Prepare 0.34.0 release

[veenstrajelmer]

Todo:

• ** WARNING: No humidity, airtemperature, cloudiness and solar radiation forcing found, setting temperature model physics:Temperature = 1 (Only transport)>> set Temperature to 1 per default? Consider adding an additional flag to enable the heatflux model (and download+merge additional ERA5 vars)
• update installation guideline with different miniforge3 link: https://conda-forge.org/miniforge/
• fix hycom download #1048
• update dev docs installation to py312
• interpolate_tide_to_bc paths for linux #1080
• pdrive windows should include slash #1084
• Support for noon and monthly averaged data in dfmt.cmems_nc_to_ini() #1086
• Avoid pytest warnings #1065
• Consider 12-hour offset for CMEMS data #878
• remove copernicusmarine temporary isoformat conversion #1047
• Writing the illegalcells file in the model folder #1044
• test modelbuilder on EDITO datalab (has clean downloads) and locally
• codecov is not updating
• re-enable era5 tests on github #1005
• notebook: illegalcells file define separately, include space in os.path.join
• clean up code and examples #1093
• move preprocess_hydrolib_readtim.py and preprocess_hydrolib_readxyz_readxyn_readcrs.py to notebook
• cleanup comments in postprocess_mapfile_to_regulargrid.py
• compare test coverage of pytest -m "requireslocaldata" with github to see which tests to focus on for moving to github
• update xugrid (see below)
• update hydrolib-core (see below)
• easy to add fou/rst/minimalnetwork demo data? Then also increase test coverage
• CMEMS phyc not available in reanalysis #622
• repair docs workflow #1092
• post-release: notify all users who downloaded cmems data in the last years, since midnight and noon datasets are not compatible. Good to redownload anyway since the spatial extents might also have slightly changed.

After xugrid release:

• add xugrid from_structured after From structured dataset xugrid#303 (merged on 21 Oct 2024)
• remove filterwarning after resolve warnings in ci workflow xugrid#312 (merged on 30-1-2025)
• uds.close() now properly works, also for partitioned datasets?

After hydrolib-core release:

• set new minimal hydrolib-core version in the dfm_tools pyproject.toml
• remove stretchtype keyword from modelbuilder notebook (and check if EDITO run functions normally)
• Run Python 3.13 testbank #559
• Check support for numpy 2.0.0 #877
• Non-orthogonal grid and illegal cells function #909 (since higher meshkernel version can now be used)
• maybe more from: Fixes related to hydrolib-core issues #214
• Check support for numpy 2.0.0 #877

Optional:

• Unexpectedly high orthogonality for relatively high latitudes #1078 >> high orthogonality with refinement for higher latitudes MeshKernelPy#199 >> because for high latitudes the dx/dy ratio should not be 1. Fix in dfm_tools or request feature in meshkernel?
• create_model_exec_files flexibility #1081
• copernicusmarine speed up retrieval of time extents #1082
• Flexible varkeys in dfmt.preprocess_merge_meteofiles_era5() #887
• Extrapolation of CMEMS initial field #1021
• snap observation points to nearest cell #898
• prevent calls to uds.grid in dfmt #363
• run pytest -m "requireslocaldata" every now and then
• run example scripts every now and then (acceptance tests)
• create issue template
• document all required repository secrets and what for