Here you'll find source code for kotton -- a wrapper to run algorithms computing weighted straight skeletons. Currently it can call two particular implementations -- campskeleton and triton (latter is currently also contained in this repository).
To build kotton, you'll need to build (patched) campskeleton and triton first.
Clone
and run mvn compile install in each to install to local maven repo.
For some theory on these, see
http://www.twak.co.uk/2009/05/engineering-weighted-straight-skeleton.html
and
http://www.twak.co.uk/2011/01/degeneracy-in-weighted-straight.html .
The code found under triton should be buildable with gradle.
It requires manual download of one library -- see triton/lib/EXTRA_LIBS.
Run gradle publishToMavenLocal to install it.
Once you've installed the dependencies, kotton can be built via gradle build
inside kotton directory. You can then run
java -jar build/libs/kotton-all.jar --help to see available runtime options.
For better performance, consider setting -Xmx4096M or more :).
You still can run triton standalone:
-
Main class starts the interactive swing applet (default in the built jar file).
-
SingleRun class runs the algorithm for a given data file and writes results to given output file(s) -- but consider using kotton instead.
-
BatchRun expects a file where each (non-empty) line contains a list of input and output parameters, and runs the algorithm for every such line (see
bin/run_batch.shanddata/batches/simple-batch.txt). It is advisable to split very large batches into chunks of under 10k items each, as the speed seems to deteriorate steadily here (due to garbage collection and swing issues?). Again, consider using kotton instead.
From triton directory, those can be invoked via gradle run (see build.gradle file for details),
or (after a successful gradle build) from the built jar (build/libs/triton-all.jar) directly.