Added hitmap wrapper

[GraceAHall]

HiT-MaP is a tool for analysis of high-resolution MALDI imaging samples.

To pass the single test in the wrapper, a node with 64Gb mem is required.
Attempts have been made to reduce this using an even smaller dataset, but it seems to have little impact on memory usage.

[bgruening]

@GraceAHall thanks a lot!

Do you know already this project: https://github.com/galaxyproject/galaxy-language-server

This might help you a bit with formatting. Given an input section, it can bootstrap you a test section and a command section. It will show you nice hits and has auto-completion of parameters etc. Moreover, it has common snippets for conditionals, integers ...

[bgruening]

The test data is too big, unfortunately. We try to be below 1MB. Can you maybe use the ibd/imzml files from the Cardinal tools?

[bgruening]

is ln enough?

[GraceAHall]

Unfortunatenly not. Hitmap needs the file extension to be intact, but all galaxy data has extension '.dat'

[bgruening]

But instead of using cp you can do ln -l '${file_inputs.imzml_file}' expdata/sample.imzML

That saves a copy operation. And works in 95% of all cases. In some cases, the tool does not follow symlinks ... I hope this is not the case here.

[GraceAHall]

Mmm gave this a shot, but since it's a container, it was having difficulty creating the link.

I think this is an improvement which could appear in a later version of the wrapper.

HiTMaP has been very temperamental and hard to get working, so I'd rather just keep as is rather that spending time on this optimization.

[bgruening]

the license refers, I think, to code that you ship with this tool not the license of the tool

Can you please also include a profile="20.05"

[bgruening]

if you like please include a bio.tools identifier. The edam ontology is not strictly needed anymore, we will get those with the bio.tools IDs. See the xref tag.

[GraceAHall]

HiT-MaP doesn't have a listing in bio.tools. Should I leave this as edam topics?

[bgruening]

In this case, leave it! We could also add it to https://github.com/bio-tools/content, but this repo is not yet in production.

[bgruening]

single quote for the path

[GraceAHall]

do you mean $tool_directory/hitmap_gen_html.py -> '$tool_directory/hitmap_gen_html.py'?

[bgruening]

yes :) the reason is that you might have "spaces" in your tool_directory path .. in this case the tool breaks.

[GraceAHall]

haha oh yeah that's a great point. Can happen!
Added to the next commit :)

[bgruening]

The Galaxy datatype imzml is a composite datatype, which means it contains 2 files. You can see how this can be used with the Cardinal tools.

https://github.com/galaxyproteomics/tools-galaxyp/blob/master/tools/cardinal/macros.xml#L18

[GraceAHall]

I see. I originally tried to do it that way, but couldn't work out how to upload two files (.imzml + .ibd) together as a composite type in the galaxy uploader.

I just went with the approach that the users upload the imzml and ibd as seperate files into their history, then select those when performing the run.

What do you think about that approach?

Also thank you again so much for looking over these submissions and helping me out. I really appreciate it!

[bgruening]

What do you think about that approach?

I think its can result in more errors from the user side, picking the wrong combination of files etc.
Do you know why this did not work for you? Can you try this on EU. I think its working on EU.

[GraceAHall]

Ohhhh I see!
Yes this is working fine, I just wasn't aware of that option. Very neat!
Added to the next commit

[bgruening]

Hi @GraceAHall! Do we want to work on this one in 2 week? There is a new version on dockerhub: 968278fe1272192cff0f598df5981fd7301d8a4d394370765033409eb3a5053e

[bgruening]

@GraceAHall I guess we forgot this. Do we still want to do it?