Connect to AI-on-Demand in Python.

The AI-on-Demand (AIOD) platform empowers AI research and innovation for industry and academia. At its core if the metadata catalogue which indexes countless AI resources, such as datasets, papers, and educational material, from many different platforms such as Zenodo, OpenML, and AIDA.

This package is mainly intended for users that want to programmatically access the platform to, e.g., build a service, fetch resources in their scripts, or write a connector that registers metadata of another platform with AI-on-Demand. A web interface for browsing and registering assets is available on AI-on-Demand, through MyLibrary and the Metadata Catalogue Editor services (to be deployed), respectively.

Installation

The aiondemand package is on PyPI:

$ pip install aiondemand

Tip: install your dependencies in a virtual environment.

Usage

You can directly access endpoints through the Python API, for example to browse datasets:

import aiod

aiod.datasets.get_list()

And results will be returned as a Pandas dataframe (though the data_format may be used to get JSON instead):

      platform platform_resource_identifier                    name       date_published                                            same_as  is_accessible_for_free  ...  relevant_link  relevant_resource relevant_to research_area scientific_domain identifier
0  huggingface       acronym_identification  acronym_identification  2022-03-02T23:29:22  https://huggingface.co/datasets/acronym_identi...                    True  ...             []                 []          []            []                []          1
...
9  huggingface              allegro_reviews         allegro_reviews  2022-03-02T23:29:22    https://huggingface.co/datasets/allegro_reviews                    True  ...             []                 []          []            []                []         10

[10 rows x 30 columns]

You can even query the elastic search endpoints:

aiod.publications.search(query="Robotics")

      platform platform_resource_identifier                                               name date_published                                            same_as is_accessible_for_free  ... relevant_resource relevant_to      research_area  scientific_domain  type  identifier
0  robotics4eu                         1803  Responsible Robotics &amp; non-tech barriers t...           None  https://www.robotics4eu.eu/publications/respon...                   None  ...                []          []  [other materials]  [other materials]  None           4

[1 rows x 36 columns]

Contributing

Interested in contributing? Check out the contributing guidelines. By contributing to this project, you agree to abide by our Code of Conduct.

Credits

The aiondemand package is being developed with funding from EU’s Horizon Europe research and innovation program under grant agreement No. 101070000 (AI4EUROPE). Not all contributors need be affiliated with this funding.

cookiecutter and the py-pkgs-cookiecutter template were used to create the repository structure.