* Change from tqdm import tqdm

to from tqdm.auto import tqdm
And update notebook 2

* Fix nbconvert version

* Fix codacy issues

And revert tqdm import

* Add ipywidgets as dependency for the docs-build

* Add docstring and tests to some ahn-methods

Landflag was calculated wrongly. A value of 0 was assigned to the first active cell even if it was active. This commit fixes this.

Two issues are solved:
- If values for a large area are prompted, the query url gets very long and is not accepted by HHNK polderpeil server. Is now queried in chunks
- An error in the database of HHNK resulted in duplicate peil definitions. Only use the unambigious entries and leave a note in the logs.

* HHNK revert to old url

* Fix failing test (unrelated to other change in this PR)

---------

Co-authored-by: Ruben Caljé <r.calje@artesia-water.nl>

* Add support for NHI GWO database

with extraction data

* Add timeout to solve codacy-issues

* docs typos

* Make nhi gwo methods more robust

Add checks for empty empty content
Determine empty lines in get_gwo_wells, to be used in skiprows

* Fix codacy issue

---------

Co-authored-by: Davíd Brakenhoff <d.brakenhoff@artesia-water.nl>

* ds_to_gridprops() to also pass unaltered datavars

ds_to_gridprops() is called by refine(). If the time settings were configured prior to calling the refine function, the refine function silently threw away the steady array, as it is non-numeric. Not anymore.

* Minor edit to ds_to_gridprops

…ng (#324)

* Zero layer thickness layers are considered active due to float rounding

Subtracting two botms to calculate the thickness results is some layers with a thickness of ~1e-9. This rounding later results in a problem because this positive thickness leads to an idomain of 1.

This can either be solved in `nlmod.layers.calculating_thickness()` or in `nlmod.layers.get_idomain()`. What do you prefer?

* Remove trailing whitespace

* Applied patch also to get_idomain

* Some changes to better implement grid rotation for structured grids

* Add a test for a rotated grid

* Add extra tests for grids

And fix wrong values for xc and yc for rotated vertex grids
Also fix gdf_to_count_da for rotated grids

* Test vertex_da_to_ds and fillnan_da

* Update tests to fix failing tests

* Improve docstrings of xorigin, yorigin and angrot

* Add get_flow_residuals and get_flow_lower_face

* Minor changes

Fix lays parameter for structured grids

* Fix failing tests

* Fix last failing test

* Improve  nlmod.gwf.output.get_budget_da()

So it also works for multiple packages of the same type (for example 3 DRN-packages), by summing the budgets

* Handle comments from @OnnoEbbens and replace grbfile by grb_file

Like in flopy

* Improve export to ugrid_nc-file, to support imod qgis plugin

* Improve docstring

Error in computing the add_offset

…th just an extent

…me name

* Update cache.py

The netcdf cache function validates the cache by comparing the ds argument and other function arguments to the pickled arguments. If they match, the cache can be used.

Currently, just the coordinates of the argument ds and the output ds had to match, introducing two errors:
- If the data_vars differ and are used the cache is falsely valid
- The coordintates of the ds argument has to match the coordinates of the output ds. This limits the use of the cache function.

The PR compares the hash of the coords and data_vars of the ds argument to those that were stored in the pickle together with the cached output ds.

Ideally, the cache.cache_netcdf() accepts arguments that specify specifically which data_vars and coords need to be included in the validation check. Beyond the scope of this pr.

- Included tests

Small change to cache

fix webservice HHNK for #334 and #317

remove confusing ci step

Otherwise nlmod.read.regis.get_regis() would return ModuleNotFoundError: No module named 'bottleneck'

And not just csv-files

* Creating new cache used full ds instead of reduced ds

Required to ensure that the cached function does not use data_vars that not explicitly required

* Format cache to please  codacy

* Better support for vertex grids

* remove option to store delr/delc as attrs

* remove option to store delr/delc as attrs

* update polygons_from_model_ds to to use new delr/delc datavars

* add delr and delc to datavars in ds_to_structured_grid

* remove delr/delc attr check in structured_da_to_ds

* get_recharge uses coords3d to determine first active layer

* add delr/delc to datavars for surface water functions

* formatting

* formatting

* remove coords that number from 0 to N (these are not listed in coords)

* remove delr and delc from datavars is vertexgrid

* do not check time attrs start and time_units

* do not check time coord attributes

* formatting

* skip delr/delc when writing files

* rename bathymetry tests

* comment still in dutch: bathymetry

* comment still in dutch: bathymetry

* use last delr to set extent

* use brackets (was failing for me without)

* Cache: Implemented inline suggestions made by David

---------

Co-authored-by: Davíd Brakenhoff <d.brakenhoff@artesia-water.nl>

* Cache: added shorthand for model ds attributes

* Update cache.py

* Please Codacy

* Update cache.py

* Treat transport flag as required model ds attribute

* Cache: Exclude "created_on" attr

* Added get_exe_path with more features

Use previously installed anywhere. If doesn't exist, download it.

* Minor change to docs

* test

* Ensure version also with nlmod install

* Update util.py

* Fixes suggested by Davíd

* Only kwargs allowed

* update function calls to get_exe+path()

* New syntax of get_exe_path compatible with prior syntax

* Github token needed for CI when downloading executables

* Envvar NLMOD_SUPPRESS_VERION_CHECK

* Renamed envvar

* Let's see if we can now get by without passing the GITHUB_TOKEN env

* Default to not check version

* Consume version_tag arg when creating model_ds

* Removing env SUPPRESS_EXE_VERSION_CHECK

* Removed dangling argument

* Added version_tag to refine/gridgen

* Also auto-download modpath_provisional for other platforms than win

* Remove delr and delc from ds, first commit

* import grid again in gwf

* Some more removal of delr and delc and simplify _get_delr_from_x

* Handle comments by @OnnoEbbens

Animate cross section

Links from https://www.dinoloket.nl/modelbestanden-aanvragen/netcdf

Fix regis URL

- allow reading heads da using ds that has no time discretization.

* Patch get_bin_directory()

* Removed trailing whitespaces

* Remove explicit download_mfbinaries in favor of auto download

* Write metadata in case metadata is not written

* Add warning

* Please codacy

* Please codacy with adding encoding to writing metadatafile

* No warning when running tests

as it is expected behavior

* run ruff check --fix, docformatter
add ruff settings to pyproject.toml

* update workflows
- up actions versions
- remove flake8 check

* restore the position of NLMOD_DATADIR

* skip failing KNMI Data Platform tests for now

* committed import to wrong branch

* update imports in examples

* more style thingies

* fix nb metadata

* restore import

* actually skip knmidp tests

* forgot parentheses

add modelextent plot

- convert ds time coordinate to pandas DateTimeIndex, optionally with the starting date.
- can be useful for resampling time series to model period

- fixes the warning
- allows flox to be used for faster groupby operations

- Improve docststring get_exe_path
- lower log-level get_flopy_bin_directories to DEBUG

- Improve docststring get_exe_path
- lower log-level get_flopy_bin_directories to DEBUG

- typos
- numpy 2.0
- show pkg versions at top

…ng to file)

Which would result the following error

ValueError: Dimensions {'band'} do not exist. Expected one or more of FrozenMappingWarningOnValuesAccess({'y': 600, 'x': 700})

- rendering of numbered/bulleted lists
- fix pyogrio warnings
- other minor things

- suppress pyogrio no CRS warning

Test skips were added in #358 because they failed.

Co-Authored-By: Martin Vonk <66305055+martinvonk@users.noreply.github.com>

- comments onno
- new link with info on EPSG 28992
- add trap for duplicate indices in gdf_to_grid (better error message for users)