Issue #344 part1 imod refactor primod update #348
Conversation
| @@ -29,6 +28,11 @@ scipy = "*" | |||
| tomli = "*" | |||
| tomli-w = "*" | |||
| xmipy = "*" | |||
| filelock = "*" | |||
There was a problem hiding this comment.
Could you sort the dependencies?
There was a problem hiding this comment.
I defer from that now, as this is a temporary measure. After iMOD Python has been released with the API update, I will revert these changes to the pixi.toml. Instead, I will then pin iMOD Python to be above or higher a specific version number.
| @@ -71,7 +73,7 @@ def _create_well_id( | |||
| layer = np.tile(well_layer, (n_subunit, 1)) | |||
| layer_1d = layer[well_active] | |||
|
|
|||
| well_id = self.well.dataset.coords["index"] + 1 | |||
| well_id = well_cellid.coords["ncellid"] + 1 | |||
There was a problem hiding this comment.
ncellid is actually an zero based index array right? otherwise the +1 is now unnecessary.
There was a problem hiding this comment.
That is correct!
| @@ -17,7 +17,7 @@ def make_msw_model( | |||
|
|
|||
| idomain = gwf["GWF_1"]["dis"]["idomain"] | |||
| if "layer" in idomain.dims: | |||
| idomain_layer1 = idomain.sel(layer=1) | |||
| idomain_layer1 = idomain.sel(layer=1, drop=True) | |||
There was a problem hiding this comment.
I ran into an issue with some tests, where iMOD Python was complaining about faulty coordinates. It turned out as the layer coord was passed on to imod.util.spatial.get_cell_area, which throws an error if there is any coord more than x, y, dx, and dy. This could be easily fixed by dropping the layer coordinate upon selecting the first layer in the fixture. It did not seem like the layer coord was necessary in the first place, so I think dropping it is a slight improvement regardless.
…using the dtype of like, which is an integer. This causes problem for the fill value "nan" which the ribamod coupling depends on.
JoerivanEngelen
merged commit bbc941e
into
issue_#342_imod_refactor_primod_update
JoerivanEngelen
deleted the
issue_#344_part1_imod_refactor_primod_update
branch
UPDATE: This is a new attempt for a PR, in the other one I accidentily merged locally and pushed to feature branch. I reset that now.
Closes #344, this is the first task to update
primodfor the iMOD Python refactor.Plans are layed out in #342. In this first task, I changed the use of the deprecated
WellDisStructuredobject, to the Mf6Wel object, which is a direct translation of MODFLOW 6's cellids to a file. As this iMOD Python refactor is still work in progress, I temporarily introduce a install-imod task, which pip installs imod from a git branch, instead of a release on conda-forge. After I have accomplished part 2 of the refactors, I'll make an iMOD Python release, and this can be reverted back.It does the following things:
Mf6Welobject instead of theWellDisStructuredobjectdtypein calls to imod.prepare.rasterize to float, to work around this issue Primod's basin mapping can't deal with integer basin id's with latest Rasterio dependecy #333. Thedtypeargument is only included in the the dev version of iMOD Python yet.