Getting images embeddings samples from robotoff

get_images_empbeddings.py

# run it using docker-compose run --rm worker_high_1 python3

from robotoff.prediction.category.neural.keras_category_classifier_3_0 import *
from robotoff.off import *
from robotoff.models import *
import json
import os

product_id = ProductIdentifier(barcode="3017620422003", server_type=ServerType.off)

product = get_product(product_id, fields=["images"])
image_ids_int = sorted((int(image_id) for image_id in product.get("images", {}) if image_id.isdigit()),reverse=True,)[:MAX_IMAGE_EMBEDDING]
image_ids = [str(image_id) for image_id in image_ids_int]

with db:
    embeddings_by_id = fetch_cached_image_embeddings(product_id, image_ids)

os.mkdir("/opt/robotoff/data/tmp")
json.dump({k: v.tolist() for k, v in embeddings_by_id.items()}, open(f"/opt/robotoff/data/tmp/{product_id.barcode}_embeddings.json", 'w'), sort_keys=True, indent=4)

# get images in ./data/tmp