Create a catalog of plasma and fusion software

[namurphy]

Description

We should create a page on the site that includes a catalog of research software that is used in plasma and fusion energy sciences.

Motivation

It is presently very difficult to find research software for plasma and fusion energy sciences. It would be really helpful to have a single site that people can go to in order to find out what software exists, how to access documentation for it, whether it's open source, and so on.

Implementation details

Here are my initial thoughts regarding the implementation, though we should continue to discuss this below. We should also create sub-issues as needed since we will probably need to do multiple pull requests to implement this.

• Project data should be stored in a standard format like YAML or TOML. Ideally, I'd like to find an appropriate metadata standard that we could use.
• The table should be dynamically generated from the data files.
• We should include software regardless of whether it is open source or not (or more specifically, whether it is available under an OSI-approved license).
• The site would probably need a taxonomy to describe the software. There could be high-level tags like simulation, data analysis, & visualization; with more fine-grained tags like particle-in-cell and gyrokinetic to provide more detail.
• There should be an easy way to list projects that have a certain tag. For example, a visitor might want to show or search for only projects that are open source.
• The site would need clear instructions on how someone could go about adding a project via a pull request.

Note

We'd eventually want to apply a validation tool to make sure that each entry meets the schema and includes all required fields (probably along the lines of name, description, url). This shouldn't be in the first pull request, but doing this sooner rather than later will ultimately make our lives easier.

An example implementation from the heliopythoniverse

We could get some inspiration from the Python in Heliophysics Community (PyHC) website, which has a page that includes a table of projects in the heliopythoniverse.

There are different fields informational fields such as:

• Name
• Description
• Code (link to Git{Hub,Lab} repo)
• Docs
• Project site
• Contact

It also has columns where projects do a self-evaluation and assess their status (good, partially met, requires improvement) around:

• Community
• Documentation
• Testing
• Software Maturity
• License

Information about the projects is stored in projects.yml, projects_core.yml, and projects_unevaluated.yml.

The fields from the PyHC website are similar to what we need, but there are some differences. The PyHC projects are all based on Python, so there isn't as much of a need to list the language(s) of the project or provide a direct link to installation instructions (since nearly all can be installed with a command like pip install package-name).

There has also been work done in the heliopythoniverse about creating a taxonomy, though I don't have a link handy at the moment.

[namurphy]

We've also attempted to list affiliated and community packages on the PlasmaPy website, but we haven't had the resources to put a lot of time into this unfortunately. I think it will ultimately be more sensible to have a page to list projects on this site.

[ZedThree]

@namurphy This is something I have been meaning to do for PlasmaFAIR, and you've basically laid out everything I have been thinking!

I have previously looked at different software catalogues, need to dig out my notes again. I don't think I found any completely obvious candidates. There's https://research-software-directory.org which is more for institutions to show off their research software, and really we want to signpost people.

Fortran-lang does a similar thing to PyHC, with a yaml file as the store for the metadata, and webpages generated from that file.

In terms of metadata schema, DataCite has an example for software, and it looks very reasonable. I'm not too concerned about getting the schema right initially, they usually have ways to convert between schema.

What I propose to do to get the ball rolling is to make a PR now* starting to collect software in a yaml file, probably just very basic and bare bones to begin with, while we also investigate what a suitable front end should look like.

*when I have time