tiara tool still not working reliably

[wm75]

@Minamehr @bgruening after #1526 there still seem to be issues remaining with the tiara xml.

Apparently their can be cases when no files get produced for a selected taxonomy filter and in that case the mv/rename part of the wrapper is failing.

I don't know much about the tool, but if it's producing .dat files, then it's likely just keeping the input file suffix?
In that case a better approach than trying to rename all outputs, would be to symlink the input to get a proper suffix from the start?

[wm75]

I guess I would also not discover the outputs, but instead replace missing files with empty ones so that the number of outputs is deterministic. You do know what files could potentially be produced, correct?

[wm75]

Here's an actual bug report illustrating the issue with the current version:

mkdir ./results/ && tiara -t ${GALAXY_SLOTS:-4} -i '/data/dnb11/galaxy_db/files/1/4/e/dataset_14ea7add-9a69-475b-a5af-1b37d414b4f2.dat' -o ./results/main_result.txt  --pr '--pr' -m '3000' -p 0.65 0.65 --tf mit  pla  bac  arc  euk && ls -l ./results/ &&  mv ./results/mit*.dat ./results/mit.fasta && ls -l ./results/ &&  mv ./results/pla*.dat ./results/pla.fasta && ls -l ./results/ &&  mv ./results/bac*.dat ./results/bac.fasta && ls -l ./results/ &&  mv ./results/arc*.dat ./results/arc.fasta && ls -l ./results/ &&  mv ./results/euk*.dat ./results/euk.fasta

stderr

mv: can't rename './results/mit*.dat': No such file or directory

WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'
WARNING:galaxy.model:Datatype class not found for extension 'dat'

stdout

Classification took 363.01111817359924 seconds.
17.412689814577995 sequences per second.
5751729.072390296 base pairs per second.
Classification done.
First iteration statistics:
	archaea: 7
	bacteria: 233
	eukarya: 6023
	organelle: 2
	prokarya: 11
	unknown: 45
Second iteration statistics:
	plastid: 2

Output saved to ./results/main_result.txt.
Log file saved to ./results/log_main_result.txt.

total 1189376
-rw-r--r--    1 galaxy   galaxy      248737 Mar 13 11:21 archaea_dataset_14ea7add-9a69-475b-a5af-1b37d414b4f2.dat
-rw-r--r--    1 galaxy   galaxy    16114507 Mar 13 11:21 bacteria_dataset_14ea7add-9a69-475b-a5af-1b37d414b4f2.dat
-rw-r--r--    1 galaxy   galaxy   2060626184 Mar 13 11:22 eukarya_dataset_14ea7add-9a69-475b-a5af-1b37d414b4f2.dat
-rw-r--r--    1 galaxy   galaxy         876 Mar 13 11:21 log_main_result.txt
-rw-r--r--    1 galaxy   galaxy      506044 Mar 13 11:21 main_result.txt
-rw-r--r--    1 galaxy   galaxy      111255 Mar 13 11:21 plastid_dataset_14ea7add-9a69-475b-a5af-1b37d414b4f2.dat

no mitochondrial output got produced and the wrapper is not prepared to handle this.

[Minamehr]

Hi, we made a small modification in #1595, but I will check the details to see why it is still failing for empty groups.

In general, Tiara classifies sequences into different groups such as eukaryotes, bacteria, and others. Depending on the dataset, some of these groups may simply not be present. For example, a dataset may contain only bacterial sequences, so no eukaryotic output would be produced.

In such cases, the tool should not fail; it should simply produce no output for the absent group, or the wrapper should handle this case by creating an empty expected output file.