create new command line tool with add_OGS_track.py

[mpoelchau]

Two scripts to convert:

• https://github.com/NAL-i5K/ColorByType/blob/master/scripts/add_OGS_track.py
• https://github.com/NAL-i5K/ColorByType/blob/master/scripts/add_NCBI_annotation_track.py

I think we would eventually need to make 2-3 separate 'final' workflows, depending on the OGS type that we're loading. Would be nice if we could use an argument instead though. Whichever way it's done, the new command line tool would substitute flow_apollo2/flatfile-to-json.cwl in the current implementation

This issue might require updating the original add_OGS_track.py script to play better with cwl.

Related to #26.

[r06942072]

I think it might have two options to manage the cwl directory structure

• first, 3 folders
  • flow_apollo2 (original one, use flat-file-to-json.cwl)
  • flow_apollo2_OGS
  • flow_apollo2_NCBI
• second, keep it one folder flow_apollo2, inside this folder, create three different workflow.cwl file
  • flow_apollo2/workflow.cwl
  • flow_apollo2/workflow_OGS.cwl
  • flow_apollo2/workflow_NCBI.cwl

---

I think it might be more elegant and wrangler-friendly to keep only single final-workflow.cwl file, because it is no need to choose between which commands to use.

So, we should add some condition in cwl, to let cwl smart enough to choose script in step flow_apollo2

[mpoelchau]

We discussed how to 1) autodetect whether the gff file is from NCBI or an OGS, and 2) whether this should be re-written from python to CWL.

1) @mpoelchau will connect with NCBI to see if there's a standard pragma line for the eukaryotic annotation pipeline; 2) @mpoelchau will review https://github.com/NAL-i5K/ColorByType/blob/master/scripts/add_NCBI_annotation_track.py to see if it can be rewritten.

Also, we agree that it would be better to have multiple workflow scripts than multiple folders (option 2 from above comment)

[mpoelchau]

@childers can you share a flatfile-to-json command that you've been using to load multi-type tracks to Apollo 2.6+ so we can start planning how to modify this issue?

[mpoelchau]

@childers says the main change is to specify "className": "{type}". We'll update add_OGS_track.py to reflect this and will test it.

[mpoelchau]

After discussions with @childers, sounds like we need to:

• update https://github.com/NAL-i5K/ColorByType/blob/master/scripts/add_OGS_track.py to reflect "className": "{type}"
• Make a decision whether that script should go in a new repo since it won't use colorbytype anymore?

[mpoelchau]

@childers can you link to an example track where "className": "{type}" is already implemented? That way I can try to move this issue forward.

[childers]

here is an example track, taken from a Varroa browser config file ...Vdes_3.0/jbrowse/data/trackList.json

Here is the track

        {
         "category" : "TEST",
         "compress" : 0,
         "key" : "TEST_feature_by_type",
         "label" : "TEST_feature_by_type",
         "storeClass" : "JBrowse/Store/SeqFeature/NCList",
         "style" : {
            "className": "{type}",
            "description" : "product, note, description",
            "subfeatureClasses" : {
               "CDS" : "gnomon_CDS",
               "exon" : "container-100pct",
               "wholeCDS" : null
            }
         },
         "trackType" : "WebApollo/View/Track/DraggableNeatHTMLFeatures",
         "type" : "WebApollo/View/Track/DraggableNeatHTMLFeatures",
         "urlTemplate" : "tracks/NCBI_Annotation_Release_100_Gene/{refseq}/trackData.json"
      }

[mpoelchau]

Thanks @childers !

[mpoelchau]

replaced by #99, closing.