Documentation

Full documentation of CVEDIA API is available on https://docs.cvedia.com To see all options cvedia_api.py provides, run python cvedia_api.py --help

Quick Start

1. Register the app: python cvedia_api.py --register
2. Create a dataset: python cvedia_api.py --create_dataset examples/create_dataset.json -- You might have an error if the name already exists
3. Upload images to dataset: python cvedia_api.py --dataset_index <index name> --upload examples/image.json
4. Upload annotations to the image: python cvedia_api.py --dataset_index <index name> --upload examples/annotation.json
5. Bind custom arbitrary values to a image entity: python cvedia_api.py --dataset_index <index name> --upload examples/image_custom_meta.json
   • Query arbitrary values for exporting: python cvedia_api.py --dataset_index <index name> --export examples/export_custom_mql.json
6. Replace an existing image with a different one: python cvedia_api.py --dataset_index <index name> --upload examples/image_replace.json
7. Output cifar-10 dataset with augmentations and filters: python cvedia_api.py --dataset_index cifar-10 --export examples/export_cifar_mql.json

What is this

This is a collection of tools that shows the API functionality of CVEDIA API.

How to install

You can test either in python or php.

Linux install

On a ubuntu system you probably already have that installed, to make sure just run:

apt-get install python -y

Windows install

Download python 2.7 (python 3 would also work) here:

https://www.python.org/ftp/python/2.7/python-2.7.amd64.msi

Running

python cvedia_api.py --help

Shows all possible options.

Before using this you need to register an app, otherwise you won't be able to access your private content, for that you can run:

python cvedia_api.py --register

This will show a set of instructions you will need to follow in order to register the app and be able to authenticate it. This process only needs to happen once as after the app is registered it will save a configuration file with the tokens.

Note that by registering an app you will allow it to access all your user data on cvedia, including private datasets.

Examples

python cvedia_api.py --datasets

Lists all datasets from the system plus the ones you own / have access to

python cvedia_api.py --projects mydataset-hrao

Lists all projects from mydataset-hrao

python cvedia_api.py --dataset_index mydataset-hrao --dataset_type train --upload /storage/path/to/dataset /storage/path/to/dataset2 --threads 64

Will recusevely upload all the files from folders /storage/path/to/dataset /storage/path/to/dataset2 in 64 threads to dataset index mydataset-hrao on type train

Note

Dataset name is different from dataset index, usually api commands only use index, name only used on the frontend. You should be able to find the name using the dataset list function.