BrooksLabUCSC / flair

Full-Length Alternative Isoform analysis of RNA
Other
208 stars 71 forks source link

flair 123 is horribly hard to use #269

Open diekhans opened 1 year ago

diekhans commented 1 year ago

"flair 123" module magic can never work correctly given the way python argument parsing works. The arguments for "flair 23" are not the same as "flair 123". This is horribly confusing and the behavior is undocumented and I will buy anyone a drink if they can accurately and clearly document it.

It is simply not worth the complexity just to save three clear commands (for a run that will take hours) to try to pack this confusion in one command. It can take hours to figure out the right way to use it.

Please delete fialr 123, it is fundamentally flawed and confusing.

Copy and paste the exact command you tried to run

../flair/flair.py 23 -q flair.aligned.bed -g
/hive/users/markd/projs/silvia/projs/flair/long+short-errors/refs/GRCh38.primary_assembly.genome.fa -f
/hive/users/markd/projs/silvia/projs/flair/long+short-errors/refs/gencode.v43.primary_assembly.annotation\
.gtf --shortread
/hive/users/markd/projs/silvia/projs/flair/long+short-errors/refs/junctions_from_sam_junctions.bed
--threads 8 -r flair.aligned.bam

How did you install Flair? source tree

What happened?

Starting correct...
Correct unrecognized arguments: -r flair.aligned.bam

What else do we need to know?

diekhans commented 7 months ago

the problem flair123 requiring command-line errors to be warnings is demonstrated here

https://github.com/BrooksLabUCSC/flair/issues/327#issuecomment-1957315199)