docs: add concepts guide on Datatypes and Datashapes

[NickCrews]

inspired by #8358

[gforsyth]

Quick pass, but this is really nice @NickCrews -- thanks for putting it together! I've flagged a few typos and some styling nits, nothing major.

I do think that most of the code examples can be made executable (I'm not sure about the one that raises the RelationError...)

[ibis-docs-bot]

Docs preview: https://pr-8557-3de83120fe54bd5f4daa390a3fc0154049938e4e--ibis-quarto.netlify.app

[lostmygithubaccount]

I would make the code blocks runnable; use a try/except to catch the error and print it out

[cpcloud]

You can also set error: true for a given cell

[NickCrews]

Thanks for all the thoughts! I agree with all of them.

Two lingering thoughts:

1. I totally ignored Tables. This could be read that Tables also have a datatype and datashape, which is wrong. But I thought it might just be confusing? if you have a good way of putting that in there, would be great
2. curious what you think about how I described "flavor". Is that the term we want to use? Are there other aspects of that that are important? I almost left it out, but I think it's not too distracting.

[ibis-docs-bot]

Docs preview: https://pr-8557-fb53cc3f9ac98031d4bcbef3ce89d8251716edb6--ibis-quarto.netlify.app

[ibis-docs-bot]

Docs preview: https://pr-8557-344547446d1043e1ed6e18960354b04aa0613ec5--ibis-quarto.netlify.app

[gforsyth]

1. I totally ignored Tables. This could be read that Tables also have a datatype and datashape, which is wrong. But I thought it might just be confusing? if you have a good way of putting that in there, would be great

I think this is ok to ignore -- if you wanted to be explicit and make sure it's covered, I think I would introduce it in a callout note and say something like (unpolished):

A table is a collection of one or more columns. Because each column can have a different datatype and the number of columns is highly variable, a table itself does not have a datatype or a datashape.  

2. curious what you think about how I described "flavor". Is that the term we want to use? Are there other aspects of that that are important? I almost left it out, but I think it's not too distracting.

I liked it and I think it's a good balance of a word that will both communicate the idea to readers less familiar with datatypes but still correct enough for more familiar readers.

[gforsyth]

I think this can go in, but I'll hold off on merging in case you have any other tweaks you'd like to make @NickCrews . Thanks again for putting this together!

[NickCrews]

Thanks for the reviews! This looks good to me, merge it!

[cpcloud]

Gah, definitely meant to get in a review of this, but it was minor so I'll just put up a PR to change.