Ancestor span heatmap

[savitakartik]

Partial fix for issue #20

This is a naive attempt at making a heatmap of average ancestor spans in time and position bins. I'm using nested for-loops to iterate over x and y axis bins and filtering values in the edges df.

Perhaps a quicker way to track genomic coordinates for the node spans is to either (a) modify child_bounds to return all edge_left and edge_right for a child or (b) assume the node persists in all trees between child_right and child_left .

[savitakartik]

ancestor_span_heatmap.mp4

[savitakartik]

We've decided to assume contiguous ancestor nodes for now, as this doesn't happen very often, at least in inferred trees. It could be slightly miselading for simulated genomes, but we can state our assumption.

We can then calculate the plotting data by using an array for the heatmap bins, looping over the nodes, getting the bounds from the existing child_bounds method referenced in an above comment, and incrementing all bins that fall within these bounds and time by 1. Should be feasible with numpy and can Numba-fy if needed.