Add "insert sites" method to TableCollection (and possibly TS) #675
Description
It might be useful to have an insert_sites method which adds new (empty) sites at user-specified positions and updates the mutations table to keep everything sane? I can imagine this being useful as a prerequisite step for e.g. combining 2 tree sequences (or maybe VCF files) that have accumulated mutations at unique sites, or adding in new sites that have just been reported in an updated sequencing effort, etc. Other existing functions such as map_mutations could then be used to add mutations at any specific (currently empty) site.
It would be useful to sort the sites and return the new ids of the inserted sites, in the same order that the positions were passed in. I imagine it something like this:
def insert_sites(self, positions, ancestral_states, metadata=None):
if metadata is None:
self.sites.append_columns(positions, *tskit.pack_bytes(ancestral_states))
else:
self.sites.append_columns(
positions,
*tskit.pack_bytes(ancestral_states),
*tskit.pack_bytes(metadata))
self.sort()
# Positions are required to be unique, so we can use these to identify the new ones
new_locations = np.searchsorted(self.sites.position[:], positions)
assert np.all(self.sites.position[:][new_locations] == positions)
return new_locations