The Serverless Image Data Repository contains datasets of randomly-generated images.
(The names of the datasets and images are also random.)
Examples
|
|
|
|
Data Access
This repository demonstrates how to host dataset files in Globus collections while providing a publicly available site for data discovery. Here are the notes on setting up the Globus Connect Server and Guest Collection.
The site is hosted using GitHub Pages a free service that can deploy sites written in Markdown. Because the Globus collection hosting the datasets provides HTTPS access, images from the collection are embedded in catalog pages.
This site represents the catalog, with metadata about the catalog itself and the datasets in the repository. The catalog is completely public while the datasets are grouped into three levels of access:
Public Access
Public is completely public, including anonymous access via HTTPS.
Accept Terms & Conditions (Register)
Accept Terms & Conditions means that a user must join the the Serverless Data Users Globus Group, provided they accept the terms & conditions. All Globus users can join the Group without any actions by repository personnel. This is the “just sign up” model where joining the Group provides a list of registered users.
Moderated Access (Restricted)
Moderated Access means that the user can ask to join the Serverless Data Project One Globus Group, but a group manager must approve members. This is the “request access” model.
To support datasets with different access requirements, datasets could be gather into folders with access tied to various Globus Groups. This also allows different projects to curate access to their datasets, where PIs or their delegates can be assigned as Group managers. Then they receive and decide on access requests.
Prior Art
An inspiration for this concept is the X-ray Tomography Data Bank. TomoBank uses a Globus collection at Argonne to store files and publishes descriptions of its datasets on ReadTheDocs. Because the dataset metadata is on Github in the TomoBank repo, users can submit new datasets through pull requests.
Datasets
| Access | Name | Creator | Created | Number of Files |
| Public | (Mock) BOUNDEDNESS285G | Rick Wagner | 2022-08-30 | 28 |
| Public | (Mock) PRONENESS405F | Rick Wagner | 2022-09-06 | 28 |
| All Users | (Mock) NONMUCILAGINOUS197C | Rick Wagner | 2022-09-03 | 28 |
| All Users | (Mock) ORLEWISE861S | Rick Wagner | 2022-09-08 | 28 |
| All Users | (Mock) PSYCHOSOCIAL631N | Rick Wagner | 2022-08-24 | 28 |
| All Users | (Mock) CENTILLION440W | Rick Wagner | 2022-09-12 | 28 |
| All Users | (Mock) MAGA679S | Rick Wagner | 2022-08-29 | 28 |
| Restricted | (Mock) POLYBORINAE498S | Rick Wagner | 2022-08-24 | 28 |
| Restricted | (Mock) AUTACOIDAL812S | Rick Wagner | 2022-09-11 | 28 |
| Restricted | (Mock) MAGNETOMOTIVE835Q | Rick Wagner | 2022-09-06 | 28 |
| Restricted | (Mock) DECORABLE727P | Rick Wagner | 2022-09-06 | 28 |
| Restricted | (Mock) HISTOZOIC885G | Rick Wagner | 2022-09-14 | 28 |