Get working with iris-grib-with-eccodes; Python 3 included.#2887
Get working with iris-grib-with-eccodes; Python 3 included.#2887pelson merged 3 commits intoSciTools:masterfrom
Conversation
pp-mo
changed the title
conda-requirements.txt
Outdated
|
|
||
| # Iris extensions (i.e. key tools that depend on Iris) | ||
| # iris_grib | ||
| iris_grib>=0.12 |
There was a problem hiding this comment.
So I thought this was commented out for a reason!? iris extensions shouldn't be installed by iris because they themselves depend on Iris.
| * see :func:`iris.fileformats.netcdf.save` | ||
| * GRIB2 - the WMO GRIdded Binary data format: | ||
| * see :func:`iris.fileformats.grib.save_grib2`. | ||
| * see :func:`iris_grib.save_grib2`. |
| 'SKIPPING test for this cube, as save/load will ' | ||
| 'not currently work.') | ||
| warnings.warn(msg.format(i, name_cube.name())) | ||
| continue |
There was a problem hiding this comment.
(I know it 's not your change but... ) We have more sophisticated means for skipping tests...
There was a problem hiding this comment.
But this isn't skipping a test, it's skipping one of the cubes tested within a single testcase.
As the set of cases is defined by content from a file, rather than the test code, I can't really see a better way ...
|
|
||
| if tests.GRIB_AVAILABLE: | ||
| import gribapi | ||
| from iris_grib._load_convert import _MDI as MDI |
There was a problem hiding this comment.
As a private API, we obviously we shouldn't be relying on this, but happy enough for now.
|
FWIW: Technically, we could not merge this in 2.0.0 and still have iris' GRIB capability using eccodes the moment we update iris-grib. I consider that to be a healthy sign of separation. As it stands, there are some good doc improvements in here that make sense to have in v2.0.0 |
pelson
force-pushed
the
post_iris2_grib_fixes
branch
from
4833ba3 to
94ae63b
Compare
2 participants
This won't work until a new iris-grib>=0.12 is available on conda-forge
I'm assuming that will have masked dask support AND eccodes
There are changes here specific to the eccodes transition, ported across from #2607