = (len(aligned_faces) + batch_size - 1) // batch_size
for i in tqdm(range(0, len(aligned_faces), batch_size),
desc="Extracting embeddings", total=num_batches):
batch = aligned_faces[i:i + batch_size]
batch_embeddings = extractor.extract_embeddings_batch(batch)
embeddings.append(batch_embeddings)
embeddings_array = np.vstack(embeddings)
else:
logger.info(f"Extracting embeddings sequentially (device={device})...")
embeddings = []
for aligned_face in tqdm(aligned_faces, desc="Extracting embeddings"):
embedding = extractor.extract_embedding(aligned_face)
embeddings.append(embedding)
embeddings_array = np.array(embeddings, dtype=np.float32)
logger.info(f"Extracted {len(embeddings_array)} embeddings with shape {embeddings_array.shape}")
return embeddings_array
def populate_database(
names: List[str],
image_paths: List[str],
original_paths: List[str],
embeddings: np.ndarray,
db_path: str,
collection_name: str,
batch_size: int = 100,
reset_database: bool = False
):
"""
Populate Milvus database with face embeddings
Args:
names: List of person names
image_paths: List of cropped face image paths
original_paths: List of original downloaded image paths
embeddings: Numpy array of embeddings
db_path: Path to Milvus database
collection_name: Name of collection
batch_size: Batch size for insertion
reset_database: If True, drop existing collection and create new one
"""
logger.info("Populating Milvus database...")
# Initialize vector database
db = VectorDatabase(db_path=db_path, collection_name=collection_name)
# Create collection (drop existing if reset_database=True)
db.create_collection(drop_existing=reset_database)
# Create index if doesn't exist
db.create_index()
# Load collection
db.load_collection()
# Insert embeddings in batches
total_inserted = 0
for i in range(0, len(names), batch_size):
batch_names = names[i:i+batch_size]
batch_paths = image_paths[i:i+batch_size]
batch_original_paths = original_paths[i:i+batch_size]
batch_embeddings = embeddings[i:i+batch_size]
num_inserted = db.insert_batch(batch_names, batch_paths, batch_original_paths, batch_embeddings)
total_inserted += num_inserted
# Get stats
stats = db.get_stats()
logger.info(f"Database populated successfully!")
logger.info(f"Total entities: {stats['num_entities']}")
db.close()
def main():
parser = argparse.ArgumentParser(
description='Prepare Face Matching Database',
formatter_class=argparse.ArgumentDefaultsHelpFormatter
)
parser.add_argument(
'--csv',
type=str,
default=config.data.csv_path,
help='Path to metadata CSV file'
)
parser.add_argument(
'--output_dir',
type=str,
default=config.data.output_dir,
help='Output directory for processed images'
)
parser.add_argument(
'--detector',
type=str,
choices=['retinaface', 'haarcascade'],
default=config.detector.default_detector,
help='Face detector to use'
)
parser.add_argument(
'--model',
type=str,
default=config.model.model_path,
help='Path to ONNX model'
)
parser.add_argument(
'--db_path',
type=str,
default=config.database.db_path,
help='Path to Milvus database'
)
parser.add_argument(
'--max_images',
type=int,
default=config.data.max_images,
help='Maximum number of images to process'
)
parser.add_argument(
'--skip_existing',
action='store_true',
default=config.data.skip_existing,
help='Skip processing if aligned face already exists'
)
parser.add_argument(
'--reset_database',
action='store_true',
default=False,
help='Reset database by dropping existing collection and creating new one'
)
parser.add_argument(
'--device',
type=str,
choices=['cuda', 'cpu'],
default='cuda',
help='Device for ONNX inference: cuda (GPU) or cpu (default: cuda)'
)
parser.add_argument(
'--parallel',
action='store_true',
default=True,
help='Use parallel processing for downloads and batch inference (default: True)'
)
parser.add_argument(
'--no-parallel',
dest='parallel',
action='store_false',
help='Disable parallel processing, use sequential mode'
)
parser.add_argument(
'--batch_size',
type=int,
default=32,
help='Batch size for embedding extraction (default: 32, only used with --parallel)'
)
parser.add_argument(
'--max_workers',
type=int,
default=20,
help='Number of parallel workers for image downloads (default: 20, only used with --parallel)'
)
args = parser.parse_args()
logger.info("="*60)
logger.info("Face Matching MVP - Database Preparation")
logger.info("="*60)
# Create necessary directories
Config.create_directories()
# Log configuration
logger.info(f"Configuration:")
logger.info(f" Device: {args.device}")
logger.info(f" Parallel mode: {args.parallel}")
if args.parallel:
logger.info(f" Batch size: {args.batch_size}")
logger.info(f" Max workers: {args.max_workers}")
# Process dataset
names, image_paths, original_paths, aligned_faces = process_dataset(
csv_path=args.csv,
output_dir=args.output_dir,
detector_type=args.detector,
max_images=args.max_images,
skip_existing=args.skip_existing,
parallel=args.parallel,
max_workers=args.max_workers
)
if len(names) == 0:
logger.error("No faces detected. Exiting.")
return
# Extract embeddings
embeddings = extract_embeddings(
aligned_faces,
model_path=args.model,
device=args.device,
batch_size=args.batch_size if args.parallel else 1,
use_batch=args.parallel
)
# Populate database
populate_database(
names,
image_paths,
original_paths,
embeddings,
db_path=args.db_path,
collection_name=config.database.collection_name,
batch_size=config.data.batch_size,
reset_database=args.reset_database
)
logger.info("="*60)
logger.info("Database preparation completed!")
logger.info("="*60)
if __name__ == "__main__":
main()