Transcript glyphs

[garrettjstevens]

This adds a ProecessedTranscript glyph, as well as some other glyphs that are used to make it (Chevron, Segments, and Subfeature). Here's a side-by-side of Jbrowse 1 & 2. More work could be done in the future to port more of the config information from JB1 to JB2 (like colors, etc.), but the main information is there.

resolves #240

[cmdcolin]

I'm not a big fan of the little angle things... I think it's visually noisy and makes it actually confusing where the true boundary of a feature is.

[cmdcolin]

Here are alternative styles that could be considered

UCSC
https://www.researchgate.net/profile/Shan_Gao13/publication/268117317/figure/fig2/AS:289267885592580@1445978247004/Example-custom-annotation-on-UCSC-Genome-Browser-Clicking-on-the-identifier-UN09857.png

Tip arrow (which can also slide with feature a la jbrowse1)
http://gmod.org/mediawiki/images/thumb/0/0b/Menu-customization2.png/600px-Menu-customization2.png

Gene label contains arrow
http://locuszoom.org/images/news_201606_lzjs_sample.jpg

[garrettjstevens]

I like having directionality indicated by the glyph, I think it's informative, but I agree it does need a bit of work. I though about making only one end pointed, like this:

The only thing I don't like about that is if you're zoomed in to the flat end, you can't tell if it's just a box or has direction.

[garrettjstevens]

I'd like to keep this PR focused on the SVG glyph architecture and not specific glyphs themselves, so I created #460 to keep discussions on glyphs.

[codecov]

Codecov Report

❗ No coverage uploaded for pull request base (master@00b0596). Click here to learn what that means.
The diff coverage is 62.46%.

@@           Coverage Diff            @@
##             master    #458   +/-   ##
========================================
  Coverage          ?   64.1%           
========================================
  Files             ?     238           
  Lines             ?    7753           
  Branches          ?    1785           
========================================
  Hits              ?    4970           
  Misses            ?    2557           
  Partials          ?     226

Impacted Files	Coverage Δ
packages/core/components/ResizeHandleHorizontal.js	36.66% <ø> (ø)
...leElementTypes/renderers/ServerSideRendererType.js	100% <ø> (ø)
...iew/src/BasicTrack/util/serverSideRenderedBlock.js	91.42% <ø> (ø)
packages/protein-widget/src/Viewer.js	71.66% <ø> (ø)
packages/jbrowse-web/src/rpc.worker.js	45.16% <ø> (ø)
packages/core/BaseAdapter.ts	89.47% <ø> (ø)
packages/core/rpc/RpcManager.js	91.89% <ø> (ø)
packages/core/util/index.ts	59.68% <ø> (ø)
...w/src/CircularView/models/viewportVisibleRegion.js	84.09% <ø> (ø)
...ckages/jbrowse1/src/JBrowse1Connection/jb1ToJb2.ts	0% <0%> (ø)
... and 19 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 00b0596...4d93587. Read the comment docs.

[cmdcolin]

Currently the human Gencode 19 track is failing with a TypeError

I can see certain things like hasSubSub and the Subfeatures glyph and I don't really understand the purpose of these?

We need to standardize a strand representation so that we do not run into having to check for -1,1,+,and - in all cases, just one of the other.

Sort of concerned about mRNA and transcript both being ProcessedTranscript by default.

[garrettjstevens]

I'm fine with standardizing, it looks like everything so far uses 1 and -1. Should we standardize on that?

Also, I was using JB1 as a guide in determining glyph types, but there could be something I'm misunderstanding:
https://github.com/GMOD/jbrowse/blob/c10e55ba2924be0f72fde0b36c57a9881ddf3038/src/JBrowse/View/Track/CanvasFeatures.js#L212-L220.

The Subfeatures glyph lays out its child glyphs vertically. This glyph gets used whenever one of the subfeatures has subfeatures itself.

I can see the Gencode v19 error, I'll see what's going on there. I also realized that reversing does not work properly, so that will definitely need to get fixed, too.

[garrettjstevens]

The Gencode track doesn't error now, and the glyphs now display properly when the view is reversed.

[rbuels]

A function on the function? Is this a thing?

[garrettjstevens]

With the original Box glyph, layOut was a static method on the class component. When I refactored to hooks, this seemed to be the equivalent way to do it (similarly to how propTypes can be a static property of a class component or defined on the function).

[rbuels]

Could you add a little explanation of the glyph architecture?

[cmdcolin]

If possible it would be good to have a rendering of the non-coding features that isn't just a line, currently it is just a line for the non-coding features I think on the Gencode test data

[garrettjstevens]

@rbuels I've added a readme

@cmdcolin I added a little arrow to the end so it's not just a line, similar to JB1.

[cmdcolin]

Some of the features like ENSG00000227169.2 have feature.get('type') undefined?

MODEL.views[0].setNewView(5.377899045020464,857364.7635717909)

[garrettjstevens]

I'm not sure what you mean. That feature's type is 'gene'.

[cmdcolin]

Oh sorry I was looking at it console.log(feature.get('name')) on it and thought I was logging type..

But I guess I am wondering why does that feature not have exons in the displayed glyph?

[garrettjstevens]

I thought exons would be part of #241. The current behavior does seem consistent with JB1 under certain configurations, though:

My JB1 track is configured as

[tracks.TestTrack]
category=Other
key=Gencode v19
label=ATestLabel
storeClass=JBrowse/Store/SeqFeature/NCList
type=JBrowse/View/Track/CanvasFeatures
transcriptType=transcript
urlTemplate=http://s3.amazonaws.com/jbrowse.org/genomes/hg19/gencode/{refseq}/trackData.json

[cmdcolin]

Ok so it filters so it only gets CDS and utr subfeatures and plots those and this feature only has exon subfeatures