pyDev4G2S

The goal of this G2S extension is to quickly prototype new sequential simulation algorithms.

To install

To install, clone the repository in the build/extension folder of G2S. Then execute make extension in build of G2S.

How to use

To use the code, setup '-a','pyDev', the setup the python code '-pa',pythonAlgo. pythonAlgo is a string structured as 'path/to/the/fileWithoutExtension:className' If 'className' is by default the same name as the directory. And the default path is to the PyDevAlgo directory in the build folder.

Structure of a class

class nameOfTheClass:
	def __init__(self,TIs,kernel,di,dt,path):
		# all parameters are read-only
		# keep simple operation in this section
		#TIs			==> list of training images if available
		#kernel			==> kernel if available
		#di				==> original destination image
		#dt				==> list of data type
		#path			==> the path, if explicitly set at call time
		pass

	def build(self,n):
		# use this section for more complex initialization
		#n				==> the maximum number neighbors as set in the call
		pass

	def sample(self,coordinates,values,randUniform,indexInThePath,threadId, fullStationary, variableOfInterest, idTI4Sampling):
		# return a sample value as a list of multivariate and index value
		# coordinates			==> list of coordinates
		# values				==> list of values
		# randUniform			==> a random value uniform between 0 and 1
		# indexInThePath		==> index in the path
		# threadId				==> thread id for multithreading, to not use currently !
		# fullStationary		==> boolean about stationarity
		# variableOfInterest	==> the variable to simulate in case of full simulation
		# idTI4Sampling			==> index to choose the training image
		return (aSampleAsAList,AnInfromativeIndex);

	def __del__(self):
		# if need to destroy any element
		pass

Examples

Exmaple SGS

[data,t]=g2s('-a','pyDev','-pa','sgs','-di',destination,'-dt',zeros(1,1),'-n',50); imagesc(data);

Exmaple SGS with Constant path

[data,t]=g2s('-a','pyDev','-pa','sgsCP','-di',destination,'-dt',zeros(1,1),'-n',50,'-sp',path); imagesc(data);

Example SGS with Constant path and TI as model

kernel=exp(-0.01*bwdist(padarray(1,[100,100])));
	
[data,t]=g2s('-a','pyDev','-pa','sgsTiCP','-di',destination,'-dt',zeros(1,1),'-n',50,'-sp',path,'-ti',kernel);
imagesc(data);

Example SGS with Constant path and multiple TI as model and non-stationarity

kernels={};
for alpha=2.^(-12:0.5:-5)
	kernels{numel(kernels)+1}=exp(-alpha*bwdist(padarray(1,[100,100])));
end
	
indexImage=floor((0:size(destination,1)-1)/size(destination,1)*numel(kernels)).*ones(size(destination,2),1);
	
[data,t]=g2s('-a','pyDev','-pa','sgsTiCP','-di',destination,'-dt',zeros(1,1),'-n',50,'-sp',path,'-ti',kernels{:},'-ii',indexImage);
imagesc(data);