Custom output format raises error in subprocess.run

[johanneswerner]

Hi,

I am trying to run diamond as part of a luigi workflow and need a custom output (#395). Therefore, I am using subprocess.run() to run diamond which fails due to the --outfmt format, although I am not clear why (running that command from bash works).

subprocess.run(['diamond',
                'blastp',
                '-d', 'refdb.dmnd',
                '-o', 'results.tsv',
                '-f', '6 qseqid sseqid pident length evalue bitscore full_sseq',
                '-q', 'query_file.fasta'],
               check=True)

This raises the following error:

Opening the database...  [0.151s]
Error: Invalid output format. Allowed values: 0,5,6,100,101,102
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/ubuntu/miniconda3/envs/query_genes/lib/python3.8/subprocess.py", line 512, in run
    raise CalledProcessError(retcode, process.args,

I am not sure where this comes from, maybe it has something to do with the quoting that subprocess.run() performs.

Did anyone have similar problems or knows a solution for that? Thank you very much.

[bbuchfink]

I would guess subprocess.run will insert additional quotation marks around 6 qseqid sseqid pident length evalue bitscore full_sseq. Not sure how to best work around that. If needed I can modify Diamond accordingly to allow this.

[johanneswerner]

Is it possible to return the "invalid output format" to the screen? I have the feeling that something else is happening, because if I am combining parameter and argument (i.e. '-f 6 qseqid sseqid pident length evalue bitscore'), I still get the same error.

I just found out, I can use it like this as an alternative (which would not require any modification):

subprocess.run('diamond blastp -d refdb.dmnd -o results.tsv -f 6 qseqid sseqid pident length evalue bitscore full_sseq -q query_file.fasta', shell=True, check=True)

This also works for me but I am curious where the error with the list calling comes from (but of course, since the second version works, this is also fine for me). Thank you for your answer.

[bbuchfink]

I added a message like that: 810b6834a4a807cfa053112cfbd3d6979a0af43c:

I think you probably need to do this: '6', 'qseqid', 'sseqid', 'pident' etc.