Fix sampling of profiles & add mfd workflow

[kejohnso]

This PR makes some small changes to the subduction tools:

• adjusts profile sampling so that imprecision doesn't prevent the last edge (i.e. the one at the lower depth) from being written
• adds the possibility of kite fault or complex fault for surfaces used as reference in classification

And adds a workflow that chooses three MFDs representative of a large set of possible ones in the following steps:

• creates N catalogues by sampling within the sigmaMagnitude
• declusters these according to the configs and rules by TRT (initial catalogue has been classified beforehand)
• solves for completeness and MFD simultaneously for each declustered catalogue
• evaluates the covariance of the best a- and b- values for all instances, returning the center and limit values (user defined sigma)
• creates a bunch of plots to demonstrate this

Most of the steps in the workflow come from other functions, so to start there are only two tests for the workflow:

1. checks catalogue generator
2. tests end to end the workflow, checking the final MFD values