create_jsonl_file_for_vertx_ai.py

import json

from google.cloud import storage

bucket_name = ""
folder_name = "unchecked_checkbox"


def list_items_in_folder(folder_path):
    client = storage.Client()
    bucket = client.get_bucket(bucket_name)

    # List objects in the specified folder path
    blobs = bucket.list_blobs(prefix=folder_path)

    # Print the names of the items in the folder
    list_of_items = []
    for blob in blobs:
        data = {
            "imageGcsUri": f"gs://{bucket_name}/{blob.name}",
            "classificationAnnotation": {
                "displayName": f"{folder_name}"
            }
        }
        list_of_items.append(data)
    with open(f'{folder_name}.jsonl', 'w') as file:
        for entry in list_of_items:
            json.dump(entry, file)
            file.write('\n')


list_items_in_folder(
    'images_for_vision_ai/new_dataset/cropped_images_v2/unchecked'
)