Collection of generally useful functions. More...

Detailed Description

Collection of generally useful functions.

Author: Daniel Wirtz

Date: 11.10.2010

Change in 0.6:: (Daniel Wirtz, 2012-06-08) Bugfix: Now the logarithmic norms are computed correctly. ARGH that took a while to detect..

New in 0.6:: (Daniel Wirtz, 2012-04-13) New method "getTube" that allows to draw random vectors from spaces of arbitrary dimension but restricted to a specified tube.

Change in 0.6:: (Daniel Wirtz, 2011-11-17) Moved the getObjectConfig to a separate file.

Change in 0.6:: (Daniel Wirtz, 2011-11-16) Using mex CalcMD5 now to compute hash values for vectors. Source downloaded from http://www.mathworks.com/matlabcentral/fileexchange/25921. Also updated the KerMor.setup script to automatically compile the CalcMD5 mex file.

Change in 0.5:

(Daniel Wirtz, 2011-09-15)

saveAxes and saveFigure now store the last save location in the preferences and reuse them.
removeMargin now properly works, together with saveFigure or saveAxes.

New in 0.5:: (Daniel Wirtz, 2011-07-05) Added the Utils.implode function.

New in 0.3:: (Daniel Wirtz, 2011-04-20) Added a new function Utils.getHelpShort to extract the first line(s) of a help text in matlab style (text until first emtpy line = short)

New in 0.3:: (Daniel Wirtz, 2011-04-18) Added the saveFigure and saveAxes methods from SegMedix.

Change in 0.3:: (Daniel Wirtz, 2011-04-04) Moved the Utils.getObjectConfig method here from models.BaseModel

New in 0.3:

(Daniel Wirtz, 2011-04-01)

Added the Utils.getBoundingBox function.
Added the Utils.findVecInMatrix function.

This class is part of the framework

KerMor - Model Order Reduction using Kernels

Homepage http://www.morepas.org/software/index.html
Documentation http://www.morepas.org/software/kermor/index.html
License KerMor license conditions

Definition at line 17 of file Utils.m.

Static Public Member Functions
static function char short =	getHelpShort (char txt)
	Extracts the help short subtext from a given text. More...

static function [ double bmin , double bmax ] =	getBoundingBox (double vectors)
	Gets the bounding box for a matrix containing column vectors. More...

static function comb =	createCombinations (ranges, varargin)
	Creates the cartesian product of the vectors passed as a matrix containing elements of each vector per row. More...

static function struct target =	copyStructFields (struct source,struct target)
	Recursively copies struct fields from one struct to another. More...

static function str =	getLatexStr (value, digits)

static function char str =	implode (char\|rowvec data,char glue,char format)
	Implodes the elements of data using glue. More...

static function idx =	findVecInMatrix (A, b)
	Finds column vectors inside a matrix. More...

static function y =	preparePlainPlot (y)
	Memory-saving plotting for plain result plots. More...

static function	saveFigure (fig, filename, ext)
	Opens a matlab save dialog and saves the given figure to the file selected. More...

static function	saveAxes (handle ax, varargin)
	Convenience function. Allows to save a custom axes instead of a whole figure which allows to drop any unwanted uiobjects contained on the source figure. More...

static function	removeMargin (f)
	Requires the axes and figure units to be the same. More...

static function h =	getHash (vec)
	Returns a hash code for the given vector. More...

static function [ matrix < double > x , colvec< double > farend ] =	getTube (integer dim,integer num,double length,double spread,integer seed)
	Computes \(n=\)`num` random vectors inside a tube of length \(l=\)`length` inside a \(d=\)`dim` dimensional space. The tube starts at zero and ends at a random point \(x_e\in\R^d\) of length \(l\). The spread \(s=\)`spread` determines the maximum distance \(ls\) by which the random vectors might be away from the tube center \(\{x\in\R^d~\|~ x = \alpha x_e,\alpha\in[0,1]\}\). More...

static function [ double ln , colvec < double > v ] =	logNorm (matrix< double > A,matrix< double > G,colvec< double > v0)
	Computes the logarithmic norm of a matrix \(\vA\), optionally using a positive definite symmetric matrix \(\vG\) inducing the matrix norm to use. More...

static function S =	sprand (integer n,integer m,double dens,RandStream rs)
	Creates a random sparse matrix with given density (approximately). More...

static function	closeAllFigures ()

static function folder =	getDir (caption, initial_path)
	Prompts the user to select a directory. More...

static function logical success =	ensureDir (char dir)
	Ensures that a directory exists. More...

static function	copyPrefGroup (char from,char to)
	Copies the preferences from one group to another. More...

static function entries =	getLegendFromVector (values, format_string)

static function res =	test_Tube ()

static function res =	test_createCombinations ()
	Tests the createCombinations function. More...

static function res =	test_findVec ()
	Tests the findVecInMatrix function. More...

Member Function Documentation

static function Utils.closeAllFigures ( )

static

Definition at line 658 of file Utils.m.

function Utils.copyPrefGroup	(	char	from,
		char	to
	)

static

Copies the preferences from one group to another.

Parameters

from	The preference group to copy from
to	The preference group to copy to

Definition at line 730 of file Utils.m.

References k.

function struct target = Utils.copyStructFields	(	struct	source,
		struct	target
	)

static

Recursively copies struct fields from one struct to another.

Effectively implements a struct.clone() method.

Author: Daniel Wirtz

Date: 2010-11-03

Parameters

source	The source struct.
target	The target struct.

Definition at line 159 of file Utils.m.

References copyStructFields().

Referenced by copyStructFields().

Here is the call graph for this function:

Here is the caller graph for this function:

function comb = Utils.createCombinations	(	ranges,
		varargin
	)

static

Creates the cartesian product of the vectors passed as a matrix containing elements of each vector per row.

Author: Daniel Wirtz

Date: 2010-10-11

Parameters

ranges	Can either be a cell array of vectors or a vector.
varargin	If the first argument is a vector, an arbitrary number of additional vectors can be passed to build the cartesian product from.

Return values

comb	A matrix containing the combinations, each row corresponds to an input vector's range.

Definition at line 114 of file Utils.m.

Referenced by demos.Basics5_KernelsAndApprox(), models.motoneuron.experiments.FrequencyLearning(), models.BaseFullModel.off2_genTrainingData(), sampling.WeightedRandomSampler.performSampling(), sampling.GridSampler.performSampling(), test_createCombinations(), models.muscle.functions.MarkertLaw.test_FitToOriginal(), models.muscle.functions.MarkertLaw.test_Linarization(), and approx.algorithms.VKOGA.test_VKOGA2D1D().

Here is the caller graph for this function:

function logical success = Utils.ensureDir ( char dir )

static

Ensures that a directory exists.

If no return argument is wanted, an exception is thrown when creation of a nonexistent directory fails.

Parameters

dir	The target directory

Return values

success True if the directory exists or has been created, false otherwise

Definition at line 700 of file Utils.m.

Referenced by kernels.KernelExpansion.exportToDirectory().

Here is the caller graph for this function:

function idx = Utils.findVecInMatrix	(	A,
		b
	)

static

Finds column vectors inside a matrix.

For multiple occurences, the first found index is used.

See http://www.mathworks.com/matlabcentral/newsreader/view_thread/174277 and the test function testing.find_vec_in_matrix_speedtest for further information.

Change in 0.3:

(Daniel Wirtz, 2011-04-12) Added support for multi vector search.

(Daniel Wirtz, 2011-04-13) Fixed errors when multiple occurences appear.

Parameters

A	A \(n\times m\) matrix of \(m\) column vectors
b	A \(n\times p\) vector, where each column is regarded as one vector to search

Return values

idx	A \(1 \times p\) vector containing the first found positions indices if a vector from b is contained in A, zero otherwise.

Definition at line 259 of file Utils.m.

References t.

Referenced by fem.geometry.BaseGeometry.getCommonNodesWith(), spacereduction.BaseSpaceReducer.getInitialSpace(), demos.ResponseSurfaceApprox.iterationPlots(), ApproxVisualizer>plotCurrent(), FunVis2D>updateCenterPoints(), fem.geometry.RegularHex8Grid(), test_findVec(), and fem.geometry.BaseGeometry.toCMGUI().

Here is the caller graph for this function:

function [ double bmin , double bmax ] = Utils.getBoundingBox ( double vectors )

static

Gets the bounding box for a matrix containing column vectors.

Parameters

vectors A \(n\times m\) matrix containing \(m\) column vectors

Return values

bmin	A \(n\times 1\) vector representing the minimum value corner of the bounding box
bmax	A \(n\times 1\) vector representing the maximum value corner of the bounding box

Definition at line 96 of file Utils.m.

Referenced by data.ApproxTrainData.ApproxTrainData(), data.MemoryTrajectoryData.getBoundingBox(), FunVis2D>FunVis2D_OpeningFcn(), FunVis2D>rangesFromATD(), error.IterationCompLemmaEstimator.offlineComputations(), and data.FileMatrix.test_FileMatrix().

Here is the caller graph for this function:

function folder = Utils.getDir	(	caption,
		initial_path
	)

static

Prompts the user to select a directory.

If no screen is used, a text input prompt will be used to let the user enter the path manually.

Parameters

caption	The caption for the dialog/prompt
initial_path	The path to start from (only for UI)

Definition at line 669 of file Utils.m.

References all().

Referenced by data.ModelData.relocate().

Here is the call graph for this function:

Here is the caller graph for this function:

function h = Utils.getHash ( vec )

static

Returns a hash code for the given vector.

Currently using the CalcMD5 routine from 3rdparty/calcmd5, which is included in KerMor as 3rd party code. Original source: http://www.mathworks.com/matlabcentral/fileexchange/25921

Definition at line 493 of file Utils.m.

References CalcMD5().

Referenced by data.FileDataCollection.addData(), data.MemoryTrajectoryData.addTrajectory(), data.FileDataCollection.getData(), data.MemoryTrajectoryData.getTrajectory(), and data.FileDataCollection.hasData().

Here is the call graph for this function:

Here is the caller graph for this function:

function char short = Utils.getHelpShort ( char txt )

static

Extracts the help short subtext from a given text.

Gets the first block of a text that goes until the first blank line.

Parameters

txt	The text to use.

Return values

short The subtext.

Definition at line 68 of file Utils.m.

static function str = Utils.getLatexStr	(	value,
		digits
	)

static

Definition at line 194 of file Utils.m.

Referenced by EstimatorAnalyzer.getResultTable().

Here is the caller graph for this function:

static function entries = Utils.getLegendFromVector	(	values,
		format_string
	)

static

Definition at line 747 of file Utils.m.

function [ matrix< double > x , colvec< double > farend ] = Utils.getTube	(	integer	dim,
		integer	num,
		double	length,
		double	spread,
		integer	seed
	)

static

Computes \(n=\)num random vectors inside a tube of length \(l=\)length inside a \(d=\)dim dimensional space. The tube starts at zero and ends at a random point \(x_e\in\R^d\) of length \(l\). The spread \(s=\)spread determines the maximum distance \(ls\) by which the random vectors might be away from the tube center \(\{x\in\R^d~|~ x = \alpha x_e,\alpha\in[0,1]\}\).

Parameters

dim	The dimension of the tube.
num	The number of desired random vectors.
length	The length \(l\) of the tube. Default: 1
spread	The spread \(s\) of the tube vectors around the tube core. Default: 0.15
seed	The seed to use for the random number generator. Default: cputime*10

Return values

x	An \(d\times n\) matrix with random vectors from a random tube.
farend	The vector from the far end away from zero.

Definition at line 510 of file Utils.m.

References Norm.L2().

Referenced by test_Tube().

Here is the call graph for this function:

Here is the caller graph for this function:

function char str = Utils.implode	(	char\|rowvec	data,
		char	glue,
		char	format
	)

static

Implodes the elements of data using glue.

Either transforms a cell array of strings into one string or implodes a numeric vector using the specified format.

New in 0.6:: (Daniel Wirtz, 2011-11-17) Added the possibility to pass a numeric vector plus a format string.

Parameters

data	A cell array of strings/chars or a numeric vector
glue	A string that is inserted between any element string representation Default: `,`
format	The sprintf format string if data is a vector Default: `%2.3e`

Return values

str	The concatented string of all `data` strings glued together with the string `glue`

Definition at line 208 of file Utils.m.

Referenced by IClassConfig.addRange(), ModelAnalyzer.compareRedFull(), Java.exportFunctions(), IClassConfig.getAxisLabels(), kernels.config.ParamTimeExpansionConfig.getConfiguredPropertiesString(), general.functions.FuncSum.getFunction(), general.functions.PiecewiseLinear.getFunction(), general.DEIM.getInterpolationPoints(), fem.AFEMConfig.getOptionStr(), models.pcd.PCDSystem1D.plot(), and MonomialIterator.test_MonomialIterator().

Here is the caller graph for this function:

function [ double ln , colvec< double > v ] = Utils.logNorm	(	matrix< double >	A,
		matrix< double >	G,
		colvec< double >	v0
	)

static

Computes the logarithmic norm of a matrix \(\vA\), optionally using a positive definite symmetric matrix \(\vG\) inducing the matrix norm to use.

Change in 0.6:: (Daniel Wirtz, 2012-07-18) Added a re-try with 6 eigenvalues if no convergence for one eigenvalue takes place. Added an optional v0 parameter as start vector for eigs.

Parameters

A	The target matrix \(\vA\)
G	The positive definite, symmetric matrix \(\vG\) inducing the norm and hence matrix norm to use. Defaults to identity matrix (Euclidean/L2 norm) Default: \(\vI_d\)
v0	An initial vector for eigs to start with. Default: []

Return values

ln	The logarithmic norm with respect to the given matrix norm induced by \(\vG\).
v	The eigenvector of the largest eigenvalue (=log norm) of the symmetric part of \(\vA\).

Definition at line 551 of file Utils.m.

References KerMor.App(), and t.

Referenced by testing.DEIM.compareDEIM_Full_Jacobian(), testing.LogNorm.compareSimTransDEIMJac_FullJac(), testing.LogNorm.compareSimTransJac_FullJac(), testing.LogNorm.getApproxLogNormsAtPos(), error.DEIMEstimator.getBeta(), and error.DEIMEstimator.offlineComputations().

Here is the call graph for this function:

Here is the caller graph for this function:

function y = Utils.preparePlainPlot ( y )

static

Memory-saving plotting for plain result plots.

Parameters

y	A result matrix \(y\) with rows corresponding to single dimensions and columns corresponding to time-steps.

Return values

y	If there are more than 1000 dimensions, the subset with distinct (via unique) last values are extracted. If this still results in more than 4000 plots, the first 4000 dimensions are selected.

Definition at line 297 of file Utils.m.

Referenced by ModelAnalyzer.analyzeError(), models.BaseModel.plot(), and models.BaseModel.plotState().

Here is the caller graph for this function:

function Utils.removeMargin ( f )

static

Requires the axes and figure units to be the same.

Todo:: tailor this method so that subplots are also supported (so far only one axis!) and change saveFigure again.

Definition at line 465 of file Utils.m.

function Utils.saveAxes	(	handle	ax,
			varargin
	)

static

Convenience function. Allows to save a custom axes instead of a whole figure which allows to drop any unwanted uiobjects contained on the source figure.

Parameters