Video tutorials

[wolny]

Add short videos with description for key functionalities:

• proofreading (Lorenzo)
• network prediction (Adrian)
• segmentation (Adrian)
• pre-/post-processing (Lorenzo)

[qin-yu]

When you think it's time, I can quickly put them in GitHub Pages by Actions in readthedoc style. I've been using that for another project. If you want the changes to be reflected in commits, we can start migrating from the existing Wiki now.

We can do either:

1. we update the doc in main branch, and when we want to officially update the doc, we sync the publication branch with our main, which triggers the Action that compiles it into a real doc on GitHub Pages.
2. we update the doc in main branch, and the Action updates the doc webpage on each commit.

The raw doc can be a standalone repo, or it can be a /doc folder.

[lorenzocerrone]

Hey @qin-yu that would be amazing, do you have some time to setup the template? If yes I agree to move the docs to github pages

[qin-yu]

@lorenzocerrone maybe by Monday. Is it fine?

[lorenzocerrone]

That would awesome! But I would say that anytime next week would be good, no need to crunch for finishing it by monday :)

[qin-yu]

@lorenzocerrone I got covid on Saturday : )

• See #179

[lorenzocerrone]

Oh no! :( I hope you have speedy recovery!

[qin-yu]

@lorenzocerrone @wolny

• https://github.com/hci-unihd/plant-seg/tree/master/docs/chapters contains the markdown files for documentation
• https://github.com/hci-unihd/plant-seg/blob/master/docs/_toc.yml is the table of content <-- this needs to be modified if you want to add/remove/rename markdowns in docs/chapters/
• Website of docs updates on every new push on master, see https://github.com/hci-unihd/plant-seg/blob/master/.github/workflows/build-deploy-book.yml

[qin-yu]

Oops forgot to tell you the local steps @lorenzocerrone @wolny :

1. create a new env and pip install -r docs/requirements.txt to use Jupyter Book locally for visualisation
2. run jupyter-book build . --all to build and use a browser to check

[qin-yu]

Sphinx systems allows API reference from docstrings. For the API section of our documentation, we can refer to this section: https://jupyterbook.org/en/stable/advanced/developers.html#api-reference-from-docstrings

[qin-yu]

Current jupyter book documentation site is decent. I used it with the expectation that code blocks could run at build. But it may not be as practical as I thought.

Here I found many packages uses this mkdocs framework, which looks better though it's more static:

[lorenzocerrone]

I have a very small preference for mkdocs, but both are equally good to me