Open pimentel opened 8 years ago
Hey, so if I see it correctly an exit status 1 has been implemented when 0 reads align? I understand that this will probably not be changed back but I would argue against raising an exit code 1 when 0 reads align but just printing a warning or using a different exit code.
The issue with exit 1 in this case is that there can be genuine use cases where 0 reads align such as:
When Kallisto then exits with an exit status 1 it becomes difficult to implement pipelines (bash or nextflow) around around these use cases as there is no way to differentiate between 0 mapping reads and e.g. an index issue. The only way to currently do this is to simply ignore exit status 1, which I seems dangerous as other, unexpected issues might give rise to the same exit status.
Currently, we do not check at the end of a run if no reads aligned. As a result, some people have been getting 0 reads mapping and 'nan' everywhere in their results. We should error, and give a helpful message such as: 'no reads mapped, you should probably check that your k is less than the read length by at least %.'