Arvados

Currently, the convention I use to group paired end reads together would be to group by (.*)[12]. so foo_1 and foo_2 can be paired as well as foo.1 and foo.2 in their group. I run into a problem where I cannot specify which read I would like in what place when running a program because of a couple reasons: the groups are not alphabetical, sometimes the group indexes foo_2 and then foo_1 (for example: qr1hi-8i9sb-kdry51qdiy9tb42). If I want to run a program s.t. prog.py foo_1 foo_2 is necessary, I'm unable to do so. I also cannot grab the basename of foo_1 and foo_2 because the basename function does not separate by "_". The current workaround would be to glob the basename of files but that only works with periods. i.e. prog.py foo.1 foo.2 is possible because i can grab foo and search for foo.1 and search for foo.2.