This project is an experiment of mine to show how generative testing can be used to solve the algorithmic problem of detecting the x,y pairs on an island of black points in an m x n grid.
Grid.java contains the main logic for initializing a grid with black pairs and detecting an island. This is Algorithms 101 and not very exciting.
ProbabilisticIslandGenerator is the main driver for generative testing in this project. It operates in a reverse fashion of island detection in that it selects a random starting x,y pair then probabilistically determines if a neighbor should be black or not, recursively visiting black nodes. ProbabilisticIslandGenerator takes 3 parameters:
- height of the grid
- width of the grid
- probability that a neighbor is black or not
Here's an example grid generated with height=20, width=20, and probability=0.5:
--------------------
--------------------
--------------------
--------------------
--------------------
--------------------
--------------------
--------------------
--------------------
------------------BB
------------------BB
----------BB-----BBB
--------BBBB--BBBBBB
------BBBBBBB---BBBB
------BBBBBBB-BBBBBB
------BBBBBBBBBBBBBB
----BBBBB--BBBBBBBBB
BBBBBB-BB----BBBBBB-
BBBBBBBBBB--BBBBBBBB
BBBBBBBB-B----BBBBBB
The unit test in GridSpec.groovy asserts that the blackPairs passed to the Grid instance are the same that make up the island when starting from a random x,y pair in blackPairs.