HRA outputs always high-risk when E/C pairs are entirely 0 for a habitat

[phargogh]

A user on the forums (link to private convo with @phargogh and @davemfish ) has an interesting setup for their run where all of the criteria for all stressors affecting a single habitat are being ignored by setting those table values to 0. This is resulting in some invalid values (numpy WARNINGs from divide-by-zero), which I suspect is then throwing off the results.

[phargogh]

Well this is a fun case! From the user's exposure/consequence table:

No wonder there's a division by zero ... I didn't think the user would request that all pairs be skipped.