[
https://issues.apache.org/jira/browse/CLIMATE-914?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16005673#comment-16005673
]
ASF GitHub Bot commented on CLIMATE-914:
----------------------------------------
Github user agoodm commented on the issue:
https://github.com/apache/climate/pull/441
@huikyole Looks like your changes don't work for 2D latlon arrays, as per
CI build logs:
```
======================================================================
ERROR: test_two_dimensional_lats_lons
(ocw.tests.test_dataset_processor.TestSpatialRegrid)
----------------------------------------------------------------------
Traceback (most recent call last):
File
"/home/travis/build/apache/climate/ocw/tests/test_dataset_processor.py", line
399, in test_two_dimensional_lats_lons
self.input_dataset, self.new_lats, self.new_lons)
File "/home/travis/build/apache/climate/ocw/dataset_processor.py", line
278, in spatial_regrid
new_values_temp = griddata((lons[mask_index], lats[mask_index]),
values_original[mask_index],
IndexError: index 2 is out of bounds for axis 0 with size 2
```
Given that regridding is probably one of the largest bottlenecks in our
RCMES workflows, I think it is also worth profiling both this version (once the
above error is fixed) and the old map_coordinates version to see if there are
any performance impacts, since the spatial_regridding tests seem to be running
more slowly.
> Update dataset_processor.spatial_regrid module
> ----------------------------------------------
>
> Key: CLIMATE-914
> URL: https://issues.apache.org/jira/browse/CLIMATE-914
> Project: Apache Open Climate Workbench
> Issue Type: Bug
> Components: regridding/data processing
> Affects Versions: 1.2.0
> Reporter: Huikyo Lee
> Assignee: Huikyo Lee
> Priority: Blocker
> Fix For: 1.3.0
>
>
> OCW's spatial regridding has used scipy.ndimage.map_coordinates. For some
> unknown reasons, the map_coordinates function outputs -9999. as a missing
> value without a mask. We will replace map_coordinates with
> scipy.interpolate.griddata that is widely used in the climate science
> communities.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
