dependency issues

[johntruckenbrodt]

Hi @Achaad,
currently I cannot install all packages together because this package needs very new versions of all dependencies. In particular, my packages do not support numpy>=2 yet. Is it possible for you to drop it? In general, are you strictly depending on these recent versions or did you just pin the versions in your environment?

Further questions:

• is loguru compatible with the standard logging library? My packages use the latter and we should ensure that everything can be logged.
• what are pandas-stubs and pytest-stub for?
• why are you installing rasterio and rioxarray using pip in the environment.yaml file?

[johntruckenbrodt]

As discussed here the dependencies of relevant packages:
https://github.com/johntruckenbrodt/spatialist/blob/main/environment.yml
https://github.com/johntruckenbrodt/pyroSAR/blob/main/environment.yml
https://github.com/SAR-ARD/s1ard/blob/main/environment.yaml

[Achaad]

Hi, @johntruckenbrodt!

• There should not be a problem with loguru and logging libraries interaction. The envisat_sarsen uses logging but imports loguru with asar-xarray, and the log output seems to be complete. However, the logs are not homogenous and it can be seen which logs are generated by which library.

• The project enforces strict type rules using mypy and requires the imported packages to provide type hints as well. The -stubs packages provide the type hints which are somewhy missing from the original packages.

• At the time of testing, there were some issues with importing correct version of rasterio and rioxarray from conda, so they were added as pip dependencies. I will verify whether that is still the case.

• I have used the version of the libraries provided by you and there were no issues with running the asar-xarray and sarsen. I will update the environment.yaml file and merge the changes the next week.

[johntruckenbrodt]

Thanks @Achaad for the explanations.

• it would be nice to have homogeneous logs but for now this is low priority
• please add some note like "for type hints" behind the stubs packages in the dependency files
• same for rasterio and rioxarray, add some note like "could not get version X from conda"
• great, looking forward to the new environment.yaml file

[Achaad]

@johntruckenbrodt
I have finally merged the changes, and the new version has been released. Feel free to reopen the issue if there are still some dependency issues.

The issue is closed with #55

[johntruckenbrodt]

@Achaad the higher numpy version is still pinned in the pyproject.toml.