Bug in count-duplicates feature

[edurand]

Running fastqp input.bam -o output -d raises UnboundLocalError: local variable 'ScalableBloomFilter' referenced before assignment

The bug can be traced to cli.py lines 132-133 and 138-139. The order of two blocks need to be reversed (or better, the two blocks need to be merged since they are testing for the same thing), so that ScalableBloomFilter is imported before it's used.

[Matt Shirley]

Thanks, @edurand. This is kind of a silly error and I guess I should point out that the duplicate calculations are probably wildly underestimated if you're using the default sampling of 2M reads. I guess if you're evaluating all the reads in a file then this is a useful feature, so maybe I'll add a warning or mutually exclusive arguments that prevent misinterpretation of this value.