Filterable DataLad cheat sheet

[jsheunis]

@datalad/developers what do you think about having a page on the website with content that covers (an updated version of) the handbook cheat sheet, but that also adds filtering. For those times when, as a user, you quickly want to check the available options of a command, or find the right command for your use case. It could also include some useful git commands in the context of datalad.

I don't think this is really important, more a nice to have. And of course it might be a burden to keep up to date with datalad updates.

[yarikoptic]

we definetely probably would think thrice if to create another resource to maintain/keep in sync manually to be aligned with current interfaces etc. But we do have all interfaces programmed etc, so as a part of the release process (or in master, or any other version) can produce whatever we would like to produce from code to populate whatever else desired to populate ;)

Since I do not yet see what you envision @jsheunis - I find it hard to suggest anything more specific, but the point is -- may be whatever you envision could be automated from what we have in the datalad core codebase? we even have _examples_ for many commands for both CLI and Python API. e.g. https://github.com/datalad/datalad/blob/HEAD/datalad/distributed/drop.py#L79 -- would be also nice to expose more

[jsheunis]

Good suggestions, thanks. I was thinking that a quick lookup mechanism for datalad commands/options/examples could be handy, as opposed to first trying to navigate the docs/handbook for an answer. Similar to the filtering mechanism here: https://www.datalad.org/in-the-wild.html. If, as a user, I'm e.g. searching for the right way to use datalad clone, I'd type in clone and it will immediately display a clear command use + options +example. But I agree it wouldn't make sense to duplicate anything unless there's (straightforward) automation involved.

[yarikoptic]

overall -- love the idea.

I always love simplicity of google -- just type in the "keywords" and get something to glance through or to restrict query to become more specific (quiz: what is my favorite datalad search query? ;-) ) . Similarish is github -- search for something, it groups it into issues, users, etc.
So, if there is a dump of all commands and all examples, all those could be listed with commands leading to e.g. the ones under https://docs.datalad.org/en/latest/cmdline.html, with examples also leading to corresponding commands they belong to, etc.?

But then overall it starts to feel like I am suggesting to have a search engine over our content on docs.datalad.org -- may be that is what we need, the search over docs.datalad.org, handbook.datalad.org, and datalad.org (may be even neurostars posts for datalad ;-) ) ? could even (ab)use google for that, e.g. could probably do via google like what https://git-annex.branchable.com/ and https://www.repronim.org/search.html do? but that derails from original aim I think too much ;-)