Pathlinker Subnetwork Generation - Conflict with Phenoscape

[jlaw9]

If the original network has node and edge styles set, the PathLinker subnetwork should have those styles as well (except for the blue diamond and yellow square sources and targets). This would allow the user to better interpret the PathLinker subnetwork. For example:

A cytoscape user has edge colors set according to specific types of edges (physical, phosphorylation, ubiquitination, etc.). In our PathLinker subnetwork results, their original styles would be helpful for them to see if there are any paths of more interest to them.

[jlaw9]

I performed another test, and discovered that the default style values are applied correctly to the ksp subentwork, but the "bypass" styles are not. We need to figure out how to fix this.

[jlaw9]

Found the documentation for setting a VisualMappingFunction which is how the bypass styles are set in the VizMapper section in the app cookbook.

Looks like all we need to do is loop through the nodes and edges of the subnetwork and use the apply function to copy the "bypass" or visual mapping styles of the original network to the subnetwork.

[jlaw9]

Scooter Morris gave an answer

[jlaw9]

The cyApplicationManager.getCurrentNetworkView(); function did not work for us. For now we are keeping the hack to select the newly-created subnetwork (see the files changed in pull request #45). In the future it will be better to implement Scooter's second suggestion and "copy" the code to generate the subnetwork.

[jlaw9]

Li implemented the Scooter's first suggestion by setting the sub-network to the selected network as the generated sub-network is automatically selected with _kspSubgraph = _applicationManager.getCurrentNetworkView().getModel();.

However, this current implementation is buggy if the Phenoscape App is open.

Specifically:

1. A different network is selected instead of the newly-created sub-network because of Phenoscape's network selection dropdown.
2. The style of the other network is applied to the sub-network

These issues should go away if we use Scooter's second suggestion to generate the sub-network.

[jlaw9]

Issue persists with PathLinker app v1.4 and Cytoscape v3.6, even though we fixed what we could on our end