The rate at which reads are created with sequencing error is lower than desirable. Approximately 10x slower than simple error-free reads.
Although some profiling has been done, I suspect that occasions where numpy.argwhere() or numpy.where() is called could be replaced with numpy.searchsorted(). This method if faster when the aim is to recall a single index, whereas the "where" methods return all indices meeting a criteria.
The rate at which reads are created with sequencing error is lower than desirable. Approximately 10x slower than simple error-free reads.
Although some profiling has been done, I suspect that occasions where numpy.argwhere() or numpy.where() is called could be replaced with numpy.searchsorted(). This method if faster when the aim is to recall a single index, whereas the "where" methods return all indices meeting a criteria.