Picrust2

[EngyNasr]

The 6 wrappers of Picrust2 https://github.com/picrust/picrust2/wiki are all added:

• 4 Main (metagenome_pipeline, pathway_pipeline, hsp, place_seqs)
• 2 Others (add_description, shuffle_predictions)

Issues:

• 2 Main:

1. in place_seqs, the --placement_tool when epa-ng is selected, I got errors even in the tools command line (May be because of the test-data available is not suitable)
2. in pathway_pipeline, the --no_regroup option when set to true, I got errors even in the tools command line (May be because of the test-data available is not suitable)

(Temp solutions removing this options for now, create an issue on picrust2 github)

• 1 not important:
  shuffle prediction, the tool itself doesn't output the files in the output directory specified in the --outdir argument, so nothing to present on usegalaxy

Extras:

• Add more tests (I have tested all options manually, so I will just add them in the wrappers tests)

Thanks a lot,
Engy

[bernt-matthias]

The .shed.yml file seems missing

[bernt-matthias]

create an issue on picrust2 github

sounds like a good idea. maybe the developers have nice and small test data that you can use.

[bernt-matthias]

Cool. Only one error: For picrust2_place_seqs you will need to add zip as requirement.

[bernt-matthias]

Looks really good. Added a few comments.

[bgruening]

@EngyNasr what is the status here? Do you need help?

[EngyNasr]

@bgruening, Back to picrust2 🎉, I have created an issue to picrust2 main github page for the issues above, however we can have the tool as it is (after removing the 2 above arguments that are causing issues), then update the tool later based on their response (add them back)

[EngyNasr]

Sooo :)

The first two issues where totally solved, where the only problem where the data sets.

The third regarding the output of shuffled prediction, this is a tool problem, however thanks to Pietro he taught me how to play around to reach the temp path of the input file where the output is also saved and now I can present the output on Galaxy

Now all issues are solved, just testing last time and will commit all changes 🎉 🎉 🎉

[EngyNasr]

I have changed as per comments but still have a problem with the test data size will solve that and commit changes

[EngyNasr]

Step 3 ;) ?

[bernt-matthias]

Thanks for working on this @EngyNasr . Few more comments from me

[bgruening]

Your last commit is cheating ;) You do not test the output and it's actually a valid error.

/tmp/tmpuv7474tj/job_working_directory/000/6/tool_script.sh: line 22: zip: command not found

You are missing zip or gunzip, probably both as a dependency.

[bebatut]

Thanks a lot @EngyNasr for this work!

In addition to my almost 100 (😱) inline comments, 2 general comments/questions:

• I did not see any wrapper for picrust2_pipeline.py. Do you plan to wrap it?
• I have mixed feeling about the intermediate files. I am not sure if it is really useful to output them as zip file. Either we expose them individually (specially if they are always the same), or we hide them. For the first version of the wrapper, I would hide them (so comment everything related to that).

[EngyNasr]

the full pipeline is all arguments together in one wrapper, is it really important to have ?

[bgruening]

I don't think so, we have workflows for that. At least in the first version we can skip this imho.

[bebatut]

the full pipeline is all arguments together in one wrapper, is it really important to have ?

I expected that people knowing Picrust may use this tool first (as they may not know the different steps). So I would say yes

I don't think so, we have workflows for that. At least in the first version we can skip this imho.

@bgruening it is the script that Picrust developers put first (https://github.com/picrust/picrust2/wiki/Full-pipeline-script). So I think it may be the most important tool to have there. The other ones are for running the steps one by one.
But I may be wrong there

[EngyNasr]

its not very urgent to have now, so I will currently close the PR

[EngyNasr]

The tool needs a follow up on this Issue

[bernt-matthias]

Hey all. Marked add comments as resolved where I thought that they are. Also did some first edits. Hope that I can continue the next days. I get more and more request by my users for this tool. So I'm happy to jump in.

TODO:

• test run time
• some parameters could be put in macros
  • hsp
• allow tabular.gz in and output
  • leave it for later

[bernt-matthias]

Hi @bgruening good to merge?

[bgruening]

Oh yes, very cool! Thanks a lot everyone!