sourmash
sourmash copied to clipboard
document containment direction for `compare` (and standardize for containment ani)
When running sourmash compare
, the matrix will be asymmetric. But which direction is which? And is it the same for containment
and containment-ani
?
Based on code here: https://github.com/sourmash-bio/sourmash/blob/latest/src/sourmash/compare.py#L75-L90
For --containment
, we do containments[i][j] = siglist[j].contained_by(siglist[i])
while for --containment-ani
, we do siglist[j].containment_ani(siglist[i])
...which will give accurate values for each but populate the matrix in opposite directions (ref #2193) This directional change will likely be fixed by changes for #2193, but we should document the final shared direction.
Sticking with the --containment
direction, with i
=row, j
=col, the matrix will be the column genome contained by the row genome.
see also #1835 😁