Added cube project function to iris.analysis.cartography

[esc24]

No description provided.

[pelson]

Is it your intention that this will always be the case?

[esc24]

No, but I found limiting the extent in a general, robust way was non trivial.

[pelson]

I think this is really neat, nice work!

Some observations:

• Could do with a couple more comments to explain the high level strategy at certain blocks of code
• Docs should clarify that, in reality, you would only want to do this if you cannot visualise the data directly (i.e. you have 2d lons and lats). For most other cases, cartopy should just handle it.
• I would love to see a visual test (might need to rebase to get those).

Apart from those, +1.

[rhattersley]

FYI. This looks to be a duplicate of @bblay's CoordSystem <-> CRS conversion.

[esc24]

Where is that code? I had a look in case someone had done it, but could only find a single method on the OSGB projection class.

[esc24]

@bblay and @pelson - the visual test work is great!

[bblay]

There's no explanation of what this keyword is for. Why do we need this when we can get it from the cube?

[bblay]

Just one outstanding issue for me: the shared array discussion, above

[pelson]

Looks great! +1

The only missing step is the ability for users to actually plot their newly created cube without having to manually get hold of the coordinates and data. @esc24 : do you have anything in the pipeline for this?

[esc24]

@pelson - That's is certainly the plan. I'll do it as soon as the pycarto++ branch is merged and things have settled down.

[pelson]

That's is certainly the plan. I'll do it as soon as the pycarto++ branch is merged and things have settled down.

Good idea. My opinion is that we should be targeting this functionality for 1.0. Given the time-scales, do you think that is achievable?

[pelson]

Needs an entry in the CHANGELOG and then I will merge with pleasure.