However, sometimes a user may wish to pass in data that is already normalised between 0-1, and therefore set normalise to False. However, this still raises the warning, which can produce a lot of spammy messages.
Hmm, this is complicated by the fact that ds can be Dask, and we don't want to trigger computation early to work out if its values are outside the 0-1.0 range...
Describe the bug/issue The
calculate_indices
function currently provides a warning ifnormalise
is set to False, but an index is requested that needs normalisation. This prevents users from running an index that assumes values in the range of 0-1 on data with values in the range of 0-10,000. https://github.com/GeoscienceAustralia/dea-notebooks/blob/develop/Tools/dea_tools/bandindices.py#L316-L324However, sometimes a user may wish to pass in data that is already normalised between 0-1, and therefore set
normalise
to False. However, this still raises the warning, which can produce a lot of spammy messages.Solution This line should include a check to see if values are not in the range 0-1, and only raise the warning if that is the case. https://github.com/GeoscienceAustralia/dea-notebooks/blob/develop/Tools/dea_tools/bandindices.py#L316-L324