Conversation
|
addresses #604 |
lib/iris/fileformats/rules.py
Outdated
There was a problem hiding this comment.
Sorry, but what's this for?
There was a problem hiding this comment.
this unique merge was running during orography processing and for a number of files it made the whole load process fail. The Fields file contained two orography fields which appeared the same to merge.
This seemed an odd error to get so I tweaked the merge conditions so the file would load.
This code change is not central to this pull request, but it does highlight that the processing of orography can trip up the whole load process for some files.
ghost
assigned
|
As mentioned in SciTools/iris-test-data#14, this could do with tests for the three sub-grids. |
|
arakawa c grid staggering for the test data was an excellent idea this and https://groups.google.com/forum/#!topic/scitools-iris/5nsb8O_dSho highlighted that the logic for arakawa staggering is flawed this will need a code update to fix |
|
arakawa staggering is now handled pull request on code-generators: |
lib/iris/fileformats/ff.py
Outdated
There was a problem hiding this comment.
I'd rather not have the extra indirection of using a dict - we should just store the grid directly in the stash_refs mapping:
stash_refs = {'m01s00i001': 1, ...}
If we need to extend at a later date, then we can do so at a later date ... that's what later dates are for. 😉
There was a problem hiding this comment.
no problem, I have updated here and in iris-code-generators.
I have changed the name of the dict to stash_grid to improve clarity
|
hopefully this addresses the review feedback test now pass apart from data file missing |
|
https://groups.google.com/forum/#!topic/scitools-iris/5nsb8O_dSho ''' Martin Dix 2:28 AM (5 hours ago) |
|
I'm going to take this under my wing as I think this will be useful to a number of people. Relevant PRs: |
|
P.S. - this works properly for my tests too... |
lib/iris/tests/test_ff.py
Outdated
6 participants
The Fields file loader is enhanced to capture data stored and referenced by pointers.
The row and column dependent constants, if present, are added to the PPHeader instance as x and y vectors
rules are updated to recognise missing data values for bdx and bdy