Code finds *additional* barcodes, filters existing about emptydrops

[nh3]

Thanks for your feedback. The intention of this package is to give easy access to those code to replicate CellRanger's cell calling behaviour. So whether the behaviour matches exactly that of the R implementation was not given priority. In this particular case, find_nonambient_barcodes() finds additional non-ambient barcodes, which is the behaviour of this function in CellRanger. The reason I didn't further wrap the barcode merging part is that it's trivial for the user to do so, and it allows users to inspect the returned p values to choose an appropriate cutoff which provides more flexibility than just returning a set of barcodes.

[CowanCS1]

Fair enough, and I can understand why CellRanger would not want to evaluate and possibly exclude cell barcodes included with the previous method (even though this could perform poorly in some edge cases).

It would be very helpful if you add the description above to the readme, as the current description relies on knowing the details of CellRanger's emptydrops implementation. If the user doesn't have this knowledge a priori, they have to dig through the CellRanger code to understand that this is the expected behavior.

[nh3]

README updated in 45330a9
Thank you for your feedback and suggestion!