org.epics.graphene.profile.utils

Class ProfileAnalysis

java.lang.Object

org.epics.graphene.profile.utils.ProfileAnalysis

public class ProfileAnalysis
extends Object

Provides the tools necessary for comparing the output tables of profiling to determine changes in the results.

Provides options to analyze the two tables types:

• ProfileGraph2D 1D tables
• MultiLevelProfiler 2D tables

The ProfileAnalysis class is purely static and unable to be instantiated.

WARNING: the analysis operations assume the preconditions are satisfied with only minor checks. Therefore it essential that the appropriate file types are selected.

Author:

asbarber

Field Summary

Fields

Modifier and Type	Field and Description
static double	STATISTICALLY_SIGNIFICANT Percent level in which difference in values are considered statistically significant and thus take appropriate (warning) action.

Method Summary

Methods

Modifier and Type	Method and Description
static List<String>	analyzeTables1D() Computes the difference between the last two records in each 1D table and analyzes whether the change was significant.
static void	compareTables2D() Brings a dialog box to ask the user to select two table based CSV files and compares the differences of the tables.
static void	compareTables2D(File fileA, File fileB) Computes the difference of each cell between the first file selected and the second file (first - second).
static double	percentChange(double valInit, double valFinal) Calculates the percent change from initial to final.
static String	performanceChange(double valInit, double valFinal) Determines the significance of the change from the initial value to the final value.

Methods inherited from class java.lang.Object

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

Field Detail

STATISTICALLY_SIGNIFICANT

public static final double STATISTICALLY_SIGNIFICANT

Percent level in which difference in values are considered statistically significant and thus take appropriate (warning) action.

See Also:

Constant Field Values

Method Detail

compareTables2D

public static void compareTables2D()

Brings a dialog box to ask the user to select two table based CSV files and compares the differences of the tables.

Computes the difference of each cell between the first file selected and the second file (first - second).

Precondition: both files selected are MultiLevelProfiler table files of the same graph type.

compareTables2D

public static void compareTables2D(File fileA,
                   File fileB)

Computes the difference of each cell between the first file selected and the second file (first - second).

Precondition: both files selected are MultiLevelProfiler table files of the same graph type.

Parameters:

fileA - .CSV MultiLevelProfiler formatted file, considered initial file (compares fileA - fileB)

fileB - .CSV MultiLevelProfiler formatted file, considered initial file (compares fileA - fileB)

analyzeTables1D

public static List<String> analyzeTables1D()

Computes the difference between the last two records in each 1D table and analyzes whether the change was significant.

Conditions to Analyze:

• Supported files can be found
• The 1D tables are in the right .CSV format
• The file has two rows of data to compare

Returns:

a list of: for each 1D table analyzed, a generated analysis message including the graph profile and the the results of the analysis (significant increase/decrease)

percentChange

public static double percentChange(double valInit,
                   double valFinal)

Calculates the percent change from initial to final.

Parameters:

valInit - initial value

valFinal - final value

Returns:

percent change (as decimal)

performanceChange

public static String performanceChange(double valInit,
                       double valFinal)

Determines the significance of the change from the initial value to the final value.

Parameters:

valInit - initial value

valFinal - final value

Returns:

message about the change in performance