Grib1 loading defines a GRIB_PARAM attribute, like grib2 load.

[pp-mo]

Closes #330 #402 #398

TODO:

• needs at least some doc

Noting that current docs is virtually a single page,

Also, version there on main already needs bumping from 0.18 to 0.19
-- or, it is missing a mergeback from latest release (v0.19.x)
(update : fixed that)

[pp-mo]

Note: while fixing + documenting this, I have reworked the single-page docs quite extensively.
Hence now addresses all-of #330 #402 #398
I have also fixed it so all doctests pass -- but I believe this is not enabled at present (#274 ?)

[trexfeathers]

Note to self, I have previously reviewed up to this point:

368596a Line-length fix.
0c08268 Further docs rework : doctests now working.
089b5b9 Various documentation improvements.
5cff713 fix mock usage.
e60697c Update cube-load result snapshots.
e02158d Shorten line.
30dc4d3 Grib1 loading defines a GRIB_PARAM attribute, like grib2 load.

[pp-mo]

STATUS UPDATE 2024-04-16 13:25
pushed a bunch of further changes, to run all checks + chase any additional problems .

Intending when this is fixed to rebase / recast, as this seems to have acquired some merging problems.

[pp-mo]

NOTE:
Now re-picked and rebased onto current main.
I think the total resulting change is now easier to review than it was.

The doctest problem remains : it is failing to import iris.tests.stock.
Not clear what triggers this, maybe somehow we are running the doctests differently from Iris ?
May need fixing in Iris itself. Ignore for now ??

[trexfeathers]

Exhaustive is probably the right word!

Here are a few comments. Cheers!

[trexfeathers]

The doctest problem remains : it is failing to import iris.tests.stock.
Not clear what triggers this, maybe somehow we are running the doctests differently from Iris ?
May need fixing in Iris itself. Ignore for now ??

Even in Iris we have never tried running a doctest that includes content from iris.tests, as far as I can see. I've always strictly stuck to content from iris-sample-data, and that is available to us now thanks to #413. Please could you adjust your doctest to use something from there instead?

[pp-mo]

Even in Iris we have never tried running a doctest that includes content from iris.tests, as far as I can see. I've always strictly stuck to content from iris-sample-data, and that is available to us now thanks to #413. Please could you adjust your doctest to use something from there instead?

Good thinking. 💯
I'll take a look at that.

[trexfeathers]

I have now reviewed up to: Fix doctests: don't use iris.tests

[pp-mo]

I have now reviewed up to: Fix doctests: don't use iris.tests

That's good!
I think I may have some nasty merge/rebasing to sort out now, but if you've reviewed up to the latest that I can force push again if I need to.
I also have the unadressed points above to cover, but I think that all that can be done afterwards without the references to existing commits. I'll sort out those changes, bank them somewhere + re-apply when I've fixed the conflicts.

[pp-mo]

@trexfeathers Thanks so much for sticking with it !
🎖️