-
Notifications
You must be signed in to change notification settings - Fork 1
Progress list
lilit-nersisyan edited this page Jun 10, 2014
·
8 revisions
Allow the user to modify sorting results - i.e. if they wish to place one node before the other.
- Examples of envisioned config files for rule mapping (Due 10 June)
- Network motif paper revision, test case construction (Due 10 June)
- Specification of GUI panel functional items (Due 10 June)
- Code revision, clean-up (Due June 6)
- Attribute mapping in Cytoscape (Due June 5)
I've incorporated JGraphT java library for graph based algorithms. Basically, the use case until now is the following:
- current network in Cytoscape is mapped to Graph (psfc).
- each Graph object contains a jgrapht instance, which takes the topology of the graph and returns results of several traversal algorithms (breadth first search, depth first search, and closest first iteration). From these three, only closest first iteration is suitable for ordering the nodes according to their distance to the start node. Since the traversal output is simply the order of nodes, I have to do extra-work for assigning levels to each of the nodes according to their distance to start. Another alternative (which I prefer for now) is to use dijkstra's shortest path algorithm implemented in JGraphT, and assign level values to each node, according to it's shortest path distance to the start node (the start node is the unique node in the graph, who's in-degree is 0. If there are multiple 0 in-degree nodes, the algorithm firstly creates another node and connects it to all 0 in-degree nodes). Should the start node be seen in Cytoscape after creation?
- Thoughts: implementing traversal algorithm myself would result in less redundant code, but more of redundancy in implementation and will probably be more error-prone.
- The levels assigned to each node will be populated in CyTable for nodes in the network (will be implemented shortly).
-
Sorting (Due by June 3, 2014)
- Basic sorting algorithms (at least one) implemented.
- The algorithm should work on Graph class. It should be tested with use cases in Cytoscape: this way CyNetwork - PSF-Graph communication will also be tested.
-
Cytoscape-PSF communication (Due by May 30, 2014)
- This milestone basically consists of the following two sub-tasks:
- 1. Basic structures for Graph implemented: with a couple of test-data.
- 2. CyNetwork -> PSF graph, and reverse communication basically established.
- This milestone basically consists of the following two sub-tasks:
-
Framework and code design specification
- Initial simple graph structure and psf algorithm implementation - first step for logical part of the code design specification
- Thoughts on GUI look and functionality