ekuiter / SATGraf

Headless fork of SATGraf for visualizing SAT community structure

Headless SATGraf

This fork of SATGraf visualizes community structure of propositional formulas in DIMACS format. Operation can be automated on the shell.

git clone https://github.com/ekuiter/SATGraf
cd satgraf
./gradlew shadowJar
java -jar build/libs/SATGraf-1.0-SNAPSHOT-all.jar exp -f some-formula.cnf -o some-formula.jpg

Notes

• The fgpu option is under development and may not work fully.
• The circle layout algorithm works poorly with disjoint graphs.
• The Louvian method requires a non-disjoint graph
• Either -f or -u must be specified.

Usage

java -jar SatGraf.jar [exp] <options>

com - View the static community representation of the formula.   
	-f --file(default = null) The file (either .cnf or .sb)   
	Must be an existing file For which you have: read permissions   
	-u --url(default = null) A file URL (either .cnf or .sb)   
	-c --community(default = ol) The community detection algorithm   
	Must be one of:    
		l - The Louvian community detection algorithm   
		ol - The Online community detection algorithm   
		cnm - The Clauset-Neuman-Moore community detection algorithm   
	-p --pattern(default = null) A list of regex expressions to group variables (not yet implemented)   
	-l --layout(default = f) The layout algorithm to use   
	Must be one of:    
		f - An implementation of the Fruchtermon and Reingold force directed layout algorithm   
		fgpu - A GPU implementation of the Fruchtermon and Reingold force directed layout algorithm   
		kk - An implementation of the Kamada-Kawai placement algorithm   
		c - A basic layout algorithm, nodes are layed out within a community using the FR layout algorithm. The communities are then layed out in a circle, with those most connected next to each other.   
		grid - A basic layout algorithm, nodes are layed out within a community in a grid pattern. The communities are then layed out on a grid   
		gkk - A basic layout algorithm, nodes are layed out within a community using the KK algorithm. The communities are then layed out on a grid   
		jung - use Jung (ISOM)
		fr - An implementation of the Fruchtermon and Reingold force directed layout algorithm, that selects a random set of nodes (rather than all pairs)
	-m --format(default = auto) The format of the file, and desired graph representation
	Must be one of:
		auto - accepts: 
			.cnf - Load a DIMACS format file into a VIG and compute communities
			.sb - Load a graph representation from SB (JSON) file
		literal - accepts: 
			.cnf - Load a DIMACS format file into a LIG and compute communities
	-n --node-color(default = auto) node edge colouring implementation to use
	Must be one of: 
		auto - Nodes are blue
	-e --edge-color(default = auto) The edge colouring implementation to use
	Must be one of: 
		auto - Edges are coloured according to their community (or white for intercommunity)