Skip to content

Commit

Permalink
Merge pull request #3 from rchristie/working
Browse files Browse the repository at this point in the history 
Match groups and markers ignoring case and whitespace differences
  • Loading branch information
@rchristie
rchristie authored Apr 22, 2020
2 parents e70ec44 + 78910ec commit f77bfe9
Show file tree
Hide file tree
Showing 2 changed files with 72 additions and 44 deletions.
82 changes: 54 additions & 28 deletions src/scaffoldfitter/fitter.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -105,24 +105,50 @@ def _loadModel(self):
def _loadData(self):
"""
Load zinc data file into self._rawDataRegion.
Rename data groups to exactly match model groups where they differ by case and whitespace only.
Transfer data points (and converted nodes) into self._region.
"""
result = self._rawDataRegion.readFile(self._zincDataFileName)
assert result == RESULT_OK, "Failed to load data file " + str(self._zincDataFileName)
# if there both nodes and datapoints, offset datapoint identifiers to ensure different
fieldmodule = self._rawDataRegion.getFieldmodule()
nodes = fieldmodule.findNodesetByFieldDomainType(Field.DOMAIN_TYPE_NODES)
if nodes.getSize() > 0:
datapoints = fieldmodule.findNodesetByFieldDomainType(Field.DOMAIN_TYPE_DATAPOINTS)
if datapoints.getSize() > 0:
maximumDatapointIdentifier = max(0, getMaximumNodeIdentifier(datapoints))
maximumNodeIdentifier = max(0, getMaximumNodeIdentifier(nodes))
# this assumes identifiers are in low ranges and can be improved if there is a problem:
identifierOffset = 100000
while (maximumDatapointIdentifier > identifierOffset) or (maximumNodeIdentifier > identifierOffset):
assert identifierOffset < 1000000000, "Invalid node and datapoint identifier ranges"
identifierOffset *= 10
with ChangeManager(fieldmodule):
with ChangeManager(fieldmodule):
# rename data groups to match model
# future: match with annotation terms
modelGroupNames = [ group.getName() for group in getGroupList(self._fieldmodule) ]
writeDiagnostics = self.getDiagnosticLevel() > 0
for dataGroup in getGroupList(fieldmodule):
dataGroupName = dataGroup.getName()
compareName = dataGroupName.strip().casefold()
for modelGroupName in modelGroupNames:
if modelGroupName == dataGroupName:
if writeDiagnostics:
print("Load data: Data group '" + dataGroupName + "' found in model")
break
elif modelGroupName.strip().casefold() == compareName:
result = dataGroup.setName(modelGroupName)
if result == RESULT_OK:
if writeDiagnostics:
print("Load data: Data group '" + dataGroupName + "' found in model as '" + modelGroupName + "'. Renaming to match.")
else:
print("Error: Load data: Data group '" + dataGroupName + "' found in model as '" + modelGroupName + "'. Renaming to match FAILED.")
if fieldmodule.findFieldByName(modelGroupName).isValid():
print(" Reason: field of that name already exists.")
break
else:
if writeDiagnostics:
print("Load data: Data group '" + dataGroupName + "' not found in model")
# if there both nodes and datapoints, offset datapoint identifiers to ensure different
nodes = fieldmodule.findNodesetByFieldDomainType(Field.DOMAIN_TYPE_NODES)
if nodes.getSize() > 0:
datapoints = fieldmodule.findNodesetByFieldDomainType(Field.DOMAIN_TYPE_DATAPOINTS)
if datapoints.getSize() > 0:
maximumDatapointIdentifier = max(0, getMaximumNodeIdentifier(datapoints))
maximumNodeIdentifier = max(0, getMaximumNodeIdentifier(nodes))
# this assumes identifiers are in low ranges and can be improved if there is a problem:
identifierOffset = 100000
while (maximumDatapointIdentifier > identifierOffset) or (maximumNodeIdentifier > identifierOffset):
assert identifierOffset < 1000000000, "Invalid node and datapoint identifier ranges"
identifierOffset *= 10
while True:
# logic relies on datapoints being in identifier order
datapoint = datapoints.createNodeiterator().next()
Expand All @@ -131,18 +157,18 @@ def _loadData(self):
break;
result = datapoint.setIdentifier(identifier + identifierOffset)
assert result == RESULT_OK, "Failed to offset datapoint identifier"
# transfer nodes as datapoints to self._region
sir = self._rawDataRegion.createStreaminformationRegion()
srm = sir.createStreamresourceMemory()
sir.setResourceDomainTypes(srm, Field.DOMAIN_TYPE_NODES)
self._rawDataRegion.write(sir)
result, buffer = srm.getBuffer()
assert result == RESULT_OK, "Failed to write nodes"
buffer = buffer.replace(bytes("!#nodeset nodes", "utf-8"), bytes("!#nodeset datapoints", "utf-8"))
sir = self._region.createStreaminformationRegion()
srm = sir.createStreamresourceMemoryBuffer(buffer)
result = self._region.read(sir)
assert result == RESULT_OK, "Failed to load nodes as datapoints"
# transfer nodes as datapoints to self._region
sir = self._rawDataRegion.createStreaminformationRegion()
srm = sir.createStreamresourceMemory()
sir.setResourceDomainTypes(srm, Field.DOMAIN_TYPE_NODES)
self._rawDataRegion.write(sir)
result, buffer = srm.getBuffer()
assert result == RESULT_OK, "Failed to write nodes"
buffer = buffer.replace(bytes("!#nodeset nodes", "utf-8"), bytes("!#nodeset datapoints", "utf-8"))
sir = self._region.createStreaminformationRegion()
srm = sir.createStreamresourceMemoryBuffer(buffer)
result = self._region.read(sir)
assert result == RESULT_OK, "Failed to load nodes as datapoints"
# transfer datapoints to self._region
sir = self._rawDataRegion.createStreaminformationRegion()
srm = sir.createStreamresourceMemory()
Expand Down Expand Up @@ -319,8 +345,8 @@ def _calculateMarkerDataLocations(self):
fieldcache.setNode(datapoint)
name = self._markerDataNameField.evaluateString(fieldcache)
# if this is the only datapoint with name:
if name and findNodeWithName(self._markerDataGroup, self._markerDataNameField, name):
node = findNodeWithName(self._markerNodeGroup, self._markerNameField, name)
if name and findNodeWithName(self._markerDataGroup, self._markerDataNameField, name, ignore_case=True, strip_whitespace=True):
node = findNodeWithName(self._markerNodeGroup, self._markerNameField, name, ignore_case=True, strip_whitespace=True)
if node:
fieldcache.setNode(node)
element, xi = self._markerLocationField.evaluateMeshLocation(fieldcache, meshDimension)
Expand Down Expand Up @@ -539,7 +565,7 @@ def calculateDataProjections(self):
unprojectedDatapoints.removeNodesConditional(self._dataProjectionNodeGroupFields[d])
unprojectedCount = unprojectedDatapoints.getSize()
if unprojectedCount > 0:
print("Warning: " + str(unprojected) + " data points with data coordinates have not been projected")
print("Warning: " + str(unprojectedCount) + " data points with data coordinates have not been projected")
del unprojectedDatapoints

# remove temporary objects before ChangeManager exits
Expand Down
34 changes: 18 additions & 16 deletions src/scaffoldfitter/fitterstepalign.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -135,23 +135,25 @@ def _doAlignMarkers(self):
assert markerDataGroup and markerDataCoordinates and markerDataName, "Align: No marker data group, coordinates or name fields"
dataMarkers = getNodeNameCentres(markerDataGroup, markerDataCoordinates, markerDataName)

# match model and data markers, warn of missing markers
# match model and data markers, warn of unmatched markers
markerMap = {}
for name, modelx in modelMarkers.items():
datax = dataMarkers.get(name)
if datax:
markerMap[name] = ( modelx, datax )
if self.getDiagnosticLevel() > 0:
for name in modelMarkers:
datax = dataMarkers.get(name)
if datax:
print("Align: Found marker " + name + " in model and data")
for name in modelMarkers:
if not markerMap.get(name):
print("Align: Model marker " + name + " not found in data")
for name in dataMarkers:
if not markerMap.get(name):
print("Align: Data marker " + name + " not found in model")
writeDiagnostics = self.getDiagnosticLevel() > 0
for modelName in modelMarkers:
# name match allows case and whitespace differences
matchName = modelName.strip().casefold()
for dataName in dataMarkers:
if dataName.strip().casefold() == matchName:
markerMap[modelName] = ( modelMarkers[modelName], dataMarkers[dataName] )
if writeDiagnostics:
print("Align: Model marker '" + modelName + "' found in data" + (" as '" + dataName +"'" if (dataName != modelName) else ""))
dataMarkers.pop(dataName)
break
else:
if writeDiagnostics:
print("Align: Model marker '" + modelName + "' not found in data")
if writeDiagnostics:
for dataName in dataMarkers:
print("Align: Data marker '" + dataName + "' not found in model")

self._optimiseAlignment(markerMap)

Expand Down

0 comments on commit f77bfe9

Please sign in to comment.