contig-level statistics, more thoughts

[ctb]

I wanted to write down some more thoughts on evaluation and contig-level stats.

• it would be interesting to know which contigs currently filtered out by Reason 1 (gather matches outside of lineage) would also be filtered out by Reason 3 (LCA matches outside of lineage), and if it changes for different data sets.
• ref https://github.com/dib-lab/charcoal/issues/33#issuecomment-636875765 we could also look at the overlap of the LCA Reason 3 with contigs identified based on the whole-genome-prioritization approach.

these will probably need to be implemented as a separate pipeline, alas.