Meshcoord printout [AVD-1592]

[pp-mo]

🚀 Pull Request

Description

Added string representations for MeshCoord

Waiting until #4053 is in : Which this is based on.
(so it won't merge into the target branch, as yet)

I have made the str/repr versions slightly different, following discussion with @bjlittle
I'm still a little uncertain that the results are the most useful, though:

• if you enter a MeshCoord at the command line, you get "repr", which shows a mesh as <Mesh object at 0x{address}>,
  • whereas the "str" form attempts a more human-readable account of the Mesh, as Mesh('name')
• This corresponds with the "standard" Python usage ...
• ... but, for a cube, the "repr" version is the short form, whereas print(cube) gives full details.
  • .. so making repr the "more precise and less human" is against what we do for cubes.

---

Consult Iris pull request check list

[bjlittle]

* .. so making repr the "more precise and less human" is **_against_** what we do for cubes.

@pp-mo Point taken.

To be honest we should rationalise and align all our __repr__ and __str__ usage throughout iris... it's well overdue and the state of affairs that we find ouselves in is a fair reflection on the unopinionated, organic way in which iris has developed over the years

[bjlittle]

@pp-mo Some suggested changes... nothing major. See what you think.

Otherwise LGTM 👍

[pp-mo]

Rebased onto #4053 (effectively).
But, this may need doing again.

In the meantime, attending to the existing comments above (though now outdated).

[pp-mo]

I think that's it for the outstanding review issues.
I'm taking this out of Draft, but it still won't look nice here (or merge into the target) until #4053 is in.

[pp-mo]

Rebased again ( 😓 )
Ready to go @bjlittle !