More informative error message for filter_gtf_for_genes_in_genome.py (Python only task!)

[MatthiasZepper]

Description of feature

The rnaseq pipeline filters records of the transcript annotation GTF, which refer to scaffolds lacking a sequence record in the FASTA file.

If there is e.g. a name mismatch between the two files and no overlaps are found, the custom script ./bin/filter_gtf_for_genes_in_genome.py fails ungracefully:

/../../../.nextflow/assets/nf-core/rnaseq/bin/filter_gtf_for_genes_in_genome.py
 - 2023-02-28 15:56:04,858 INFO: Extracted 0 / 0 lines from
Pisum_sativum_v1a_genes.gtf matching sequences in Pisum_sativum_v1a.fa
  Traceback (most recent call last):
    File "/home/../../.nextflow/assets/nf-core/rnaseq/bin/filter_gtf_for_genes_in_genome.py", line 79, in <module>
      extract_genes_in_genome(args.fasta, args.gtf, args.output)
    File "/../../../.nextflow/assets/nf-core/rnaseq/bin/filter_gtf_for_genes_in_genome.py", line 60, in extract_genes_in_genome
      [logger.info](http://logger.info/)("All sequence IDs from GTF: " + ", ".join(sorted(x for x in seq_name_gtf)))
  UnboundLocalError: local variable 'seq_name_gtf' referenced before assignment

The reason is that the script never corroborates that seq_name_gtf was assigned by this step:

      for line in g.readlines():
          n_total_lines += 1
          seq_name_gtf = line.split("\t")[0]

There are further scenarios causing this step to fail, e.g. if g.readlines() doesn't yield a single line or if the GTF file is not tab-separated.

Thus, I suggest that the script filter_gtf_for_genes_in_genome.py should be improved to account for those cases and produce meaningful and actionable error messages. I think, this would be a good issue for the next Hackathon?

For more details, see the respective conversation on the #rnaseq Slack channel.

[pinin4fjords]

This was related to what I was doing in #1107, so I sorted it there.

@MatthiasZepper I assume it was ok to handle this even though I'm not a first-timer, since it didn't get sorted in the hackathon.

[MatthiasZepper]

@pinin4fjords: Absolutely! I am very grateful that you are working hard on the pipeline!

Because there had been a call to create issues for people to work on during the Hackathon, I went through my backlog of "I should do this somewhen" things for this pipeline and picked those that seemed doable for others without too much explanation.

I added the "First-timers only" label to highlight the easier ones for the Hackathon, but of course this is not imperative!