No error or warning is raised when required input files are missing

[luukvdmeer]

For example: When I compute netascore (Docker + OSM workflow) using a weight file that specifies a non-zero weight for gradient_bicycle, but do not provide a DEM, the computation succeeds without giving any error or warning. The same is true for specifying a noise weight without providing a noise file.

How are these situations handled internally? Does it mess up the normalization of the indices? Depending on that, I would at least give a warning but probably better an error.

[christian-werner]

@luukvdmeer Index calculation will not be messed up - the indicator will be ignored (as if weight = NULL). However, I totally agree that we should output a warning in such cases.

After discussion with @robin-wendel we propose the following solution:
In index step, for every indicator we check whether all indicator values are NULL and in this case issue a warning (loglevel=1 / major info). Furthermore, we plan to generally include a quality indicator. This indicator would also be reduced by the weight that is assigned to the missing indicator.
This solution would work for any missing indicator and should also catch cases of e.g. a DEM provided but spatially not intersecting the network.