Visualize.py - Error in finding input files

[GiliWolf]

Hi, I encountred some problems when running -
python visualize.py --viz_output <output_dir> --gene_list <gene_list> <main_output_dir>

After running isoquant using the command -
isoquant.py --yaml <single_sample>.yml --data_type pacbio_ccs --reference GRCh38.primary_assembly.genome.fa --index GRCh38.primary_assembly.genome.mmi --genedb gencode.v49.annotation.gtf --complete_genedb --force --sqanti_output --threads 8

where the tree of the main_output_dir is -

<main_output_dir>
├── GRCh38.primary_assembly.genome.fa.fai 
└── isoquant_output 
        ├── <SAMPLE>
                 ├──<SAMPLE>.corrected_reads.bed.gz
                 └── <SAMPLE>.discovered_gene_counts.tsv
                  ................
        ├── alignment.log
        ├── gencode.v49.annotation.bed
        ├── gencode.v49.annotation.db
        ├── isoquant.log
        └── .params

The first error I got was -
AssertionError: Params file not found: <main_output_dir>/.params
So, I tried to change the output directory to - "<main_output_dir>/isoquant_output "
But then I got the error -
Error parsing GTF file: Database file isoquant_output/gencode.v49.annotation.db does not exist
(since in the params file - genedb_filename=isoquant_output/gencode.v49.annotation.db,
and in the "post_process.py" script: _process_params function, it's extracted using -
self.genedb_filename = params.get("genedb_filename")

I fixed it, and got an additional error (similar to #346 issue) -
"TypeError: expected str, bytes or os.PathLike object, not NoneType"

When looking into the code in "post_process.py" script, I've realised there is a problem when trying to run 'visualise.py' when using a yml file containing only one "experiment" (or sample), as the function "find_files_from_yaml" tries to extract count/tpm information only using the 'combined*' files, which are not generated for a single sample.

I think this is the source of both of the errors explained above.

Hope everything is clear.
Thank you,
Gili

---

my yml -

[
  data format: "unmapped_bam",
  {
    name: "sample_1",
    long read files: [
      "path/repeat1_demux.bam",
      "path/repeat2_demux.bam",
    ],
    labels: [
      "repeat1",
      "repeat2",
    ],
    prefix: "sample_1",
  },
]

[andrewprzh]

Dear @GiliWolf

Thanks a lot for the report!

The visualizer was developed by a collaborator of mine, who now seems to be busy with other projects.
Unfortunately, I myself don't have enough resources to maintain it.
I hope we will find out the way to support visualization in the future.

Best
Andrey