[BUG] --repeats argument can't be empty?

[gerverska]

Hey there! I'm working with v26.2.12 from the bioconda build, which seems to be ahead of the docs. Currently, there is an argument for --repeats listed under --help.

Optional arguments:
  -p , --proteins          protein alignments in GFF3 format [accepts multiple files: space separated] (default: [])
  -t , --transcripts       transcripts alignments in GFF3 format [accepts multiple files: space separated] (default: [])
  -r , --repeats           repeat alignments in BED or GFF3 format
  -w , --weights           user supplied source weights [accepts multiple: space separated source:weight] (default: [])
  -n , --num-processes     number of processes to use for parallel execution (default: number of CPU cores)
  -m , --minscore          minimum score to retain gene model (default: auto)
  --repeat-overlap         percent gene model overlap with repeats to remove (default: 90)
  --min-exon               minimum exon length (default: 3)
  --max-exon               maximum exon length (default: -1)
  --min-intron             minimum intron length (default: 10)
  --max-intron             maximum intron length (default: -1)
  -l , --logfile           write logs to file
  --silent                 do not write anything to terminal/stderr (default: False)
  --debug                  write/keep intermediate files (default: False)

Unlike the other optional arguments, it currently does not have a (default: []). I think this suggests an expectation for a --repeats argument that should not be there, and I've gotten this output from a test run:

[Mar 12 08:48 AM]: Python v3.12.12; GFFtk v26.2.12; numpy v1.26.4; natsort v8.4.0; interlap v0.2.6
[Mar 12 08:48 AM]: Namespace(subparser_name='consensus', fasta='test.fasta', genes=['test1.gff3', 'test2.gff3'], out='consensus.gff3', proteins=[], transcripts=[], repeats=None, weights=[], num_processes=None, minscore=None, repeat_overlap=90, min_exon=3, max_exon=-1, min_intron=10, max_intron=-1, logfile=None, silent=False, debug=False)
Traceback (most recent call last):
  File "/path/to/env/bin/gfftk", line 10, in <module>
    sys.exit(main())
             ^^^^^^
  File "/path/to/env/lib/python3.12/site-packages/gfftk/__main__.py", line 21, in main
    consensus(args)
  File "/path/to/env/lib/python3.12/site-packages/gfftk/consensus.py", line 29, in consensus
    check_inputs([args.fasta] + args.genes + args.proteins + args.transcripts + [args.repeats])
  File "/path/to/env/lib/python3.12/site-packages/gfftk/utils.py", line 40, in check_inputs
    if not is_file(filename):
           ^^^^^^^^^^^^^^^^^
  File "/path/to/env/lib/python3.12/site-packages/gfftk/utils.py", line 45, in is_file
    if os.path.isfile(f):
       ^^^^^^^^^^^^^^^^^
  File "<frozen genericpath>", line 30, in isfile
TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType

It looks like repeats, but also num_processes, minscore, and logfile are read in as having None when empty, while other optional arguments return the expected []. These then seem to produce the TypeError at the end of the output. I'll take a stab fixing consensus.py, but not sure when I'll get around to it...

[gerverska]

Ah, looks like the issue is actually with __main__.py. No defaults are given for the arguments called out in the first post.

[gerverska]

Ah, looks like the issue is actually with __main__.py. No defaults are given for the arguments called out in the first post.

Sorry--getting some help from Claude, looks like the main issue is utils.py. I changed the following around line 40:

# Change this:
   if not is_file(filename):
# To something like:
   if filename is not None and not is_file(filename):

Currently getting the following:

[Mar 12 12:21 PM]: Python v3.12.12; GFFtk v26.2.12; numpy v1.26.4; natsort v8.4.0; interlap v0.2.6
[Mar 12 12:21 PM]: Namespace(subparser_name='consensus', fasta='test.fasta', genes=['test1.gff3', 'test2.gff3'], out='consensus.gff3', proteins=[], transcripts=[], repeats=None, weights=[], num_processes=1, minscore=None, repeat_overlap=90, min_exon=3, max_exon=-1, min_intron=10, max_intron=-1, logfile=None, silent=False, debug=False)
[Mar 12 12:21 PM]: GFFtk consensus will generate the best gene model at each locus
[Mar 12 12:21 PM]: Parsing GFF3 files and clustering data into strand specific loci
[Mar 12 12:21 PM]: Merging gene predictions from 2 source files
{'predictions': {'RefSeq'}, 'evidence': set()}
Traceback (most recent call last):
  File "/path/to/env/bin/gfftk", line 10, in <module>
    sys.exit(main())
             ^^^^^^
  File "/path/to/env/lib/python3.12/site-packages/gfftk/__main__.py", line 21, in main
    consensus(args)
  File "/path/to/env/lib/python3.12/site-packages/gfftk/consensus.py", line 35, in consensus
    _ = generate_consensus(
        ^^^^^^^^^^^^^^^^^^^
  File "/path/to/env/lib/python3.12/site-packages/gfftk/consensus.py", line 158, in generate_consensus
    data = parse_data(fasta, genes, proteins, transcripts, log=log)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/path/to/env/lib/python3.12/site-packages/gfftk/consensus.py", line 1960, in parse_data
    loci, n_loci, pseudo = get_loci(Genes)
                           ^^^^^^^^^^^^^^^
  File "/path/to/env/lib/python3.12/site-packages/gfftk/consensus.py", line 1693, in get_loci
    if info["protein"][0].count("*") > 1:
       ~~~~~~~~~~~~~~~^^^
IndexError: list index out of range

I'll keep digging into this to see whether this issue is related to repeats, or whether it's a separate issue.

[gerverska]

Ah, looks like the issue is actually with __main__.py. No defaults are given for the arguments called out in the first post.

Sorry--getting some help from Claude, looks like the main issue is utils.py. I changed the following around line 40:

# Change this:
   if not is_file(filename):
# To something like:
   if filename is not None and not is_file(filename):

Currently getting the following:

[Mar 12 12:21 PM]: Python v3.12.12; GFFtk v26.2.12; numpy v1.26.4; natsort v8.4.0; interlap v0.2.6
[Mar 12 12:21 PM]: Namespace(subparser_name='consensus', fasta='test.fasta', genes=['test1.gff3', 'test2.gff3'], out='consensus.gff3', proteins=[], transcripts=[], repeats=None, weights=[], num_processes=1, minscore=None, repeat_overlap=90, min_exon=3, max_exon=-1, min_intron=10, max_intron=-1, logfile=None, silent=False, debug=False)
[Mar 12 12:21 PM]: GFFtk consensus will generate the best gene model at each locus
[Mar 12 12:21 PM]: Parsing GFF3 files and clustering data into strand specific loci
[Mar 12 12:21 PM]: Merging gene predictions from 2 source files
{'predictions': {'RefSeq'}, 'evidence': set()}
Traceback (most recent call last):
  File "/path/to/env/bin/gfftk", line 10, in <module>
    sys.exit(main())
             ^^^^^^
  File "/path/to/env/lib/python3.12/site-packages/gfftk/__main__.py", line 21, in main
    consensus(args)
  File "/path/to/env/lib/python3.12/site-packages/gfftk/consensus.py", line 35, in consensus
    _ = generate_consensus(
        ^^^^^^^^^^^^^^^^^^^
  File "/path/to/env/lib/python3.12/site-packages/gfftk/consensus.py", line 158, in generate_consensus
    data = parse_data(fasta, genes, proteins, transcripts, log=log)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/path/to/env/lib/python3.12/site-packages/gfftk/consensus.py", line 1960, in parse_data
    loci, n_loci, pseudo = get_loci(Genes)
                           ^^^^^^^^^^^^^^^
  File "/path/to/env/lib/python3.12/site-packages/gfftk/consensus.py", line 1693, in get_loci
    if info["protein"][0].count("*") > 1:
       ~~~~~~~~~~~~~~~^^^
IndexError: list index out of range

I'll keep digging into this to see whether this issue is related to repeats, or whether it's a separate issue.

Ah, this seems to be separate, so I'll open a new issue.