image_classifier/fashion_mnist/ingest.py

from abcli import file
from abcli import string
from abcli.logging import crash_report
import os.path
from abcli import logging
import tensorflow as tf
import logging

logger = logging.getLogger(__name__)


def ingest(output_path):
    import tensorflow as tf

    try:
        fashion_mnist = tf.keras.datasets.fashion_mnist

        (train_images, train_labels), (
            test_images,
            test_labels,
        ) = fashion_mnist.load_data()
    except:
        crash_report("-fashion_mnist: ingest.")
        return False

    logger.info("ingesting fashion_mnist")

    success = True
    for name, thing in zip(
        "train_images,train_labels,test_images,test_labels".split(","),
        [train_images, train_labels, test_images, test_labels],
    ):
        if file.save(os.path.join(output_path, f"{name}.pyndarray"), thing):
            logger.info(f"ingested {name}: {string.pretty_shape_of_matrix(thing)}")
        else:
            success = False

    class_names = [
        "T-shirt/top",
        "Trouser",
        "Pullover",
        "Dress",
        "Coat",
        "Sandal",
        "Shirt",
        "Sneaker",
        "Bag",
        "Ankle boot",
    ]
    if file.save_json(os.path.join(output_path, "class_names.json"), class_names):
        logger.info(
            f"ingested {len(class_names)} class name(s): {', '.join(class_names)}"
        )
    else:
        success = False

    return success