Create a basic version of a data package?

[joelostblom]

Since this template is a starting point for new users of Sprout, I think it can be confusing that there are so many files created from the start. Particularly as most of these are configuration files that are not related to working with the actual data or metadata.

Currently, a new user who copies this template, sees all of this in their directory:

.git/
.github/
.venv/
.vscode/
docs/
scripts/
.copier-answers.yml
.cz.toml
.editorconfig
.gitignore
.pre-commit-config.yaml
.python-version
.typos.toml
CITATION.cff
LICENSE.md
LICENSE-MIT.md
README.md
justfile
main.py
pyproject.toml
ruff.toml
uv.lock

Of these files, the only one that will actually be edited is main.py (I think). The rest are different configs. Are all these config files helping sprout to carry out its purpose? Or are some of them an artifact from how the Seedcase team works but are not necessarily relevant for anyone cloning the template repo? E.g. do we need to enforce the editor config settings for every user using the package or could files like .editorconfig and the .vscode folder be considered for exclusion from the template?

Other questions include if we need to enforce checking for typos, combine the two license files into one, and if the user is supposed to know what to do with the justfile; it currently have many commands in it but we don't document how they are expected to be used.

A potential alternative to deletion would be to move some of these config files to the post creation setup, where some Seedcase specific settings are currently documented.

[joelostblom]

If all these files are considered important, then maybe another approach would be to document really clearly that the user can ignore most of them. My goal with this issue is to make the "getting started" process smoother and I think it is generally difficult coming into a new project and start out having 20ish config files in the repo, before even getting to the part of creating/modifying data/metadata.

[signekb]

I agree that this amount of files will be overwhelming for most people starting to use this package.

Maybe we should have a minimal/basic version as default to get started, and then keep the possibility to add (some?) of these config files with a copier flag/question?

[lwjohnst86]

These are all development files that help in developing a Data Package in a higher-quality, more rigorous way. We shouldn't remove them, but having a "basic" version would help I think 😄

[lwjohnst86]

Another thing, this template isn't to help with "carrying out Sprout's purpose", this template is to help develop a Data Package... which happens to also use our Seedcase packages like Sprout, check-datapackage, etc etc

[joelostblom]

Yes, sorry that phrasing was unspecific. My general point is that if I were to download a package to e.g. help me with data visualization and then it also enforces that I use tabs instead of spaces in my editor, I would be a bit annoyed, even if there was an option to opt out. I do get that a certain level of opinionated sane defaults is beneficial, and here I'm asking where the line goes for what to include given that the purpose is to develop data packages in a higher quality, more rigorous manner. I think having a basic version and and opt-in would be helpful and solves this issue to some extent, but we would still need to decide what goes in the basic version.

[lwjohnst86]

I think the basic version would be very basic, just what is needed to make a data package (JSON file, main.py, etc). But (as you know) I am very opinionated about things 😛 if someone is annoyed, they don't have to use it 😛 But also, I wouldn't expect someone to use this template on top of an existing repository (in general). This would be a new project/repo, so nothing would be affected that way.

Finally, this template is for us first and foremost. See https://design.seedcase-project.org/organisation/contribution#we-are-our-first-users 😝 So whatever we have here is for us and how we work.

[lwjohnst86]

Related to using the data in the Data Package, that's a whole separate thing. The template is only to develop and store the Data Package, not for use to analyse it. That will be coming later with e.g. propagate package.