Conversation
mxkpp
left a comment
There was a problem hiding this comment.
I built using these versions and got the geomod n1 test to pass as is without modifying the current expected test data.
(with geo_test_pre_refactor checked out locally for ngen-forcing):
REPO_TAG_FCST_MGR="856fc0e1201076df909e56c7cd384f58e82965a2"
REPO_TAG_MSW_MGR="693c206a22b5e9ffcca3103166c0ca59e2b11b25"
REPO_TAG_CAL_MGR="7e56bf01477ea77e72dfb25a166ac26ff6090ecb"
REPO_TAG_NGEN_FORCING="LOCAL"
NGEN_SOURCE_MODE="ghcr"
NGEN_BASE__REMOTE_GHCR_TAG="844c5f6"
Please see initial requested changes.
|
Can these files be removed? |
mxkpp
left a comment
There was a problem hiding this comment.
All of the original regrid test data should remain, those should not be deleted. For example these n1 files and their corresponding n2s:
tests/test_data/expected_results/test_expect_regrid_aorc_aws__gauge_01123000_start20130701000000_n1_rank0_timestep0.json
tests/test_data/expected_results/test_expect_regrid_conus_hrrr__gauge_01123000_start20250710040000_n1_rank0_timestep0.json
tests/test_data/expected_results/test_expect_regrid_conus_rap__gauge_01123000_start20250710040000_n1_rank0_timestep0.json
This should be fixed now. |
mxkpp
left a comment
There was a problem hiding this comment.
The np.full calls are missing dtype in this branch. I think once those are re-added, this is ready to merge to mdeshotel_pre_dev.
mxkpp
left a comment
There was a problem hiding this comment.
Looks good. I re-created expected test results data to confirm that the committed test files don't change, and then I re-ran the tests.
This PR sets up tests for GeoMeta and InputForcings classes prior to refactoring. The intent is to establish pre-refactored results for comparison to post-refactored results.
Additions
Removals
Changes
Testing
Screenshots
Notes
Todos
Checklist
Testing checklist
Target Environment support