class RandomODCMPermutationsSvc:
###
# Initialize the service (stateless).
###
def __init__(self):
self.kfHelper = KFunctionHelper()
###
# Generate the ODCM permutations.
# @param analysisType Either Global Analysis or Cross Analysis.
# @param srcPoints The source points.
# @param destPoints The destination points. Ignored if analysisType is GLOBAL.
# @param networkDataset The network dataset to use for the ODCMs.
# @param snapDist The snap distance for points that are not directly on the network.
# @param cutoff The cutoff distance. Ignored if None.
# @param outLoc The full path to a database. The ODCM data will be written here.
# @param outFC The name of the feature class in outLoc where the ODCM data will be written.
# @param numPerms The number of permutations (string representation).
# @param outCoordSys The coordinate system to project the points into (optional).
# @param numPointsFieldName The optional name of a field in the network dataset's edge source
# from which the number of random points should be derived.
# @param messages A messages instances with addMessage() implemented (for debug output).
# @param callback A callback function(odDists, iteration) called on each iteration
# with the current OD cost matrix.
###
def generateODCMPermutations(self, analysisType, srcPoints, destPoints,
networkDataset, snapDist, cutoff, outLoc, outFC, numPerms, outCoordSys,
numPointsFieldName, messages, callback = None):
# Default no-op for the callback.
if callback is None:
callback = lambda odDists, iteration: None
# For global analysis the destination points are the same as the source points
# (e.g. destPoints is ignored).
if analysisType == "GLOBAL" or destPoints is None:
destPoints = srcPoints
# Count the number of crashes (there may be fewer points in the ODCM, but it's the total
# number of crashes that is needed).
numDests = self.kfHelper.countNumberOfFeatures(os.path.join(outLoc, destPoints))
messages.addMessage("Number of crashes: {0}".format(numDests))
# Make the observed ODCM and calculate the distance between each set of
# points. If a cross analysis is selected, find the distance between the
# source and destination points. Otherwise there is only one set of points
odDists = self._calculateDistances(networkDataset, srcPoints, destPoints, snapDist, cutoff)
self._writeODCMData(odDists, outLoc, outFC, 0)
callback(odDists, 0)
messages.addMessage("Iteration 0 (observed) complete.")
# Generate the OD Cost matrix permutations.
kfTimer = KFunctionTimer(numPerms)
for i in range(1, numPerms + 1):
if numPointsFieldName:
randPoints = self.kfHelper.generateRandomPoints(networkDataset, outCoordSys, None, numPointsFieldName)
else:
randPoints = self.kfHelper.generateRandomPoints(networkDataset, outCoordSys, numDests, None)
# See the note above: Either find the distance from the source points to the random points,
# or the distance between the random points.
if analysisType == "CROSS":
odDists = self._calculateDistances(networkDataset, srcPoints, randPoints, snapDist, cutoff)
else:
odDists = self._calculateDistances(networkDataset, randPoints, randPoints, snapDist, cutoff)
self._writeODCMData(odDists, outLoc, outFC, i)
callback(odDists, i)
# Clean up the random points table.
arcpy.Delete_management(randPoints)
# Show the progress.
kfTimer.increment()
messages.addMessage("Iteration {0} complete. Elapsed time: {1}s. ETA: {2}s.".format(
i, kfTimer.getElapsedTime(), kfTimer.getETA()))
###
# Calculate the distances between each set of points using an OD Cost Matrix.
# @param networkDataset A network dataset which the points are on.
# @param srcPoints The source points to calculate distances from.
# @param destPoints The destination points to calculate distances to.
# @param snapDist If a point is not directly on the network, it will be
# snapped to the nearset line if it is within this threshold.
# @param cutoff The cutoff distance for the ODCM (optional).
###
def _calculateDistances(self, networkDataset, srcPoints, destPoints, snapDist, cutoff):
# This is the current map, which should be an OSM base map.
curMapDoc = arcpy.mapping.MapDocument("CURRENT")
# Get the data from from the map (see the DataFrame object of arcpy).
dataFrame = arcpy.mapping.ListDataFrames(curMapDoc, "Layers")[0]
# Create the cost matrix.
costMatResult = arcpy.na.MakeODCostMatrixLayer(networkDataset, "TEMP_ODCM_NETWORK_K", "Length", cutoff)
odcmLayer = costMatResult.getOutput(0)
# The OD Cost Matrix layer will have Origins and Destinations layers. Get
# a reference to each of these.
odcmSublayers = arcpy.na.GetNAClassNames(odcmLayer)
odcmOriginLayer = odcmSublayers["Origins"]
odcmDestLayer = odcmSublayers["Destinations"]
# Add the origins and destinations to the ODCM.
arcpy.na.AddLocations(odcmLayer, odcmOriginLayer, srcPoints, "", snapDist)
arcpy.na.AddLocations(odcmLayer, odcmDestLayer, destPoints, "", snapDist)
# Solve the matrix.
arcpy.na.Solve(odcmLayer)
# Show the ODCM layer (it must be showing to open th ODLines sub layer below).
#arcpy.mapping.AddLayer(dataFrame, odcmLayer, "TOP")
#arcpy.RefreshTOC()
# Get the "Lines" layer, which has the distance between each point.
odcmLines = arcpy.mapping.ListLayers(odcmLayer, odcmSublayers["ODLines"])[0]
# This array will hold all the OD distances.
odDists = []
if srcPoints == destPoints:
# If the source points and destination points are the same, exclude the
# distance from the point to itself.
where = """{0} <> {1}""".format(
arcpy.AddFieldDelimiters(odcmLines, "originID"),
arcpy.AddFieldDelimiters(odcmLines, "destinationID"))
else:
where = ""
with arcpy.da.SearchCursor(
in_table=odcmLines,
field_names=["Total_Length", "originID", "destinationID"],
where_clause=where) as cursor:
for row in cursor:
odDists.append({"Total_Length": row[0], "OriginID": row[1], "DestinationID": row[2]})
return odDists
###
# Write the ODCM data to a table.
# @param odDists An array of raw ODCM data.
# @param outLoc The location of a database.
# @param outFC The feature class name, in outLoc, to write the data to.
# @param iteration The iteration number (0 is observed).
###
def _writeODCMData(self, odDists, outLoc, outFC, iteration):
# This is the full path to the output feature class.
outFCFullPath = os.path.join(outLoc, outFC)
if iteration == 0:
# Create the output table.
arcpy.CreateTable_management(outLoc, outFC)
arcpy.AddField_management(outFCFullPath, "Iteration_Number", "LONG")
arcpy.AddField_management(outFCFullPath, "OriginID", "LONG")
arcpy.AddField_management(outFCFullPath, "DestinationID", "LONG")
arcpy.AddField_management(outFCFullPath, "Total_Length", "DOUBLE")
with arcpy.da.InsertCursor(outFCFullPath,
["Iteration_Number", "OriginID", "DestinationID", "Total_Length"]) as cursor:
for odDist in odDists:
cursor.insertRow([iteration, odDist["OriginID"], odDist["DestinationID"], odDist["Total_Length"]])
