예제 #1
0
    def withKnowledge(self, filename):
        if not os.path.exists(filename):
            raise AttributeError("the file '%s' does not exist" % filename)

        # read knowledge file
        jsonObject = ASGCKnowledgeFormatter().deserializeFromFile(filename)
        self._simlearner.knowledge = ASGCKnowledge.fromJson(jsonObject)
        self._simlearner.iteration = self._simlearner.knowledge.getIteration()
        return self
예제 #2
0
    def __init__(self):
        # interacting objects
        self.uqSetting = None
        self.knowledge = ASGCKnowledge()
        self.sampler = None
        self.learner = None

        self.stats = ASGCStatistics()

        # test set
        self.testSet = None
        self.learnWithTest = False

        self.dataContainer = {}

        # simulation based parameters
        self.__params = None
        self._qoi = '_'
        self.__knowledgeTypes = [KnowledgeTypes.SIMPLE]
        self.__timeStepsOfInterest = [0]

        self.verbose = True
예제 #3
0
    def __init__(self):
        super(SimulationLearner, self).__init__()
        # specification
        self.specification = SimulationLearnerSpecification()
        self.knowledge = ASGCKnowledge()

        # statistics per knowledge type
        self.trainingOverall = {}
        self.trainAccuracy = {}
        self.trainCount = {}
        self.testAccuracy = {}
        self.testingOverall = {}
        self.testCount = {}

        self.numberPoints = {}
        self.level = {}

        # initialize data container
        self.dataContainer = {}

        # there is one learner per time step
        self._learners = {}

        self._verbose = True
예제 #4
0
    def __init__(self):
        super(SimulationLearner, self).__init__()
        # specification
        self.specification = SimulationLearnerSpecification()
        self.knowledge = ASGCKnowledge()

        # statistics per knowledge type
        self.trainingOverall = {}
        self.trainAccuracy = {}
        self.trainCount = {}
        self.testAccuracy = {}
        self.testingOverall = {}
        self.testCount = {}

        self.numberPoints = {}
        self.level = {}

        # initialize data container
        self.dataContainer = {}

        # there is one learner per time step
        self._learners = {}

        self._verbose = True
예제 #5
0
class SimulationLearner(Learner):
    def __init__(self):
        super(SimulationLearner, self).__init__()
        # specification
        self.specification = SimulationLearnerSpecification()
        self.knowledge = ASGCKnowledge()

        # statistics per knowledge type
        self.trainingOverall = {}
        self.trainAccuracy = {}
        self.trainCount = {}
        self.testAccuracy = {}
        self.testingOverall = {}
        self.testCount = {}

        self.numberPoints = {}
        self.level = {}

        # initialize data container
        self.dataContainer = {}

        # there is one learner per time step
        self._learners = {}

        self._verbose = True

    # ----------------------------------------------------------------
    def __getattr__(self, attr):
        """
        Overrides built-in method if method called is not a object
        method of this Descriptor, most probably it's a method of
        the learner so it tries to call the method
        from our specification
        @param attr: string method name
        @return: method call in specification
        """
        return getattr(self.specification, attr)

    def addLearner(self, learner, t):
        self._learners[t] = copy.copy(learner)

    def getGrid(self):
        return self.grid

    def getKnowledge(self):
        return self.knowledge

    def getSpecification(self):
        return self.specification

    def setSpecification(self, specification):
        self.specification = specification

    def getLearner(self, t=0):
        return self._learners[t]

    # ----------------------------------------------------------------
    def __prepareDataContainer(self, data, name):
        """
        Prepare data for learning
        @param data: dictionary loaded from UQSetting
        @return dictionary {dtype: {t: <DataContainer>}}
        """
        ans = {}
        U = self.getParameters()\
                .activeParams()\
                .getIndependentJointDistribution()
        for dtype in self.getKnowledgeTypes():
            ans[dtype] = {}
            dim = self.grid.getStorage().getDimension()

            # prepare data container depending on the given knowledge type
            tmp = KnowledgeTypes.transformData(data, U, dtype)

            # load data for all time steps
            for t, values in list(tmp.items()):
                size = len(values)
                mydata = DataMatrix(size, dim)
                sol = DataVector(size)
                for i, (sample, res) in enumerate(values.items()):
                    p = DataVector(sample.getActiveUnit())
                    mydata.setRow(i, p)
                    sol[i] = float(res)
                ans[dtype][t] = DataContainer(points=mydata,
                                              values=sol,
                                              name=name)
        return ans

    # @profile
    def setDataContainer(self, trainUQSetting, testUQSetting=None):
        """
        Sets the training dataContainerDict container given a UQSetting
        @param trainUQSetting: UQSetting
        @param testUQSetting: UQSetting

        WARNING: This method has severe performance issues. It needs
        to be improved such that it loads just the last computed
        chunk of samples.
        """
        # load time steps and quantity of interest
        toi = self.getTimeStepsOfInterest()
        qoi = self.getQoI()

        # lookup for new samples
        ps = []
        for dtype, dataDict in self.dataContainer:
            for t, dataContainer in dataDict:
                for sample in list(trainUQSetting.getSamplesStats().values()):
                    if sample.getActiveUnit() not in dataContainer:
                        ps.append(sample)

        # load the results for the new samples
        if len(self.dataContainer) == 0:
            ps = None

        resultsDict = trainUQSetting.getTimeDependentResults(toi, qoi, ps)

        # prepare the
        dataContainerDict = self.__prepareDataContainer(resultsDict, 'train')
        # set the new dataContainerDict container
        for dtype, values in list(dataContainerDict.items()):
            if dtype not in self.dataContainer:
                self.dataContainer[dtype] = {}

            for t, newDataContainer in list(values.items()):
                if t not in self.dataContainer[dtype]:
                    self.dataContainer[dtype][t] = newDataContainer
                else:
                    self.dataContainer[dtype][t].combine(newDataContainer)

        # if there is a test setting given, combine the train and the
        # test dataContainerDict container
        if testUQSetting is not None:
            dataContainerDict = testUQSetting.getTimeDependentResults(toi, qoi)
            dataContainerDict = self.__prepareDataContainer(
                dataContainerDict, 'test')
            for dtype, values in list(dataContainerDict.items()):
                for t, newDataContainer in list(values.items()):
                    self.dataContainer[dtype][t] = \
                        self.dataContainer[dtype][t].combine(newDataContainer)

    # ----------------------------------------------------------------
    def learnData(self, *args, **kws):
        # learn data
        for dtype, values in list(self.dataContainer.items()):
            knowledge = {}
            print(KnowledgeTypes.toString(dtype))
            # do the learning
            for t, dataContainer in list(values.items()):
                print(("t = %g, " % t, ))
                if dataContainer is not None:
                    learner = self._learners[t]
                    # learn data, if there is any available
                    learner.grid = self.getGrid()
                    learner.dataContainer = dataContainer
                    alpha = learner.learnData()

                    # prepare the answer
                    knowledge[t] = copyGrid(learner.grid), DataVector(alpha)
            print()
            # update results
            if len(knowledge) > 0:
                self.updateResults(knowledge, dtype)

    def learnDataWithTest(self, dataset=None, *args, **kws):
        # learn data
        for dtype, values in list(self.dataContainer.items()):
            knowledge = {}
            # do the learning
            for t, dataContainer in list(values.items()):
                print(("t = %g, " % t, ))
                if dataContainer is not None:
                    learner = self._learners[t]
                    # learn data, if there is any available
                    learner.grid = self.getGrid()
                    learner.dataContainer = dataContainer
                    alpha = learner.learnDataWithTest(dtype=dtype)

                    # prepare the answer
                    knowledge[t] = copyGrid(learner.grid), DataVector(alpha)
            print()
            # update results
            if len(knowledge) > 0:
                self.updateResults(knowledge, dtype)

    def learnDataWithFolding(self, *args, **kws):
        # learn data
        for dtype, values in list(self.dataContainer.items()):
            knowledge = {}

            # do the learning
            for t, dataContainer in list(values.items()):
                if dataContainer is not None:
                    learner = self._learners[t]
                    # learn data, if there is any available
                    learner.grid = self.getGrid()
                    learner.dataContainer = dataContainer
                    alpha = learner.learnDataWithFolding(dtype=dtype)

                    # prepare the answer
                    knowledge[t] = copyGrid(learner.grid), DataVector(alpha)

            # update results
            if len(knowledge) > 0:
                self.updateResults(knowledge, dtype)

    # ----------------------------------------------------------------
    def updateResults(self, knowledge, dtype):
        # run over all learners
        n = len(knowledge)
        for i, (t, (grid, alpha)) in enumerate(knowledge.items()):
            learner = self._learners[t]
            k = -n + i
            # store accuracies of the learning process
            self.trainAccuracy[dtype][t].append(learner.trainAccuracy[k])
            self.trainCount[dtype][t].append(learner.trainCount[k])
            self.trainingOverall[dtype][t].append(
                np.mean(self.trainAccuracy[dtype][t]))

            # check if there are testing results available
            if len(learner.testAccuracy) == len(learner.trainAccuracy):
                self.testAccuracy[dtype][t].append(learner.testAccuracy[k])
                self.testCount[dtype][t].append(learner.testCount[k])
                self.testingOverall[dtype][t].append(
                    np.mean(self.testAccuracy[dtype][t]))

            # update knowledege
            self.knowledge.update(grid, alpha, self.getQoI(), t, dtype,
                                  self.iteration)
        # update other stats
        gs = grid.getStorage()
        self.level[dtype].append(gs.getMaxLevel())
        self.numberPoints[dtype].append(gs.getSize())

    def getL2NormError(self):
        """
        calculate L2-norm of error for all learners
        @return: numpy array of L2 norm errors
        """
        return np.array(
            [learner.getL2NormError() for learner in self._learners])

    def getMaxError(self):
        """
        calculate max error for all learners
        @return: numpy array of max error
        """
        return np.array(
            [learner.getMaxError()() for learner in self._learners])

    def getMinError(self):
        """
        calculate min error for all learners
        @return: numpy array of min error
        """
        return np.array(
            [learner.getMinError()() for learner in self._learners])

    # ----------------------------------------------------------------
    def getCollocationNodes(self):
        """
        Create a set of all collocation nodes
        """
        gs = self.grid.getStorage()
        ps = np.ndarray([gs.getSize(), gs.getDimension()], dtype='float')
        p = DataVector(gs.getDimension())
        for i in range(gs.getSize()):
            gs.getCoordinates(gs.getPoint(i), p)
            ps[i, :] = p.array()

        return ps

    def refineGrid(self):
        # load the time steps we use for refinement
        # refinets = self.getRefinement().getAdaptTimeWindow()
        refinets = self.getTimeStepsOfInterest()
        oldGridSize = self.getGrid().getSize()
        oldAdmissibleSetSize = self.getRefinement().getAdmissibleSet().getSize(
        )

        # refine
        newCollocationNodes = self.getRefinement().refineGrid(self, refinets)

        # increase counter
        self.iteration += 1

        # print some information
        if self._verbose:
            print("iteration: %i" % self.iteration)
            print("old grid size: %i" % oldGridSize)
            print("old AS size: %i" % oldAdmissibleSetSize)
            print("new collocation nodes: %i" % len(newCollocationNodes))
            print(("new grid size:", self.getGrid().getSize()))
            print( "new AS size: %i" % self.getRefinement()\
                                          .getAdmissibleSet()\
                                          .getSize())

#         fig = plotGrid(self.__grid, self.__knowledge.getAlpha(self.getQoI()),
#                        self.getRefinement().getAdmissibleSetCreator()
#                                            .getAdmissibleSet(),
#                        self.getParameters(), newCollocationNodes)
#         fig.savefig('%i.png' % self._learner.iteration)

# parse them to a numpy array
        gs = self.grid.getStorage()
        p = DataVector(gs.getDimension())
        ans = np.ndarray([len(newCollocationNodes),
                          gs.getDimension()],
                         dtype='float')
        for i, gp in enumerate(newCollocationNodes):
            gs.getCoordinates(gp, p)
            ans[i, :] = p.array()

        return ans
예제 #6
0
class SimulationLearner(Learner):

    def __init__(self):
        super(SimulationLearner, self).__init__()
        # specification
        self.specification = SimulationLearnerSpecification()
        self.knowledge = ASGCKnowledge()

        # statistics per knowledge type
        self.trainingOverall = {}
        self.trainAccuracy = {}
        self.trainCount = {}
        self.testAccuracy = {}
        self.testingOverall = {}
        self.testCount = {}

        self.numberPoints = {}
        self.level = {}

        # initialize data container
        self.dataContainer = {}

        # there is one learner per time step
        self._learners = {}

        self._verbose = True

    # ----------------------------------------------------------------
    def __getattr__(self, attr):
        """
        Overrides built-in method if method called is not a object
        method of this Descriptor, most probably it's a method of
        the learner so it tries to call the method
        from our specification
        @param attr: string method name
        @return: method call in specification
        """
        return getattr(self.specification, attr)

    def addLearner(self, learner, t):
        self._learners[t] = copy.copy(learner)

    def getGrid(self):
        return self.grid

    def getKnowledge(self):
        return self.knowledge

    def getSpecification(self):
        return self.specification

    def setSpecification(self, specification):
        self.specification = specification

    def getLearner(self, t=0):
        return self._learners[t]

    # ----------------------------------------------------------------
    def __prepareDataContainer(self, data, name):
        """
        Prepare data for learning
        @param data: dictionary loaded from UQSetting
        @return dictionary {dtype: {t: <DataContainer>}}
        """
        ans = {}
        U = self.getParameters()\
                .activeParams()\
                .getIndependentJointDistribution()
        for dtype in self.getKnowledgeTypes():
            ans[dtype] = {}
            dim = self.grid.getStorage().dim()

            # prepare data container depending on the given knowledge type
            tmp = KnowledgeTypes.transformData(data, U, dtype)

            # load data for all time steps
            for t, values in tmp.items():
                size = len(values)
                mydata = DataMatrix(size, dim)
                sol = DataVector(size)
                for i, (sample, res) in enumerate(values.items()):
                    p = DataVector(sample.getActiveUnit())
                    mydata.setRow(i, p)
                    sol[i] = float(res)
                ans[dtype][t] = DataContainer(points=mydata, values=sol, name=name)
        return ans

    # @profile
    def setDataContainer(self, trainUQSetting, testUQSetting=None):
        """
        Sets the training dataContainerDict container given a UQSetting
        @param trainUQSetting: UQSetting
        @param testUQSetting: UQSetting

        WARNING: This method has severe performance issues. It needs
        to be improved such that it loads just the last computed
        chunk of samples.
        """
        # load time steps and quantity of interest
        toi = self.getTimeStepsOfInterest()
        qoi = self.getQoI()

        # lookup for new samples
        ps = []
        for dtype, dataDict in self.dataContainer:
            for t, dataContainer in dataDict:
                for sample in trainUQSetting.getSamplesStats().values():
                    if sample.getActiveUnit() not in dataContainer:
                        ps.append(sample)

        # load the results for the new samples
        if len(self.dataContainer) == 0:
            ps = None

        resultsDict = trainUQSetting.getTimeDependentResults(toi, qoi, ps)

        # prepare the
        dataContainerDict = self.__prepareDataContainer(resultsDict, 'train')
        # set the new dataContainerDict container
        for dtype, values in dataContainerDict.items():
            if dtype not in self.dataContainer:
                self.dataContainer[dtype] = {}

            for t, newDataContainer in values.items():
                if t not in self.dataContainer[dtype]:
                    self.dataContainer[dtype][t] = newDataContainer
                else:
                    self.dataContainer[dtype][t].combine(newDataContainer)

        # if there is a test setting given, combine the train and the
        # test dataContainerDict container
        if testUQSetting is not None:
            dataContainerDict = testUQSetting.getTimeDependentResults(toi, qoi)
            dataContainerDict = self.__prepareDataContainer(dataContainerDict, 'test')
            for dtype, values in dataContainerDict.items():
                for t, newDataContainer in values.items():
                    self.dataContainer[dtype][t] = \
                        self.dataContainer[dtype][t].combine(newDataContainer)

    # ----------------------------------------------------------------
    def learnData(self, *args, **kws):
        # learn data
        for dtype, values in self.dataContainer.items():
            knowledge = {}
            print KnowledgeTypes.toString(dtype)
            # do the learning
            for t, dataContainer in values.items():
                print "t = %g, " % t,
                if dataContainer is not None:
                    learner = self._learners[t]
                    # learn data, if there is any available
                    learner.grid = self.getGrid()
                    learner.dataContainer = dataContainer
                    alpha = learner.learnData()

                    # prepare the answer
                    knowledge[t] = copyGrid(learner.grid), DataVector(alpha)
            print
            # update results
            if len(knowledge) > 0:
                self.updateResults(knowledge, dtype)

    def learnDataWithTest(self, dataset=None, *args, **kws):
        # learn data
        for dtype, values in self.dataContainer.items():
            knowledge = {}
            # do the learning
            for t, dataContainer in values.items():
                print "t = %g, " % t,
                if dataContainer is not None:
                    learner = self._learners[t]
                    # learn data, if there is any available
                    learner.grid = self.getGrid()
                    learner.dataContainer = dataContainer
                    alpha = learner.learnDataWithTest(dtype=dtype)

                    # prepare the answer
                    knowledge[t] = copyGrid(learner.grid), DataVector(alpha)
            print
            # update results
            if len(knowledge) > 0:
                self.updateResults(knowledge, dtype)

    def learnDataWithFolding(self, *args, **kws):
        # learn data
        for dtype, values in self.dataContainer.items():
            knowledge = {}

            # do the learning
            for t, dataContainer in values.items():
                if dataContainer is not None:
                    learner = self._learners[t]
                    # learn data, if there is any available
                    learner.grid = self.getGrid()
                    learner.dataContainer = dataContainer
                    alpha = learner.learnDataWithFolding(dtype=dtype)

                    # prepare the answer
                    knowledge[t] = copyGrid(learner.grid), DataVector(alpha)

            # update results
            if len(knowledge) > 0:
                self.updateResults(knowledge, dtype)

    # ----------------------------------------------------------------
    def updateResults(self, knowledge, dtype):
        # run over all learners
        n = len(knowledge)
        for i, (t, (grid, alpha)) in enumerate(knowledge.items()):
            learner = self._learners[t]
            k = -n + i
            # store accuracies of the learning process
            self.trainAccuracy[dtype][t].append(learner.trainAccuracy[k])
            self.trainCount[dtype][t].append(learner.trainCount[k])
            self.trainingOverall[dtype][t].append(np.mean(self.trainAccuracy[dtype][t]))

            # check if there are testing results available
            if len(learner.testAccuracy) == len(learner.trainAccuracy):
                self.testAccuracy[dtype][t].append(learner.testAccuracy[k])
                self.testCount[dtype][t].append(learner.testCount[k])
                self.testingOverall[dtype][t].append(np.mean(self.testAccuracy[dtype][t]))

            # update knowledege
            self.knowledge.update(grid, alpha, self.getQoI(),
                                  t, dtype, self.iteration)
        # update other stats
        gs = grid.getStorage()
        self.level[dtype].append(gs.getMaxLevel())
        self.numberPoints[dtype].append(gs.size())

    def getL2NormError(self):
        """
        calculate L2-norm of error for all learners
        @return: numpy array of L2 norm errors
        """
        return np.array([learner.getL2NormError()
                         for learner in self._learners])

    def getMaxError(self):
        """
        calculate max error for all learners
        @return: numpy array of max error
        """
        return np.array([learner.getMaxError()()
                         for learner in self._learners])

    def getMinError(self):
        """
        calculate min error for all learners
        @return: numpy array of min error
        """
        return np.array([learner.getMinError()()
                         for learner in self._learners])

    # ----------------------------------------------------------------
    def getCollocationNodes(self):
        """
        Create a set of all collocation nodes
        """
        gs = self.grid.getStorage()
        ps = np.ndarray([gs.size(), gs.dim()], dtype='float32')
        p = DataVector(gs.dim())
        for i in xrange(gs.size()):
            gs.get(i).getCoords(p)
            ps[i, :] = p.array()

        return ps

    def refineGrid(self):
        # load the time steps we use for refinement
        # refinets = self.getRefinement().getAdaptTimeWindow()
        refinets = self.getTimeStepsOfInterest()
        oldGridSize = self.getGrid().getSize()
        oldAdmissibleSetSize = self.getRefinement().getAdmissibleSet().getSize()

        # refine
        newCollocationNodes = self.getRefinement().refineGrid(self, refinets)

        # increase counter
        self.iteration += 1

        # print some information
        if self._verbose:
            print "iteration: %i" % self.iteration
            print "old grid size: %i" % oldGridSize
            print "old AS size: %i" % oldAdmissibleSetSize
            print "new collocation nodes: %i" % len(newCollocationNodes)
            print "new grid size:", self.getGrid().getSize()
            print "new AS size: %i" % self.getRefinement()\
                                          .getAdmissibleSet()\
                                          .getSize()

#         fig = plotGrid(self.__grid, self.__knowledge.getAlpha(self.getQoI()),
#                        self.getRefinement().getAdmissibleSetCreator()
#                                            .getAdmissibleSet(),
#                        self.getParameters(), newCollocationNodes)
#         fig.savefig('%i.png' % self._learner.iteration)

        # parse them to a numpy array
        gs = self.grid.getStorage()
        p = DataVector(gs.dim())
        ans = np.ndarray([len(newCollocationNodes), gs.dim()], dtype='float32')
        for i, gp in enumerate(newCollocationNodes):
            gp.getCoords(p)
            ans[i, :] = p.array()

        return ans
예제 #7
0
class ASGCUQManager(object):
    def __init__(self):
        # interacting objects
        self.uqSetting = None
        self.knowledge = ASGCKnowledge()
        self.sampler = None
        self.learner = None

        self.stats = ASGCStatistics()

        # test set
        self.testSet = None
        self.learnWithTest = False

        self.dataContainer = {}

        # simulation based parameters
        self.__params = None
        self._qoi = '_'
        self.__knowledgeTypes = [KnowledgeTypes.SIMPLE]
        self.__timeStepsOfInterest = [0]

        self.verbose = True

    def hasMoreSamples(self):
        return self.sampler.hasMoreSamples()

    def runNextSamples(self):
        samples = self.sampler.nextSamples(self.knowledge, self._qoi,
                                           self.__timeStepsOfInterest)
        if len(samples) > 0:
            self.uqSetting.runSamples(samples)
        self.learnData()

    # ----------------------------------------------------------------
    def __prepareDataContainer(self, data, name):
        """
        Prepare data for learning
        @param data: dictionary loaded from UQSetting
        @return dictionary {dtype: {t: <DataContainer>}}
        """
        ans = {}
        U = self.__params.activeParams().getIndependentJointDistribution()
        for dtype in self.getKnowledgeTypes():
            ans[dtype] = {}
            dim = self.sampler.getGrid().getStorage().getDimension()

            # prepare data container depending on the given knowledge type
            tmp = KnowledgeTypes.transformData(data, U, dtype)

            # load data for all time steps
            for i, (t, values) in enumerate(tmp.items()):
                size = len(values)
                mydata = np.ndarray((size, dim))
                sol = np.ndarray(size)
                for i, (sample, res) in enumerate(values.items()):
                    mydata[i, :] = np.array(sample.getActiveUnit())
                    sol[i] = float(res)
                ans[dtype][t] = DataContainer(points=mydata,
                                              values=sol,
                                              name=name)
        return ans

    # @profile
    def updateDataContainer(self, updateTestData=False):
        """
        Sets the training dataContainerDict container given a UQSetting

        WARNING: This method has severe performance issues. It needs
        to be improved such that it loads just the last computed
        chunk of samples.
        """
        # load the results for the new samples
        if len(self.dataContainer) == 0:
            ps = None
        else:
            # load time steps and quantity of interest
            # lookup for new samples
            ps = []
            dataContainer = next(
                iter(self.dataContainer.values()).next().itervalues())
            for sample in list(self.uqSetting.getSamplesStats().values()):
                if sample.getActiveUnit() not in dataContainer:
                    ps.append(sample)
#             print( dataContainer.getSizeTrain(), " = ", self.uqSetting.getSize(), "-", len(ps) )
#             assert dataContainer.getSizeTrain() == self.uqSetting.getSize() - len(ps)

        resultsDict = self.uqSetting.getTimeDependentResults(
            self.__timeStepsOfInterest, self._qoi, ps)
        # prepare the results as dictionary
        dataContainerDict = self.__prepareDataContainer(resultsDict, 'train')
        # set the new dataContainerDict container
        for dtype, values in list(dataContainerDict.items()):
            if dtype not in self.dataContainer:
                self.dataContainer[dtype] = {}

            for t, newDataContainer in list(values.items()):
                if t not in self.dataContainer[dtype]:
                    self.dataContainer[dtype][t] = newDataContainer
                else:
                    if newDataContainer.getSize() > 0:
                        self.dataContainer[dtype][t] = \
                            self.dataContainer[dtype][t].combine(newDataContainer)

        # if there is a test setting given, combine the train and the
        # test dataContainerDict container.
        # the test set does not change over time so update it just at the beginning
        if updateTestData and self.testSet is not None:
            dataContainerDict = self.testSet.getTimeDependentResults(
                self.__timeStepsOfInterest, self._qoi)
            dataContainerDict = self.__prepareDataContainer(
                dataContainerDict, 'test')
            for dtype, values in list(dataContainerDict.items()):
                for t, newDataContainer in list(values.items()):
                    if newDataContainer.getSize() > 0:
                        self.dataContainer[dtype][t] = \
                            self.dataContainer[dtype][t].combine(newDataContainer)

    def learnData(self):
        """
        Learn the available data
        """
        # learn the data
        self.updateDataContainer(
            updateTestData=self.sampler.getCurrentIterationNumber() == 1)
        if self.learnWithTest:
            self.learnDataWithTest()
        else:
            self.learnDataWithoutTest()

    # ----------------------------------------------------------------
    def learnDataWithoutTest(self, *args, **kws):
        # learn data
        if self.verbose:
            print("learning (i=%i, gs=%i, type=%s)" %
                  (self.sampler.getCurrentIterationNumber(),
                   self.sampler.getGrid().getSize(),
                   self.sampler.getGrid().getTypeAsString()))
        self.learner.grid = self.sampler.getGrid()
        for dtype, values in list(self.dataContainer.items()):
            knowledge = {}
            if self.verbose:
                print(KnowledgeTypes.toString(dtype))
            # do the learning
            for t, dataContainer in list(values.items()):
                if self.verbose:
                    print(("t = %g, " % t, ))
                sys.stdout.flush()
                if dataContainer is not None:
                    # learn data, if there is any available
                    self.learner.dataContainer = dataContainer
                    self.learner.learnData()

                    # update the knowledge
                    self.knowledge.update(
                        copyGrid(self.learner.grid), self.learner.alpha,
                        self._qoi, t, dtype,
                        self.sampler.getCurrentIterationNumber() - 1)

                    # update results
                    self.stats.updateResults(dtype, t, self.learner)

            if self.verbose:
                print()

    def learnDataWithTest(self, dataset=None, *args, **kws):
        if self.verbose:
            print("learning with test (i=%i, gs=%i)" %
                  (self.sampler.getCurrentIterationNumber(),
                   self.sampler.getGrid().getSize()))
        # learn data
        self.learner.grid = self.sampler.getGrid()
        for dtype, values in list(self.dataContainer.items()):
            knowledge = {}
            # do the learning
            for t in np.sort(list(values.keys())):
                dataContainer = values[t]
                if self.verbose:
                    print(("t = %g, " % t, ))
                sys.stdout.flush()
                if dataContainer is not None:
                    # learn data, if there is any available
                    self.learner.dataContainer = dataContainer
                    self.learner.learnDataWithTest(dtype=dtype)

                    # update the knowledge
                    self.knowledge.update(
                        copyGrid(self.learner.grid), self.learner.alpha,
                        self._qoi, t, dtype,
                        self.sampler.getCurrentIterationNumber() - 1)

                    # update results
                    self.stats.updateResults(dtype, t, self.learner)

            if self.verbose:
                print()

    def recomputeStats(self):
        if len(self.dataContainer) == 0:
            self.updateDataContainer(True)

        self.knowledge.clearAlphas()

        for dtype, values in list(self.dataContainer.items()):
            knowledge = {}
            # do the learning
            for t, dataContainer in list(values.items()):
                if dataContainer is not None:
                    for iteration in self.knowledge.getAvailableIterations():
                        grid = self.knowledge.getGrid(self._qoi, iteration)
                        self.learner.grid = grid
                        trainSubset = dataContainer.getTrainDataset()
                        testSubset = None
                        if self.learnWithTest:
                            testSubset = dataContainer.getTestDataset()

                        # -----------------------------------------------------
                        # do the learning
                        if self.verbose:
                            print(("learning: t = %g," % t, ))
                        # learn alpha with corresponding grid
                        self.learner.grid = grid

                        if self.learnWithTest:
                            if self.verbose:
                                print("with test (i=%i, gs=%i)" %
                                      (iteration, grid.getSize()))
                            self.learner.learnDataWithTest(dataContainer,
                                                           dtype=dtype)
                        else:
                            if self.verbose:
                                print("(i=%i, gs=%i)" %
                                      (iteration, grid.getSize()))
                            self.learner.learnDataWithoutTest(dataContainer,
                                                              dtype=dtype)

                        # -----------------------------------------------------
                        # update the knowledge
                        self.knowledge.update(grid, self.learner.alpha,
                                              self._qoi, t, dtype, iteration)

                        # update stats -> copy
                        self.stats.updateResults(dtype, t, self.learner)

            if self.verbose:
                print()

    def getParameters(self):
        return self.__params

    def setParameters(self, value):
        self.__params = value

    def getRefinement(self):
        return self._refinement

    def getTestSet(self):
        return self.testSet

    def setTestSet(self, value):
        self.testSet = value

    def setLearnWithTest(self, value):
        self.learnWithTest = value

    def getKnowledgeTypes(self):
        return self.__knowledgeTypes

    def setKnowledgeTypes(self, value):
        self.__knowledgeTypes = value

    def getQoI(self):
        return self._qoi

    def getTimeStepsOfInterest(self):
        return self.__timeStepsOfInterest

    def getRefinement(self):
        return self.refinementManager

    def setRefinement(self, value):
        self.refinementManager = value

    def setQoI(self, value):
        self._qoi = value

    def setTimeStepsOfInterest(self, value):
        self.__timeStepsOfInterest = value

    def getDim(self):
        return self.__params.getDim()

    def getGrid(self):
        return self.sampler.getGrid()

    def setSampler(self, sampler):
        self.sampler = sampler

    def getKnowledge(self):
        return self.knowledge