dev.roanh.cpqkeys

Class Main

java.lang.Object

dev.roanh.cpqkeys.Main

public class Main
extends java.lang.Object

Main class for the CPQ Keys canonization algorithm evaluation framework.

Field Summary

Fields

Modifier and Type	Field and Description
static java.util.List<Algorithm>	algorithms List of algorithms to evaluate.
private static int	DATASET_SIZE Number of graphs in a data set.
private static int	LABELS Number of labels to use for the data set.
private static java.nio.file.Path	LOGS Directory to store runtime logs in.
private static int	MAX_RULES Maximum number of rule applications to use to generate a data set.
private static long	MAX_RUNTIME The maximum amount of time in nanoseconds that an algorithm is allowed to spend evaluating a complete data set.
private static int	MIN_RULES Number of rule applications to use to generate the initial data set.
static java.lang.String	PYTHON_COMMAND Command to start a Python process on the host system.
private static int	RULE_GROWTH_FACTOR Increase in data set size after an algorithm processes a data set within the set time limit.
private static int	SEED Random seed to use to generate data sets.

Constructor Summary

Constructors

Constructor and Description
Main()

Method Summary

Modifier and Type	Method and Description
private static EvaluationResults	evaluateAlgorithm(Algorithm algo) Evaluates the given algorithm on randomly generated CPQs.
private static java.lang.String	findPython() Finds a Python 3 installation on the host system.
private static void	loadNatives() Loads the compiled JNI libraries required for certain algorithm implementations.
static void	main(java.lang.String[] args) Starts the algorithm evaluation process.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

LOGS

private static final java.nio.file.Path LOGS

Directory to store runtime logs in.

MAX_RUNTIME

private static final long MAX_RUNTIME

The maximum amount of time in nanoseconds that an algorithm is allowed to spend evaluating a complete data set. If the algorithms needs more time to run on a data set then it will not be given a larger data set. Conversely, if the algorithm manages to finish processing the whole data set within this time limit, then it is given a larger data set to run on next.

DATASET_SIZE

private static final int DATASET_SIZE

Number of graphs in a data set.

See Also:

Constant Field Values

MIN_RULES

private static final int MIN_RULES

Number of rule applications to use to generate the initial data set.

See Also:

GraphDataSet.fromCPQ(int, int, int), Constant Field Values

MAX_RULES

private static final int MAX_RULES

Maximum number of rule applications to use to generate a data set. Algorithms will not be given a larger data set to work with even if they manage to complete their work on a data set of this size within the set time limit.

See Also:

GraphDataSet.fromCPQ(int, int, int), Constant Field Values

RULE_GROWTH_FACTOR

private static final int RULE_GROWTH_FACTOR

Increase in data set size after an algorithm processes a data set within the set time limit. Expressed as the number of rule applications.

See Also:

GraphDataSet.fromCPQ(int, int, int), Constant Field Values

LABELS

private static final int LABELS

Number of labels to use for the data set.

See Also:

GraphDataSet.fromCPQ(int, int, int), Constant Field Values

SEED

private static final int SEED

Random seed to use to generate data sets.

See Also:

Constant Field Values

PYTHON_COMMAND

public static final java.lang.String PYTHON_COMMAND

Command to start a Python process on the host system.

algorithms

public static final java.util.List<Algorithm> algorithms

List of algorithms to evaluate.

Constructor Detail

Main

public Main()

Method Detail

main

public static void main(java.lang.String[] args)

Starts the algorithm evaluation process.

Parameters:

args - No valid command line arguments.

evaluateAlgorithm

private static final EvaluationResults evaluateAlgorithm(Algorithm algo)

Evaluates the given algorithm on randomly generated CPQs. The number of rule applications used to generate dataset CPQs will start from MIN_RULES and go up to at most MAX_RULES at a rate of RULE_GROWTH_FACTOR. Each dataset will contain DATASET_SIZE CPQs and at most LABELS labels will be used in the CPQs. Each dataset will need to be processed by the algorithm in MAX_RUNTIME nanoseconds. If the algorithm needs more time it will not be given a larger dataset next effectively ending the evaluation. The value for SEED will be used to generate the random dataset. All tasks will be executed on a single thread.

Parameters:

algo - The algorithm to evaluate.

Returns:

The result of the algorithm evaluation.

See Also:

GraphDataSet, Algorithm, ReportSummaryStatistics, RuntimeReport, EvaluationResults

loadNatives

private static final void loadNatives()
                               throws java.io.IOException,
                                      java.lang.UnsatisfiedLinkError

Loads the compiled JNI libraries required for certain algorithm implementations. Loads JNI dynamic libraries from the 'native' folder with libraries inside the 'native/lib' being loaded first so they can be used as dependencies for other libraries without causing errors.

Throws:

java.io.IOException - When an IOException occurs.

java.lang.UnsatisfiedLinkError - When loading a native library fails.

findPython

private static final java.lang.String findPython()

Finds a Python 3 installation on the host system. This will typically return the python3 command unless an anaconda 3 installation is detected.

Returns:

The Python executable.