Issue #965 npf from imod5 #1010
Conversation
JoerivanEngelen
left a comment
JoerivanEngelen
left a comment
There was a problem hiding this comment.
Looks good, I have some suggestions to improve things
imod/mf6/npf.py
Outdated
| def fill_missing_layers( | ||
| source: xr.DataArray, full: xr.DataArray, fillvalue: Union[float | int] | ||
| ): | ||
| result = zeros_like(full) | ||
| result.values[:, :, :] = fillvalue | ||
| for l in source.coords["layer"].values: | ||
| result.loc[l, :, :] = source.sel({"layer": l}) | ||
| return result |
There was a problem hiding this comment.
Consider moving this logic to this namespace: https://github.com/Deltares/imod-python/blob/imod5_converter_feature_branch/imod/mf6/utilities/imod5_converter.py
| if regridder_types is not None: | ||
| regridder_settings.update(regridder_types) | ||
|
|
||
| regrid_context = RegridderWeightsCache(data["k"], target_grid) |
There was a problem hiding this comment.
Just a note: I think we should be able to optionally provide regrid contexts when we implement complete conversion of models.
There was a problem hiding this comment.
we can just provide an optional parameter there which has a default value of an empty RegridderCache. Then it's up to the user.
There was a problem hiding this comment.
Yeah I think that is a good resolution. You could implement immediately so we don't forget. If you decide to do so, please don't forget to update the method for the discretization package as well.
review comment Co-authored-by: Joeri van Engelen <[email protected]>
nice! Co-authored-by: Joeri van Engelen <[email protected]>
…es/imod-python into issue_#965_npf_from_imod5 # Conflicts: # imod/mf6/npf.py
| source for the layers that are in the source. For the other layers, the | ||
| fillvalue is assigned. | ||
| """ | ||
| result = zeros_like(full) |
There was a problem hiding this comment.
I'm pretty confident xarray has tooling for this, which ideally also supports out-of-memory (unlike accessing values).
Maybe align?
import xarray as xr
da1 = xr.DataArray([1, 2, 4], coords={"x": [1, 2, 4]}, dims=["x"])
da2 = xr.DataArray([0, 0, 0, 0, 0], coords={"x": [1, 2, 3, 4, 5]}, dims=["x"])
aligned, _ = xr.align(da1, da2, join="right")There was a problem hiding this comment.
This was a previous comment of me as well. I had a code suggestion using reindex, but it appears to be ignored or not pushed yet?
UPDATE: Found the comment: #1010 (comment)
There was a problem hiding this comment.
@JoerivanEngelen Indeed, I overlooked it, Now it is implemented, with reindexing
Fixes #965
Description
Adds a function to import an npf package from an imod5 project file. Adds tests.
Checklist
Issue #nr, e.g.Issue #737