1.4.0 changes and what's new by bblay · Pull Request #540 · SciTools/iris

bblay · 2013-06-05T10:26:31Z

No description provided.

bblay · 2013-06-05T10:28:54Z

Note: I removed the features and bugs headings because I couldn't determine which bugs to include. That made it look like we hadn't fixed any bugs! Hopefully, with imminent, stricter PR guidelines on updating these files as we go, that should not be a problem with future releases.

esc24 · 2013-06-05T10:41:21Z

docs/iris/src/whatsnew/1.4.rst

Why have you removed these entries? They are the result of the few PRs that remembered to add entries.

It's still in the changelog, but because it relates to the tests, i.e it's a developer issue, it doesn't belong in the what's new list.

I disagree. It is a change to the configuration that directly affect some of our users. The changelog is not part of our built documentation so you can't expect that to be a sufficient route of communication. Let me be clear - changes such as these need to be easy to find. Sysadmins and developers who have custom site.cfg files need to be informed of this change.

ah, it is there, further down: https://github.com/SciTools/iris/pull/540/files#L1R33
needs to have the new alternative stated too though. will do that now.

esc24 · 2013-06-05T10:45:57Z

I quite like the new format of CHANGES, but the short one-liners need some minor tidying. However, the What's New entries are far too brief in my opinion. Bear in mind the people reading this may not be familiar with the context and it's also our opportunity to highlight our achievements over the last few months.

Did we really not deprecate anything?

ajdawson · 2013-06-05T10:52:03Z

I agree with @esc24 the What's New page should be more detailed. Check out matplotlib's What's New page: http://matplotlib.org/users/whats_new.html, which I think is a nice example. Not every feature needs a huge amount of detail, but they should at least be clear what the feature is for unfamiliar users.

bblay · 2013-06-05T10:54:58Z

More details what's new

Ok, will do.

Did we really not deprecate anything?

That's not easy to answer just by trawling through the commit history.
I'll have another look through and see if I can see anything.

ajdawson · 2013-06-05T11:00:59Z

I know for sure that the add_custom_season_* functions from coord_categorisation.py have been deprecated in favour of adding their functionality to the add_season_* functions.

bjlittle · 2013-06-05T12:20:47Z

@bblay it might be worth while searching the code base on deprecated and you'll find a hit in coord_categorisation.py as @ajdawson said. We should be using the .. deprecated:: X.X Sphinx directive to document deprecated functionality.

Can't help feeling that we should be more strict with ourselves when we add new functionality so that the appropriate whats new files are updated as part of the PR review exit citerion. It might make life easier when it comes to a release. Thoughts ...

cpelley · 2013-06-05T12:31:44Z

docs/iris/src/whatsnew/1.4.rst

this requires a more in-depth description

See the PR description #475 ... that might help clarify.

bblay · 2013-06-05T13:59:06Z

...we should be more strict with ourselves when we add new functionality so that the appropriate whats new files are updated as part of the PR...

@bjlittle, please could I encourage you to pop a comment somewhere on here

bblay · 2013-06-05T14:04:17Z

Please could people take a quick look at the what's new page, to check it is a fair representation of the work you wrote/reviewed. Many thanks, @munslowa, @esc24, @bjlittle, @pp-mo, @ajdawson, @cpelley.

esc24 · 2013-06-05T14:09:24Z

docs/iris/src/whatsnew/1.4.rst

I'd focus on the positive here. This is a big step forward in functionality. I'd also be tempted to combine the two experimental regridding features under a single section. I'll post some suggested text.

Experimental regridding enhancements

Bilinear and area weighted regridding functions are now available in :mod:iris.experimental.regrid. The final API is still in development, but in the meantime :func:regrid_bilinear_rectilinear_src_and_grid can be used to regrid a cube onto a horizontal grid defined in a different coordinate system. The data values are calculated using bilinear interpolation. :func:regrid_area_weighted_rectilinear_src_and_grid can be used to regrid a cube such that the data values of the resulting cube are calculated using the area weighted mean. Both of these function support masked data and handle derived coordinates such as hybrid height.

ajdawson · 2013-06-05T14:28:47Z

Deprecations are missing from the changelog.

pp-mo · 2013-06-05T14:51:30Z

docs/iris/src/whatsnew/1.4.rst

"expoerted" ?

um, it means taking all the poetry out! :D

esc24 · 2013-06-05T14:58:52Z

CHANGES has a release date, but what's new doesn't.

cpelley · 2013-06-05T15:00:10Z

docs/iris/src/whatsnew/1.4.rst

I do not believe this description covers all changes made

@cpelley, could you elaborate please?

pp-mo · 2013-06-05T15:08:48Z

Checking stuff I was involved with, items seem missing ?

(*) No mention of Grib translation extensions - #482.
Levels change is mentioned, but not ...

more grib2 params recognised on input
some grib2 params may now read with different standard_name
now translates some codes on grib2 output

(*) No mention of ESMPy conservative regridding - #453
Suggest something like

experimental support for area-conservative regridding between geographical coordinate systems.
This uses the ESMF library functions, via the ESMPy interface

pp-mo · 2013-06-05T15:19:54Z

I also think the regridding extensions should be grouped together.
There are 3, I think --

I reckon the comparison would make it clearer what each one does.

bblay · 2013-06-05T16:16:45Z

Many thanks for taking a look, peeps. 👀
All addressed except @cpelley's comment that the cube merge description is incomplete, awaiting further information.

cpelley · 2013-06-06T08:24:08Z

#475

This PR also resolves the situation when a NotYetImplementedError exception was raised when the situation of No functional relationship between separable and inseparable candidate dimensions was determined. Instead, the candidate coordinates participating within this group are all enumerated along an anonymous dimension.

Since users might be familiar with the exception that was raised rather than the reasons behind it, it might be worth mentioning the scenario for this. Just a thought really, I wouldnt hold anything back for it!

bblay · 2013-06-06T08:51:16Z

it might be worth mentioning the scenario for this

If someone can offer a suggested text I'll add it. That quote is gobbledygook 😃
I'll also have a little look at the code myself but I suggest we leave this one.

esc24 · 2013-06-06T09:00:16Z

I suggest:

"Cube merging now favours numerical coordinates over string coordinates
when identifying a coordinate to describe a dimension. Furthermore, it now favours :class:~iris.coords.DimCoord over :class:iris.coords.AuxCoord. These modifications reduce the likelihood of the error: No functional relationship between separable and inseparable candidate dimensions was determined."

cpelley · 2013-06-06T09:16:34Z

👍

bblay · 2013-06-06T11:15:45Z

(Perhaps we should merge this and discuss the changelog link elsewhere)

esc24 · 2013-06-06T12:33:41Z

I can't find the error it refers to

See https://github.com/SciTools/iris/blob/v1.3.0/lib/iris/_merge.py#L687

I got the wording slightly wrong. Please copy it from the above link.

esc24 · 2013-06-06T12:45:09Z

~~Please make the release dates consistent in changes and what's new~~. I suggest tomorrow i.e. 07 June 2013

bblay · 2013-06-06T12:46:20Z

See https://github.com/SciTools/iris/blob/v1.3.0/lib/iris/_merge.py#L687
I got the wording slightly wrong. Please copy it from the above link.

That's not in the trunk:
~~https://github.com/SciTools/iris/blob/v1.3.0/lib/iris/_merge.py#L687~~
https://github.com/SciTools/iris/blob/master/lib/iris/_merge.py#L687

esc24 · 2013-06-06T12:47:33Z

That's because the fix fixed it.

esc24 · 2013-06-06T12:49:46Z

docs/iris/src/whatsnew/1.4.rst

calculaus -> calculus

bblay · 2013-06-06T12:56:59Z

All review actions complete, except the cube merge text because I still can't find the error it refers to.

esc24 · 2013-06-06T12:58:06Z

docs/iris/src/whatsnew/1.4.rst

Move this "Both ..." up to the top paragraph.

bblay · 2013-06-06T12:59:50Z

That's because the fix fixed it.

Hehehe, ok, then we can't really say, "These modifications reduce the likelihood of the error: No functional relationship between separable and inseparable candidate dimensions was determined."

May we drop this one?

esc24 · 2013-06-06T12:59:53Z

docs/iris/src/whatsnew/1.4.rst

This returns a cube list. Either use cubes = ... or use load_cube()

Nice spot, thanks.

esc24 · 2013-06-06T13:03:56Z

May we drop this one?

OK by me but @cpelley wanted something so I suggest adding "These modifications prevent the error: No functional relationship between separable and inseparable candidate dimensions."

esc24 · 2013-06-06T13:06:09Z

@bblay - I'm done. Please see comments above.

bblay · 2013-06-06T13:23:51Z

Thanks, I think everything's been addressed.

cpelley · 2013-06-06T13:41:13Z

I'm happy to let it be, I only though that users who were familiar with this message for their case would identify the changes as being relevant to them, but its not important enough to hold this PR for.

esc24 · 2013-06-06T13:41:31Z

CHANGES

Propogate -> Propagate

bblay · 2013-06-06T13:45:10Z

@cpelley, it made it in: bblay@fe9b83a#L1R144

bblay · 2013-06-06T13:46:41Z

(I think you were both right to push for it)

esc24 · 2013-06-06T13:47:04Z

CHANGES

Change to:
"Improved error message within cube.aggregated_by()"

or similar

bblay · 2013-06-06T14:14:52Z

Thanks for the high quality review work, looking really good now.

cpelley · 2013-06-06T14:24:07Z

👍

1.4.0 changes and what's new

esc24 reviewed Jun 5, 2013
View reviewed changes

cpelley reviewed Jun 5, 2013
View reviewed changes

esc24 reviewed Jun 5, 2013
View reviewed changes

pp-mo reviewed Jun 5, 2013
View reviewed changes

cpelley reviewed Jun 5, 2013
View reviewed changes

ghost assigned bblay Jun 5, 2013

esc24 reviewed Jun 6, 2013
View reviewed changes

docs/iris/src/whatsnew/1.4.rst Outdated

Copy link

Member

esc24 Jun 6, 2013

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

calculaus -> calculus

esc24 reviewed Jun 6, 2013
View reviewed changes

docs/iris/src/whatsnew/1.4.rst Outdated

Copy link

Member

esc24 Jun 6, 2013

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Move this "Both ..." up to the top paragraph.

esc24 reviewed Jun 6, 2013
View reviewed changes

CHANGES Outdated

Copy link

Member

esc24 Jun 6, 2013

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Propogate -> Propagate

esc24 reviewed Jun 6, 2013
View reviewed changes

CHANGES Outdated

Copy link

Member

esc24 Jun 6, 2013

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Change to:
"Improved error message within cube.aggregated_by()"

or similar

1.4 changelog and what's new

efe2b81

esc24 added a commit that referenced this pull request Jun 6, 2013

Merge pull request #540 from bblay/suuup

fa45e64

1.4.0 changes and what's new

esc24 merged commit fa45e64 into SciTools:v1.4.x Jun 6, 2013

Conversation

bblay commented Jun 5, 2013

Uh oh!

bblay commented Jun 5, 2013

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

esc24 commented Jun 5, 2013

Uh oh!

ajdawson commented Jun 5, 2013

Uh oh!

bblay commented Jun 5, 2013

Uh oh!

ajdawson commented Jun 5, 2013

Uh oh!

bjlittle commented Jun 5, 2013

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bblay commented Jun 5, 2013

Uh oh!

bblay commented Jun 5, 2013

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Experimental regridding enhancements

Uh oh!

ajdawson commented Jun 5, 2013

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

esc24 commented Jun 5, 2013

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pp-mo commented Jun 5, 2013

Uh oh!

pp-mo commented Jun 5, 2013

Uh oh!

bblay commented Jun 5, 2013

Uh oh!

cpelley commented Jun 6, 2013

Uh oh!

bblay commented Jun 6, 2013

Uh oh!

esc24 commented Jun 6, 2013

Uh oh!

cpelley commented Jun 6, 2013

Uh oh!

bblay commented Jun 6, 2013

Uh oh!

esc24 commented Jun 6, 2013

Uh oh!

esc24 commented Jun 6, 2013

Uh oh!

bblay commented Jun 6, 2013

Uh oh!

esc24 commented Jun 6, 2013

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bblay commented Jun 6, 2013

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bblay commented Jun 6, 2013

Uh oh!