catalog format for time-only forecasts

[rinaldiap]

How should I format my catalog if I want to compare only the time evolution of my forecast?
Should I just use a fake location for "each" event?

[wsavran]

Is the forecast specified as a catalog? also, how will you define the region to filter the catalog used for the evaluation? in short, yes, you can just create dummy locations for lat/lon/depth in the forecast. by doing this, you will only be able to compute the number_test().

the other option is to create a mock class that would work with the number_test and magnitude_test() functions. this requires a bit more understanding of python. this is probably the best solution to be honest.

i haven't anticipated this use-case yet, so ill work through an example to show how this would work. it might require some code changes.

[rinaldiap]

the actual output is just number of events in time, so the first option would require a sort of wrapper/post-processing to create an ad-hoc catalog with dummy locations.
I agree that the best solution would be to have a mock class. I will think at something and maybe double check if it is then something to include officially in pyCSEP.

[wsavran]

id be in support of having the ability to accommodate a temporal only forecast. if you don't mind sharing an example output of the forecast in ascii format i could try and help you set up the class to make sure it works with everything else.

[wsavran]

hi @rinaldiap, just checking in to see if you had any luck working with time-only forecasts.

[rinaldiap]

not really. I had no much time, but I will provide an example soon