Extracting endmembers (sampling points from raster) expected behaviour

[KBodolai]

ping openjournals/joss-reviews#6692

Description

I was stress-testing the examples thinking of common mistakes an inexperienced user may make. In the first machine learning example, if you have some point endmembers with a different CRS, it returns, without any warning or complaint, a dataframe where all the values are -9999.

This is not an issue of the example per-se, but I'm thinking it could be useful for scikeo.process.extract to warn the user an attempt to reproject (or just warn them that when the CRS is not the same the results will likely be wrong).

What do you think @yotarazona ?

[yotarazona]

Hi @KBodolai, thanks again for this feedback. It was fixed in the new version of the package. See this issue please #20

[KBodolai]

fantastic, thanks!