arborworkflows / arborwebapps Goto Github PK

Need to rescale D3 dendrogram for the demo web app.

Not all of the data comes from files. Filename should be optional.

Currently you can't use control panel after starting VTKWeb view

Could be simple routines already implemented in Rpy "local server"

As far as I am aware tangelo.css and jquery are no longer included in the newest tangelo sources. Almost all the webapps require these to be in the "default tangelo web dir". Can we update these references or included them elsewhere?

We need to ensure defined order.

Stylistically, we need to make all of our Tangelo uses consistent w.r.t. common files and templates.

Analysis workflow failed during demo to Curtis. What's up with that?

Figure out why awesomeness<2 does not work in arbor Qt app

Needs to be added.

With help from @zackgalbreath and @jeffbaumes.

Zack, please take a pass at existing issues and see which you expect should be done by a milestone, or should be abandoned. I suppose a few could be left without milestone as "maybe someday" tasks.

Ports should be named.

The collapse/expand feature is done nicely in the Arbor tree display app, but when a branch is collapsed, lets also display the number of contained species (like the VTK collapse does). For example:

sagrei...carolinensis (42 species)

With help from @zackgalbreath most likely.

wesley@Wesleys-MacBook-Pro:~$ ^lsa
/Arbor.app/Contents/MacOS/Arbor
onCurrentRowChanged - InputType "TreeHeatmap"
Selected tree: "Tree"
ERROR: In /projects/Visomics-bin/VTK/Infovis/Core/vtkTreeDifferenceFilter.cxx, line 247
vtkTreeDifferenceFilter (0x112ac4680): tree #1's VertexData does not have a vtkDoubleArray named true node weight

onCurrentRowChanged - ContainerType "Geiger Tree Model Fitting 1" - voRemoteCustomAnalysis(0x104ad9a10, name = "Geiger Tree Model Fitting")
=> Analysis "Geiger Tree Model Fitting" DONE

Auto populate inputs, parameters on the web from these descriptions

The user has to refresh the page before moving from one table to another. The tag fields don't update correctly.

execution updates state in the instance

Can these be made to look the same to the user?

Could be upload of XML/R, XML/Python or equivalently upload of JSON/R, JSON/Python (or all of the above). Bonus: enable editing on the web.

Can be separate tab.

In the editor app

In Arbor, be able to replicate the analysis in workflowExamples/arbor_Picante_PhyloDiversity/GrasslandDataExample/grassland.workflow.R

Documentation/vis is in workflowExamples/arbor_Picante_PhyloDiversity/Biodiversity analysis in R Kembel.pdf

Depends on #34

Checkbox for using table transpose.

Fixed in 0933d48.

Get help from biology partners on this

Requires building shell around Celery-based analyses. With assistance from @zackgalbreath .

When fit continuous with EB parameter is run on anolis dataset, the resulting tree is not ultra metric anymore (all branches should still sum to the same value (not always 1.0, but always the same across the tree). Is the handle node branch length causing a problem?

Debugging by modifying R script to save input tree & matrix and modified tree before transmit back could isolate the algorithm and identify if algorithm is calculating wrong or if it is in the way Arbor is handling the data storage & transfers

Celery does not propagate errors back to the browser.

When importing python modules, Tangelo does not propagate import errors back.

Either warn or make unique. Right now data is overwritten, but the same name remains multiple times in the data manifest.

Is it necessary, can we get rid of it.

Including python modules to Tangelo does not preserve the path.

Skip past the handle when converting to tree for dendrogram

Curt could optionally get this to Zack and reassign to him

Can we support that type of Perl data-munging in Arbor?

It is on CRAN.

E.g. treemap, bubble map, etc. There was some discussion of a "tunnel" 3D visualization which would be down the road.