This application is intended as a very simple demonstration of how user-shared Experiences can be shared through a website and deposited into Open Humans.
The initial feature set includes:
- logging in into the application through Open Humans
- writing new experiences
- Consent management
- whether to allow research use of the experience
- whether to make this experience publicly available
- Users can toggle the research use and public status of each individual experience at any point.
- Each Experience is stored as an individual JSON file in Open Humans, annotated with additional metadata (whether the experience is designated to be made public & whether research use of it is allowed)
- When toggling the sharing/usage permission status those files will be updated to reflect the change
- Public Experiences are additionally stored in this web application's database for easier & faster display
- Experiences that are made public enter a moderation queue and need to be approved before joining a public list of experiences.
All experiences are stored on Open Humans as a simple JSON file, at this point just containing the timestamp of posting the experience and the experience text.
{
"text": "add another test",
"timestamp": "2019-07-18 10:43:56.549079"
}
Furthermore the application deposits additional metadata in Open Humans when uploading a file:
{
"id": 1234,
"basename": "testfile.json",
"created": "2019-07-18T10:43:59.709964Z",
"download_url": "https://www.openhumans.org/data-management/datafile-download/XXXXXX/?key=access_key",
"metadata":
{
"tags":
[
"viewable",
"non-research"
],
"uuid": "f1790562-a948-11e9-8161-8c859069dbc5",
"description": "this is a test file"
},
"source": "direct-sharing-267"
}
The uuid is used to correctly link publicly shared experiences to the files on Open Humans. The tags describe whether a file is marked as publicly shared or not and whether research use of the data is permitted.
When a user flags an experience as publicly availble a copy of it will be deposited in this applications own database. The Django model to store those is found in main/models.py. In addition to linking it to the user, it also stores the uuid, linking it to the deposited data in Open Humans.
In all cases the data should be stored on Open Humans and be considered the canonical copy. The data stored in this application's database (e.g. public experiences) is just done as a way to cache the data for faster loading times
In this very simplistic demo app there is no proper handling of ownership of Experiences or of moderator roles, e.g. everyone can approve/moderate experiences in the queue.
Logging into this app and storing data is handled through the Open Humans ecosystem. In order to use this application an associated project on Open Humans must be created. The following links provide further documentation on creating these projects & how to use the API.
- A general overview on OAuth2-based projects on Open Humans.
- How-to create a new project on Open Humans.
- This application uses the
django-open-humansPython library which provides a Django app that includes user management, the full OAuth2 workflow as well as additional functions to interact with the Open Humans API. - The
django-open-humanslibrary is based on a more genericopen-humans-apiPython library, which can be used in non-Django based Python applications.
This app is written in Python 3.6+, uses the Django 2.x framework and is designed to be ultimately deployed to Heroku. You will need some additional modules and packages to locally experiment with this uploader template or to develop it further. A full step-by-step guide that should work for Mac OS (and with minor differences for Linux) can be found in the INSTALL.md.
We'd love to get your contribution to this project, thanks so much for your interest in this! Please read our CONTRIBUTING.md to see how you can help and become part of our team! 🎉