#4 first vis plugin

[wilzbach]

Hi all,

so this is my first, infancy Galaxy visualization plugin as discussed in #4.

1. I don't know whether it is a good idea to commit the "binary" (compressed,minified) MSA viewer code into this repo. (There exists also a CDN version.)
2. Is there a better way to load a resource relatively to the plugin folder?
   (this is seems a bit inflexible)

${h.javascript_link( root + 'plugins/visualizations/biojs-msa/static/msa.min.js' )}

1. Any other suggestions are welcome

[bgruening]

Seb;

wonderful!
Thanks a lot. Out of curiosity, can your plugin be applied to *.msa files as well? Or the native clustalw format?

Regarding your questions, can you try:

${h.javascript_link( plugin_path + '/static/msa.min.js' )}

I think shipping the binary JS file is ok, because we can't rely on internet connection on every Galaxy host. A few Galaxy instances are not allowed to be part of the Internet (private data issue ...).

[wilzbach]

I couldn't find Clustal as input format. So I just picked sequences. Is there a special Clustal format?

plugin_path is great, but unfortunately it is an absolute path :-(

[bgruening]

@greenify what kind of input formats do you support? I don't understand the problem with the absolute path, what do you need to do, or what is not working?

[wilzbach]

what kind of input formats do you support?

Currently Fasta and Clustal

don't understand the problem with the absolute path, what do you need to do, or what is not working?

well /home/... doesn't work in a browser. It thought plugin_path would similar to this:

plugin_path = "/config/plugins/visualizations/biojs-msa/"

[jmchilton]

Exactly - plugin_path is meant for consumption by Python components - it is a file path not a URL.

[bgruening]

Here is the clustal datatype. https://github.com/galaxyproject/tools-iuc/blob/master/datatypes/msa/clustal/datatypes_conf.xml But this can be fixed after it is merged.

[carlfeberhard]

Heya, Seb.

Looks good. The config file is well formed. Might want to close the html tag in the mako template.

For loading scripts, that form is correct for now. Unfortunately, the static directory served by galaxy is set up to be explicit as possible and that makes it a bit long. I'll look into whether there's a way to make relative paths a bit easier.

I really look forward to seeing it in action

[wilzbach]

@bgruening Should I use the Clustal datatype even though it is not a standard datatype?

@carlfeberhard any news on the path problem? Can/Should I send a patch to the galaxy mercurial project? Btw it would be super awesome if Galaxy would migrate to Git(hub).

[bgruening]

@greenify I will add the clustal datatypes later. But can you add a comment which datatypes are supported? Than we can use this as documentation. Preferably as EDAM http://edamontology.org annotation.

[bgruening]

@carlfeberhard This seems to be fine, right? I will merge this tomorrow if no one complains :)

[wilzbach]

@bgruening sorry for being so unresponsive. I am not sure what you expect as EDAM - the EDAM IDs?

ClustalW http://edamontology.org/format_1982
FASTA http://edamontology.org/format_1929

(the MSA only supports FASTA and Clustal at the moment.)

[bgruening]

Hi @greenify it was just a recommendation for your project and for the documentation. At some point we will have problems to map datatype names. Is it clustal or clustalw or Clustal. If we create such a config file for every biojs visualisation we should agree on some ontology and this is my proposal :)

I will merge this now. Thanks a lot!
Bjoern