Dump HistorySet MetaData into the correct output csv files (#987)

* enable historyset to handle both scalar and vector meta data, and add check for NAN from code collections * fix * fix HistorySet * fix reading vector meta data * update tests * add unit test for meta data load, print, reload * fix * modify test to test only the mechanics
idaholab · May 10, 2019 · 9849909 · 9849909
1 parent 69f066f
commit 9849909
Show file tree

Hide file tree

Showing 12 changed files with 498 additions and 100 deletions.
diff --git a/framework/DataObjects/DataSet.py b/framework/DataObjects/DataSet.py
@@ -112,20 +112,21 @@ def addExpectedMeta(self,keys, params={}):
       @ In, keys, set(str), keys to register
       @ In, params, dict, optional, {key:[indexes]}, keys of the dictionary are the variable names,
         values of the dictionary are lists of the corresponding indexes/coordinates of given variable
-      @ Out, None
+      @ Out, keys, list(str), extra keys that has been registered
     """
     # TODO add option to skip parts of meta if user wants to
     # remove already existing keys
     keys = list(key for key in keys if key not in self.getVars()+self.indexes)
     # if no new meta, move along
     if len(keys) == 0:
-      return
+      return keys
     # CANNOT add expected meta after samples are started
     assert(self._data is None)
     assert(self._collector is None or len(self._collector) == 0)
     self._metavars.extend(keys)
     self._orderedVars.extend(keys)
     self.setPivotParams(params)
+    return keys
 
   def addMeta(self, tag, xmlDict = None, node = None):
     """
@@ -972,7 +973,7 @@ def _convertArrayListToDataset(self,array,action='return'):
       # determine dimensions for each variable
       dimsMeta = {}
       for name, var in new.variables.items():
-        if name not in self._inputs + self._outputs:
+        if name not in self._inputs + self._outputs + self._metavars:
           continue
         dims = list(var.dims)
         # don't list if only entry is sampleTag
@@ -1554,8 +1555,12 @@ def _loadCsvMeta(self,fileName):
       dims = meta.get('pivotParams',{})
       if len(dims)>0:
         self.setPivotParams(dims)
+      # vector metavars is also stored in 'DataSet/dims' node
+      metavars = meta.get('metavars',[])
+      # get dict of vector metavars
+      params = {key:val for key, val in dims.items() if key in metavars}
       # add metadata, so we get probability weights and etc
-      self.addExpectedMeta(meta.get('metavars',[]))
+      self.addExpectedMeta(metavars,params)
       # check all variables desired are available
       provided = set(meta.get('inputs',[])+meta.get('outputs',[])+meta.get('metavars',[]))
     # otherwise, if we have no meta XML to load from, infer what we can from the CSV, which is only the available variables.

diff --git a/framework/DataObjects/HistorySet.py b/framework/DataObjects/HistorySet.py
@@ -81,6 +81,8 @@ def __init__(self):
     self.printTag  = self.name
     self._tempPivotParam = None
     self._neededForReload = [] # HistorySet doesn't need anything special to load, since it's written in cluster-by-sample CSV format
+    self._inputMetaVars = [] # meta vars belong to the input of HistorySet, i.e. scalar
+    self._outputMetaVars = [] # meta vara belong to the output of HistorySet, i.e. vector
 
   def _readMoreXML(self,xmlNode):
     """
@@ -126,7 +128,7 @@ def _fromCSV(self,fileName,**kwargs):
     main = self._readPandasCSV(fileName+'.csv')
     nSamples = len(main.index)
     ## collect input space data
-    for inp in self._inputs + self._metavars:
+    for inp in self._inputs + self._inputMetaVars:
       data[inp] = main[inp].values
     ## get the sampleTag values if they're present, in case it's not just range
     if self.sampleTag in main:
@@ -136,7 +138,7 @@ def _fromCSV(self,fileName,**kwargs):
     # load subfiles for output spaces
     subFiles = main['filename'].values
     # pre-build realization spots
-    for out in self._outputs + self.indexes:
+    for out in self._outputs + self.indexes + self._outputMetaVars:
       data[out] = np.zeros(nSamples,dtype=object)
     # read in secondary CSVs
     for i,sub in enumerate(subFiles):
@@ -150,7 +152,7 @@ def _fromCSV(self,fileName,**kwargs):
       if len(set(subDat.keys()).intersection(self.indexes)) != len(self.indexes):
         self.raiseAnError(IOError,'Importing HistorySet from .csv: the pivot parameters "'+', '.join(self.indexes)+'" have not been found in the .csv file. Check that the '
                                   'correct <pivotParameter> has been specified in the dataObject or make sure the <pivotParameter> is included in the .csv files')
-      for out in self._outputs+self.indexes:
+      for out in self._outputs + self.indexes + self._outputMetaVars:
         data[out][i] = subDat[out].values
     # construct final data object
     self.load(data,style='dict',dims=self.getDimensions())
@@ -190,7 +192,7 @@ def _selectiveRealization(self,rlz):
       if not utils.isSingleValued(val):
         # treat inputs, outputs differently TODO this should extend to per-variable someday
         ## inputs
-        if var in self._inputs:
+        if var in self._inputs + self._inputMetaVars:
           method,indic = self._selectInput
         # pivot variables are included here in "else"; remove them after they're used in operators
         else:
@@ -256,7 +258,7 @@ def _toCSV(self,fileName,start=0,**kwargs):
     # specific implementation
     ## write input space CSV with pointers to history CSVs
     ### get list of input variables to keep
-    ordered = list(i for i in itertools.chain(self._inputs,self._metavars) if i in keep)
+    ordered = list(i for i in itertools.chain(self._inputs,self._inputMetaVars) if i in keep)
     ### select input part of dataset
     inpData = data[ordered]
     ### add column for realization information, pointing to the appropriate CSV
@@ -271,7 +273,7 @@ def _toCSV(self,fileName,start=0,**kwargs):
     ### write CSV
     self._usePandasWriteCSV(fileName,inpData,ordered,keepSampleTag = self.sampleTag in keep,mode=mode)
     ## obtain slices to write subset CSVs
-    ordered = list(o for o in self.getVars('output') if o in keep)
+    ordered = list(o for o in itertools.chain(self._outputs,self._outputMetaVars) if o in keep)
 
     if len(ordered):
       # hierarchical flag controls the printing/plotting of the dataobject in case it is an hierarchical one.
@@ -300,3 +302,17 @@ def _toCSV(self,fileName,start=0,**kwargs):
           self._usePandasWriteCSV(filename,rlz,ordered,keepIndex=True)
     else:
       self.raiseAWarning('No output space variables have been requested for DataObject "{}"! No history files will be printed!'.format(self.name))
+
+  def addExpectedMeta(self,keys, params={}):
+    """
+      Registers meta to look for in realizations.
+      @ In, keys, set(str), keys to register
+      @ In, params, dict, optional, {key:[indexes]}, keys of the dictionary are the variable names,
+        values of the dictionary are lists of the corresponding indexes/coordinates of given variable
+      @ Out, None
+    """
+    extraKeys = DataSet.addExpectedMeta(self, keys, params)
+    self._inputMetaVars.extend(list(key for key in extraKeys if key not in params))
+    if params:
+      self._outputMetaVars.extend(list(key for key in extraKeys if key in params))
+    return extraKeys
diff --git a/framework/Models/Code.py b/framework/Models/Code.py
@@ -615,6 +615,8 @@ def evaluateSample(self, myInput, samplerType, kwargs):
 
         csvLoader = CsvLoader.CsvLoader(self.messageHandler)
         csvData = csvLoader.loadCsvFile(outFile)
+        if np.isnan(csvData).all():
+          self.raiseAnError(IOError, 'The data collected from', outputFile+'.csv', 'only contain "NAN"')
         headers = csvLoader.getAllFieldNames()
 
         ## Numpy by default iterates over rows, thus we transpose the data and

diff --git a/framework/PostProcessors/BasicStatistics.py b/framework/PostProcessors/BasicStatistics.py
@@ -254,20 +254,32 @@ def initialize(self, runInfo, inputs, initDict):
     inputObj = inputs[-1] if type(inputs) == list else inputs
     if inputObj.type == 'HistorySet':
       self.dynamic = True
-    metaKeys = []
+    inputMetaKeys = []
+    outputMetaKeys = []
     for metric, infos in self.toDo.items():
       steMetric = metric + '_ste'
       if steMetric in self.steVals:
         for info in infos:
           prefix = info['prefix']
           for target in info['targets']:
             metaVar = prefix + '_ste_' + target if not self.outputDataset else metric + '_ste'
-            metaKeys.append(metaVar)
+            metaDim = inputObj.getDimensions(target)
+            if len(metaDim[target]) == 0:
+              inputMetaKeys.append(metaVar)
+            else:
+              outputMetaKeys.append(metaVar)
+    metaParams = {}
     if not self.outputDataset:
-      metaParams = {key:[self.pivotParameter] for key in metaKeys} if self.dynamic else {}
+      if len(outputMetaKeys) > 0:
+        metaParams = {key:[self.pivotParameter] for key in outputMetaKeys}
     else:
-      metaParams = {key:[self.pivotParameter,self.steMetaIndex] for key in metaKeys} if self.dynamic else {key:[self.steMetaIndex]}
-
+      if len(outputMetaKeys) > 0:
+        params = {key:[self.pivotParameter,self.steMetaIndex] for key in outputMetaKeys + inputMetaKeys}
+        metaParams.update(params)
+      elif len(inputMetaKeys) > 0:
+        params = {key:[self.steMetaIndex] for key in inputMetaKeys}
+        metaParams.update(params)
+    metaKeys = inputMetaKeys + outputMetaKeys
     self.addMetaKeys(metaKeys,metaParams)
 
   def _localReadMoreXML(self, xmlNode):

diff --git a/tests/framework/PostProcessors/BasicStatistics/gold/timeDepMeta/Back_to_MASTER.xml b/tests/framework/PostProcessors/BasicStatistics/gold/timeDepMeta/Back_to_MASTER.xml
@@ -0,0 +1,66 @@
+<DataObjectMetadata name="SET_Back_to_MASTER">
+  <DataSet type="Static">
+    <dims>
+      <avg_EXP_MaxCtoF_a>Time_aligned</avg_EXP_MaxCtoF_a>
+      <avg_EXP_PCT_a>Time_aligned</avg_EXP_PCT_a>
+      <avg_EXP_P_a>Time_aligned</avg_EXP_P_a>
+      <avg_REP_MaxCtoF_a>Time_aligned</avg_REP_MaxCtoF_a>
+      <avg_REP_PCT_a>Time_aligned</avg_REP_PCT_a>
+      <avg_REP_P_a>Time_aligned</avg_REP_P_a>
+      <avg_ste_EXP_MaxCtoF_a>Time_aligned</avg_ste_EXP_MaxCtoF_a>
+      <avg_ste_EXP_PCT_a>Time_aligned</avg_ste_EXP_PCT_a>
+      <avg_ste_EXP_P_a>Time_aligned</avg_ste_EXP_P_a>
+      <avg_ste_REP_MaxCtoF_a>Time_aligned</avg_ste_REP_MaxCtoF_a>
+      <avg_ste_REP_PCT_a>Time_aligned</avg_ste_REP_PCT_a>
+      <avg_ste_REP_P_a>Time_aligned</avg_ste_REP_P_a>
+      <sen_EXP_MaxCtoF_a_HTcoeff>Time_aligned</sen_EXP_MaxCtoF_a_HTcoeff>
+      <sen_EXP_MaxCtoF_a_LinPow>Time_aligned</sen_EXP_MaxCtoF_a_LinPow>
+      <sen_EXP_MaxCtoF_a_RodPitch>Time_aligned</sen_EXP_MaxCtoF_a_RodPitch>
+      <sen_EXP_MaxCtoF_a_TCool>Time_aligned</sen_EXP_MaxCtoF_a_TCool>
+      <sen_EXP_MaxCtoF_a_VCool>Time_aligned</sen_EXP_MaxCtoF_a_VCool>
+      <sen_EXP_PCT_a_HTcoeff>Time_aligned</sen_EXP_PCT_a_HTcoeff>
+      <sen_EXP_PCT_a_LinPow>Time_aligned</sen_EXP_PCT_a_LinPow>
+      <sen_EXP_PCT_a_RodPitch>Time_aligned</sen_EXP_PCT_a_RodPitch>
+      <sen_EXP_PCT_a_TCool>Time_aligned</sen_EXP_PCT_a_TCool>
+      <sen_EXP_PCT_a_VCool>Time_aligned</sen_EXP_PCT_a_VCool>
+      <sen_EXP_P_a_HTcoeff>Time_aligned</sen_EXP_P_a_HTcoeff>
+      <sen_EXP_P_a_LinPow>Time_aligned</sen_EXP_P_a_LinPow>
+      <sen_EXP_P_a_RodPitch>Time_aligned</sen_EXP_P_a_RodPitch>
+      <sen_EXP_P_a_TCool>Time_aligned</sen_EXP_P_a_TCool>
+      <sen_EXP_P_a_VCool>Time_aligned</sen_EXP_P_a_VCool>
+      <sen_REP_MaxCtoF_a_HTcoeff>Time_aligned</sen_REP_MaxCtoF_a_HTcoeff>
+      <sen_REP_MaxCtoF_a_LinPow>Time_aligned</sen_REP_MaxCtoF_a_LinPow>
+      <sen_REP_MaxCtoF_a_RodPitch>Time_aligned</sen_REP_MaxCtoF_a_RodPitch>
+      <sen_REP_MaxCtoF_a_TCool>Time_aligned</sen_REP_MaxCtoF_a_TCool>
+      <sen_REP_MaxCtoF_a_VCool>Time_aligned</sen_REP_MaxCtoF_a_VCool>
+      <sen_REP_PCT_a_HTcoeff>Time_aligned</sen_REP_PCT_a_HTcoeff>
+      <sen_REP_PCT_a_LinPow>Time_aligned</sen_REP_PCT_a_LinPow>
+      <sen_REP_PCT_a_RodPitch>Time_aligned</sen_REP_PCT_a_RodPitch>
+      <sen_REP_PCT_a_TCool>Time_aligned</sen_REP_PCT_a_TCool>
+      <sen_REP_PCT_a_VCool>Time_aligned</sen_REP_PCT_a_VCool>
+      <sen_REP_P_a_HTcoeff>Time_aligned</sen_REP_P_a_HTcoeff>
+      <sen_REP_P_a_LinPow>Time_aligned</sen_REP_P_a_LinPow>
+      <sen_REP_P_a_TCool>Time_aligned</sen_REP_P_a_TCool>
+      <sen_REP_P_a_VCool>Time_aligned</sen_REP_P_a_VCool>
+      <var_EXP_MaxCtoF_a>Time_aligned</var_EXP_MaxCtoF_a>
+      <var_EXP_PCT_a>Time_aligned</var_EXP_PCT_a>
+      <var_EXP_P_a>Time_aligned</var_EXP_P_a>
+      <var_REP_MaxCtoF_a>Time_aligned</var_REP_MaxCtoF_a>
+      <var_REP_PCT_a>Time_aligned</var_REP_PCT_a>
+      <var_REP_P_a>Time_aligned</var_REP_P_a>
+      <var_ste_EXP_MaxCtoF_a>Time_aligned</var_ste_EXP_MaxCtoF_a>
+      <var_ste_EXP_PCT_a>Time_aligned</var_ste_EXP_PCT_a>
+      <var_ste_EXP_P_a>Time_aligned</var_ste_EXP_P_a>
+      <var_ste_REP_MaxCtoF_a>Time_aligned</var_ste_REP_MaxCtoF_a>
+      <var_ste_REP_PCT_a>Time_aligned</var_ste_REP_PCT_a>
+      <var_ste_REP_P_a>Time_aligned</var_ste_REP_P_a>
+    </dims>
+    <general>
+      <inputs>avg_VCool,avg_TCool,avg_LinPow,avg_HTcoeff,avg_RodPitch,var_VCool,var_TCool,var_LinPow,var_HTcoeff,var_RodPitch</inputs>
+      <outputs>sen_EXP_PCT_a_VCool,sen_EXP_MaxCtoF_a_VCool,sen_EXP_P_a_VCool,sen_EXP_PCT_a_TCool,sen_EXP_MaxCtoF_a_TCool,sen_EXP_P_a_TCool,sen_EXP_PCT_a_LinPow,sen_EXP_MaxCtoF_a_LinPow,sen_EXP_P_a_LinPow,sen_EXP_PCT_a_HTcoeff,sen_EXP_MaxCtoF_a_HTcoeff,sen_EXP_P_a_HTcoeff,sen_EXP_PCT_a_RodPitch,sen_EXP_MaxCtoF_a_RodPitch,sen_EXP_P_a_RodPitch,sen_REP_PCT_a_VCool,sen_REP_MaxCtoF_a_VCool,sen_REP_P_a_VCool,sen_REP_PCT_a_TCool,sen_REP_MaxCtoF_a_TCool,sen_REP_P_a_TCool,sen_REP_PCT_a_LinPow,sen_REP_MaxCtoF_a_LinPow,sen_REP_P_a_LinPow,sen_REP_PCT_a_HTcoeff,sen_REP_MaxCtoF_a_HTcoeff,sen_REP_P_a_HTcoeff,sen_REP_PCT_a_RodPitch,sen_REP_MaxCtoF_a_RodPitch,avg_EXP_PCT_a,avg_EXP_MaxCtoF_a,avg_EXP_P_a,avg_REP_PCT_a,avg_REP_MaxCtoF_a,avg_REP_P_a,var_EXP_PCT_a,var_EXP_MaxCtoF_a,var_EXP_P_a,var_REP_PCT_a,var_REP_MaxCtoF_a,var_REP_P_a</outputs>
+      <pointwise_meta>avg_ste_EXP_MaxCtoF_a,avg_ste_EXP_PCT_a,avg_ste_EXP_P_a,avg_ste_HTcoeff,avg_ste_LinPow,avg_ste_REP_MaxCtoF_a,avg_ste_REP_PCT_a,avg_ste_REP_P_a,avg_ste_RodPitch,avg_ste_TCool,avg_ste_VCool,var_ste_EXP_MaxCtoF_a,var_ste_EXP_PCT_a,var_ste_EXP_P_a,var_ste_HTcoeff,var_ste_LinPow,var_ste_REP_MaxCtoF_a,var_ste_REP_PCT_a,var_ste_REP_P_a,var_ste_RodPitch,var_ste_TCool,var_ste_VCool</pointwise_meta>
+      <sampleTag>RAVEN_sample_ID</sampleTag>
+    </general>
+  </DataSet>
+
+</DataObjectMetadata>
diff --git a/tests/framework/PostProcessors/BasicStatistics/test_time_dep_meta.xml b/tests/framework/PostProcessors/BasicStatistics/test_time_dep_meta.xml
@@ -0,0 +1,157 @@
+<?xml version="1.0" ?>
+<Simulation verbosity="debug">
+
+  <TestInfo>
+    <name>framework/PostProcessors/BasicStatistics.timeDepMeta</name>
+    <author>wangc</author>
+    <created>2019-05-10</created>
+    <classesTested>PostProcessors.BasicStatistics, DataObjects.HistorySet</classesTested>
+    <description>
+      This test checks the metadata printing of history set, the scalar meta data should be printed with
+      the input of HistorySet, while vector meta data should be printed with output of HistorySet. See issue #986.
+      With the fix of #986, The HistorySet will handle scalar and vector meta data separately.
+    </description>
+  </TestInfo>
+
+  <RunInfo>
+    <WorkingDir>timeDepMeta</WorkingDir>
+    <Sequence>Read_raw,  SensPost_1, Back_to_MASTER</Sequence>
+    <internalParallel>True</internalParallel>
+    <deleteOutExtension>o,plt,rst,csv,i</deleteOutExtension>
+  </RunInfo>
+
+  <Files>
+    <Input name="myinputfile" type="">RawData</Input>
+  </Files>
+
+  <VariableGroups>
+    <Group name="GRO_SensPost_in_features_scalar">
+      VCool,
+      TCool,
+      LinPow,
+      HTcoeff,
+      RodPitch
+    </Group>
+    <Group name="GRO_SensPost_in_targets_Timealigned">
+      EXP_PCT_a,
+      EXP_MaxCtoF_a,
+      EXP_P_a,
+      REP_PCT_a,
+      REP_MaxCtoF_a,
+      REP_P_a
+    </Group>
+    <Group name="GRO_SensPost_out_scalar">
+      avg_VCool,
+      avg_TCool,
+      avg_LinPow,
+      avg_HTcoeff,
+      avg_RodPitch,
+      var_VCool,
+      var_TCool,
+      var_LinPow,
+      var_HTcoeff,
+      var_RodPitch
+    </Group>
+    <Group name="GRO_SensPost_out_Timealigned">
+      sen_EXP_PCT_a_VCool,
+      sen_EXP_MaxCtoF_a_VCool,
+      sen_EXP_P_a_VCool,
+      sen_EXP_PCT_a_TCool,
+      sen_EXP_MaxCtoF_a_TCool,
+      sen_EXP_P_a_TCool,
+      sen_EXP_PCT_a_LinPow,
+      sen_EXP_MaxCtoF_a_LinPow,
+      sen_EXP_P_a_LinPow,
+      sen_EXP_PCT_a_HTcoeff,
+      sen_EXP_MaxCtoF_a_HTcoeff,
+      sen_EXP_P_a_HTcoeff,
+      sen_EXP_PCT_a_RodPitch,
+      sen_EXP_MaxCtoF_a_RodPitch,
+      sen_EXP_P_a_RodPitch,
+      sen_REP_PCT_a_VCool,
+      sen_REP_MaxCtoF_a_VCool,
+      sen_REP_P_a_VCool,
+      sen_REP_PCT_a_TCool,
+      sen_REP_MaxCtoF_a_TCool,
+      sen_REP_P_a_TCool,
+      sen_REP_PCT_a_LinPow,
+      sen_REP_MaxCtoF_a_LinPow,
+      sen_REP_P_a_LinPow,
+      sen_REP_PCT_a_HTcoeff,
+      sen_REP_MaxCtoF_a_HTcoeff,
+      sen_REP_P_a_HTcoeff,
+      sen_REP_PCT_a_RodPitch,
+      sen_REP_MaxCtoF_a_RodPitch,
+      sen_REP_P_a_HTcoeff,
+      avg_EXP_PCT_a,
+      avg_EXP_MaxCtoF_a,
+      avg_EXP_P_a,
+      avg_REP_PCT_a,
+      avg_REP_MaxCtoF_a,
+      avg_REP_P_a,
+      var_EXP_PCT_a,
+      var_EXP_MaxCtoF_a,
+      var_EXP_P_a,
+      var_REP_PCT_a,
+      var_REP_MaxCtoF_a,
+      var_REP_P_a
+    </Group>
+  </VariableGroups>
+  <Models>
+    <PostProcessor name="SensPost_1" subType="BasicStatistics" verbosity="debug">
+      <pivotParameter>Time_aligned</pivotParameter>
+      <sensitivity prefix="sen">
+        <targets>GRO_SensPost_in_targets_Timealigned</targets>
+        <features>GRO_SensPost_in_features_scalar</features>
+      </sensitivity>
+      <expectedValue prefix="avg">
+       GRO_SensPost_in_targets_Timealigned, GRO_SensPost_in_features_scalar
+    </expectedValue>
+      <variance prefix="var">
+       GRO_SensPost_in_targets_Timealigned, GRO_SensPost_in_features_scalar
+    </variance>
+    </PostProcessor>
+  </Models>
+
+  <DataObjects>
+    <HistorySet name="SET_SensPost_aligned">
+      <Input>GRO_SensPost_in_features_scalar</Input>
+      <Output>GRO_SensPost_in_targets_Timealigned</Output>
+      <options>
+        <pivotParameter>Time_aligned</pivotParameter>
+      </options>
+    </HistorySet>
+    <HistorySet name="SET_Back_to_MASTER">
+      <Input>GRO_SensPost_out_scalar</Input>
+      <Output>GRO_SensPost_out_Timealigned</Output>
+      <options>
+        <pivotParameter>Time_aligned</pivotParameter>
+      </options>
+    </HistorySet>
+  </DataObjects>
+
+  <Steps>
+    <IOStep name="Read_raw">
+      <Input class="Files" type="csv">myinputfile</Input>
+      <Output class="DataObjects" type="HiostorySet">SET_SensPost_aligned</Output>
+    </IOStep>
+
+    <PostProcess name="SensPost_1">
+      <Input class="DataObjects" type="HistorySet">SET_SensPost_aligned</Input>
+      <Model class="Models" type="PostProcessor">SensPost_1</Model>
+      <Output class="DataObjects" type="HistorySet">SET_Back_to_MASTER</Output>
+    </PostProcess>
+
+    <IOStep name="Back_to_MASTER">
+      <Input class="DataObjects" type="HistorySet">SET_Back_to_MASTER</Input>
+      <Output class="OutStreams" type="Print">Back_to_MASTER</Output>
+    </IOStep>
+  </Steps>
+
+  <OutStreams>
+    <Print name="Back_to_MASTER">
+      <type>csv</type>
+      <source>SET_Back_to_MASTER</source>
+    </Print>
+  </OutStreams>
+</Simulation>