ual-re/rem

A Python web UI built on FastAPI to gather metadata to construct a README

About Us

Research Engagement (RE) helps UA faculty, researchers, and students with every stage of their research. We have three functional units:

1. Research Incubator
2. Data Cooperative
3. Scholarly Communications

Members of the Data Cooperative:

GitHub	Name	Role/Title
	Jeff Oliver	Data Science Specialist
	Fernando Rios	Data Management Specialist
	Kiri Carini	GIS Specialist
	Jonathan Ratliff	Research Data Repository Assistant

Experience

Specialty	Tools/Resources
Data Science
Data Management & Publishing
GIS

rem's Issues

Refactor with ldcoolp-figshare

We packaged figshare curation API as ldcoolp-figshare. Let's now refactor our Figshare calls to simplify the code.

For this to work, we need to use v0.3.0 of ldcoolp-figshare

Feature: Work with Figshare curation metadata instead of published metadata

Currently we are using the metadata from published Figshare datasets. We now need to use the curation metadata to enable working with pending dataset that are under review.This will require handling with the Figshare API token.

Action items:

• Adjust figshare methods for required curation_id input
  • Add api_key and stage_api_key variable
  • Update unit tests
• Adjust intake_form for required curation_id input
  • Add curation_id to IntakeData
  • Update unit tests
• Add API tokens for Figshare (production and stage)
  • Add conftest.py to ensure parsing of tokens
  • Add configure_api() function in main.py for simplicity
  • Add settings_templates.json for production deployment
• Allow curation_id to be an optional field
• Adjust database search to use curation_id when available. Otherwise use previous results

Add examples with schema_extra

In the future, it would be great to have an example of the raw string in a comment.

Originally posted by @damian-romero in #4 (comment)

I believe these instructions will be helpful in doing so:

Rendering Markdown README.txt

Still doing some explorations but this stack overflow issue is of interest: https://stackoverflow.com/questions/35442329/visualizing-readme-md-files-in-my-website


Package code with setup.py and requirements

Initialize with:

• setup.py
• readme_tool package with __init__.py and initial version
• requirements.txt
• MANIFEST.in

Corresponding author can be filled with "nan" when no response recorded for Q6 in Qualtrics survey

When creating the readme for the following deposit, in which "N/A" appears as the value for each field of Q6 in the Deposit Agreement Qualtrics survey, the ReDATA Readme Tool generated a file with "nan" as the value for Corresponding Author.

Add form module to intake of metadata

A primary form module is needed to:

• [x] Populate a page with Figshare metadata from the figshare:metadata_get endpoint
• Ask respondents to fill in information
• Upon submission that is stored in a database (perhaps NoSQL)
• The endpoint will have input of article_id and curation_id to store it in the database
  • article_id
  • curation_id

Note: curation_id will be later given complexity with external API token auth

Chore: Change routing paths for figshare, change method names for consistent

Let's simplify the API routing paths for the figshare and intake_form module.

Also, let's change the method namespace so it's [API_METHOD]_[FUNCTION]


Module to retrieve Figshare metadata for public deposit (no token)

A figshare module is needed to retrieve Figshare metadata from its API. Allows for GET method with an input of the article_id and optional input of stage.

Responses should contain the metadata reduced to what we need for the README form. This includes:

1. Article ID
2. Title
3. Description
4. DOI
5. Preferred citation
6. License
7. References

• API route to get Figshare article data, figshare_get
• API route to extract only metadata for README form, metadata_get
• Function to output only metadata for README form, figshare_metadata_readme
• main.py for unicorn run
• pytest and coverage configuration files
• unit testing of figshare module

In addition:

• GitHub actions python CI build and testing

Switch from uvicorn.Server to router approach for multi modules

We are currently using uvicorn.Server to handle the different modules and having them on different ports. It would be better to consolidate them and use multiple routers.


Feature: Containerized for deployment

Here are some resource for nginx deployment:
Blob: https://levelup.gitconnected.com/deploying-an-asynchronous-fastapi-on-nginx-unit-b038288bec5
nginx unit: https://unit.nginx.org/howto/fastapi/