def BRAINSCutCreateVector(configurationFilename,
probabilityMapDict,
normalization,
outputXmlFilename,
outputVectorFilename):
print(BRAINSCutCreateVector)
print(BRAINSCutCreateVector)
print(BRAINSCutCreateVector)
print(BRAINSCutCreateVector)
print(BRAINSCutCreateVector)
print(BRAINSCutCreateVector)
import os
import sys
for roi in probabilityMapDict.iterkeys():
if not os.path.exists(probabilityMapDict[roi]):
print( """ ERROR
{fn} does not exist.
""".format( fn=probabilityMapDict[roi]) )
sys.exit()
vectorDir = os.path.dirname(os.path.abspath(outputVectorFilename))
if not os.path.exists(vectorDir):
os.mkdirs(vectorDir)
generateProbabilityMap = False
createVectors = True
trainModel = False
applyModel = False
dummyMethodParameter = {}
dummyOutputDirDict = {}
dummyModelFilename = "na"
probabilityMapGaussianSigma = 1
from ConfigurationParser import BRAINSCutCMDFromConfigFile
returnList = BRAINSCutCMDFromConfigFile(configurationFilename,
outputXmlFilename,
probabilityMapDict,
probabilityMapGaussianSigma,
outputVectorFilename,
dummyModelFilename,
generateProbabilityMap,
createVectors,
normalization,
trainModel,
applyModel,
dummyOutputDirDict,
dummyMethodParameter)
outputVectorFilename = returnList['inputVectorFilename']
outputVectorHdrFilename = outputVectorFilename + ".hdr"
outputXmlFilename = os.path.abspath(outputXmlFilename)
outputNormalization = normalization
print("""Output of BRAINSCutCreateVector
outputVectorFilename = {ovf}
outputVectorHdrFilename = {ovh}
outputNormalization = {on}
outputXmlFilename = {oxf}
""".format( ovf=outputVectorFilename,
ovh=outputVectorHdrFilename,
on=outputNormalization,
oxf=outputXmlFilename))
return outputVectorFilename, outputVectorHdrFilename, outputNormalization, outputXmlFilename
def BRAINSCutTrainModel(
configurationFilename,
inputVectorFilename,
outputModelFilenamePrefix,
outputXmlFilename,
methodParameter,
):
import os.path
import sys
generateProbabilityMap = False
createVectors = False
trainModel = True
applyModel = False
dummyProbMapDict = {}
p_inputVectorFilename = os.path.abspath(inputVectorFilename)
p_inputVectorFilename = p_inputVectorFilename[:-3] # truncate ANN
print(p_inputVectorFilename)
print(p_inputVectorFilename)
print(p_inputVectorFilename)
print(p_inputVectorFilename)
print(p_inputVectorFilename)
print(p_inputVectorFilename)
print(p_inputVectorFilename)
p_outputModelFilenamePrefix = os.path.abspath(outputModelFilenamePrefix)
dummyOutputDirDict = {}
createvectorsNormalization = "Dummy"
probabilityMapGaussianSigma = 1
from ConfigurationParser import BRAINSCutCMDFromConfigFile
returnList = BRAINSCutCMDFromConfigFile(
configurationFilename,
outputXmlFilename,
dummyProbMapDict,
probabilityMapGaussianSigma,
p_inputVectorFilename,
p_outputModelFilenamePrefix,
generateProbabilityMap,
createVectors,
createvectorsNormalization,
trainModel,
applyModel,
dummyOutputDirDict,
methodParameter,
)
outputModelFileSearchStr = (p_outputModelFilenamePrefix + "*" +
str(methodParameter["--numberOfTrees"]) + "*" +
"*gz")
import glob
trainedModelFile = glob.glob(outputModelFileSearchStr)
return os.path.abspath(trainedModelFile[0]), methodParameter
def BRAINSCutGenerateProbabilityMap(configurationFilename, probabilityMapDict,
gaussianSigma, outputXmlFilename):
print("""*****************************
BRAINSCutGenerateProbabilityMap
""")
generateProbabilityMap = True
createVectors = False
trainModel = False
applyModel = False
dummyMethodParameter = {}
print(("""generate probability map
{str}
""".format(str=probabilityMapDict)))
import os
for roi in list(probabilityMapDict.keys()):
print((os.path.abspath(probabilityMapDict[roi])))
probDir = os.path.dirname(os.path.abspath(probabilityMapDict[roi]))
if not os.path.exists(probDir):
os.mkdirs(probDir)
dummyFilename = "na"
dummyDict = {}
createVectorsNormalization = "dummy"
from ConfigurationParser import BRAINSCutCMDFromConfigFile
returnList = BRAINSCutCMDFromConfigFile(
configurationFilename,
outputXmlFilename,
probabilityMapDict,
gaussianSigma,
dummyFilename,
dummyFilename,
generateProbabilityMap,
createVectors,
createVectorsNormalization,
trainModel,
applyModel,
dummyDict,
dummyMethodParameter,
)
returnProbMapList = returnList["probabilityMap"]
import sys
if list(returnProbMapList.keys()) != list(probabilityMapDict.keys()):
print("""ERROR
returnProbMapList has to match probabilityMapDict
in BRAINSCutGenerateProbabilityMap
""")
sys.exit()
outputXmlFilename = os.path.abspath(outputXmlFilename)
return returnProbMapList, outputXmlFilename, configurationFilename
def BRAINSCutApplyModel(
configurationFilename,
probabilityMapDict,
normalization,
inputModelFilename,
methodParameter,
outputXmlFilename,
):
import os.path
import sys
if isinstance(normalization, list):
if len(set(normalization)) == 1:
normalization = normalization[0]
else:
print(("""ERROR::
Improper normalization input.
Currently only one normalization type for
all ROIs are supported.
Normalization input is not same for all ROIs.
{fn}""".format(fn=normalization)))
sys.exit()
p_probMapDict = {}
for roi in list(probabilityMapDict.keys()):
if not os.path.exists(probabilityMapDict[roi]):
print(""" ERROR
probabilityMapDict[ roi ] does not exist.
""")
sys.exit()
p_probMapDict[roi] = os.path.abspath(probabilityMapDict[roi])
import ConfigurationParser as this
outputDirDict = this.getOutputDirDict(configurationFilename)
generateProbabilityMap = False
createVectors = False
trainModel = False
applyModel = True
dummyInputVectorFilename = ""
probabilityMapGaussianSigma = 1
from ConfigurationParser import BRAINSCutCMDFromConfigFile
returnList = BRAINSCutCMDFromConfigFile(
configurationFilename,
outputXmlFilename,
p_probMapDict,
probabilityMapGaussianSigma,
dummyInputVectorFilename,
inputModelFilename,
generateProbabilityMap,
createVectors,
normalization,
trainModel,
applyModel,
outputDirDict,
methodParameter,
)
import glob
outputLabelDict = {}
for session in list(returnList["outputDirDict"].keys()):
outputDir = returnList["outputDirDict"][session]
print(outputDir)
print(outputDir)
print(outputDir)
print(outputDir)
print(outputDir)
print(outputDir)
outputSessionDict = {}
outputSessionDict["label"] = glob.glob(outputDir +
"/*_ANNLabel_seg.nii.gz")
outputSessionDict["ambigiousLabel"] = glob.glob(
outputDir + "/*AmbiguousMap.nii.gz")
outputSessionDict["defPrior"] = glob.glob(outputDir + "/*def.nii.gz")
outputSessionDict["rawOutput"] = glob.glob(outputDir +
"/ANNContinuous*.nii.gz")
outputLabelDict[session] = outputSessionDict
returnList["outputLabelDict"] = outputLabelDict
outputLabelDict = returnList["outputLabelDict"]
return outputLabelDict