Bug: assign() generates KeyError when using read_pd() samples based on min_dist

[sgsutcliffe]

Problem

In the read_pd() function of the dist_reader class there is an option to drop samples based on a minimum distance (i.e. the lowest threshold set for clusters).

           if self.filter:
                if self.min_dist is not None:
                    if d < self.min_dist:
                        continue
                if self.max_dist is not None:
                     if d > self.max_dist:
                        continue

This creates an issue in the assign class when the id is dropped from dist

                            for id in addr_members:
                                addr_dists.append(dists[qid][id])

Which creates a KeyError

Proposed Solution

Not the most efficient, but easy to implement, is to add a conditional statement prior to appending the address:

                            for id in addr_members:
                                if id in dists[qid]: #<--- add conditional statement
                                    addr_dists.append(dists[qid][id])