def testCluster(self):
# preparation
clusteringBuilder = clustering.ClusteringBuilder()
clusteringParams = {"type": "rmsd",
"params": {"ligandResname": "AIN",
"contactThresholdDistance": 8}}
clusteringInstance = clusteringBuilder.buildClustering(clusteringParams,
"ain_report", 3)
trajNames = ["tests/data/aspirin_data/traj*"]
# function to test
clusteringInstance.cluster(trajNames)
# assertion
allClusters = clusteringInstance.clusters.clusters
goldenNumberOfClusters = 2
goldenEnergyCluster1 = -8424.8
goldenEnergyCluster2 = -8453.29
goldenElementsCluster1 = 2
goldenElementsCluster2 = 1
self.assertEqual(len(allClusters), goldenNumberOfClusters)
self.assertAlmostEqual(allClusters[0].getMetric(), goldenEnergyCluster1, 2)
self.assertAlmostEqual(allClusters[1].getMetric(), goldenEnergyCluster2, 2)
self.assertEqual(allClusters[0].elements, goldenElementsCluster1)
self.assertEqual(allClusters[1].elements, goldenElementsCluster2)
def testCluster_protein_protein(self):
# preparation
clusteringBuilder = clustering.ClusteringBuilder()
clusteringParams = {"type": "rmsd",
"params": {"ligandChain": "B",
"contactThresholdDistance": 8},
"thresholdCalculator" : {
"type" : "constant",
"params" : { "value" : 4.0 }
}}
clusteringInstance = clusteringBuilder.buildClustering(clusteringParams,
"report", 3)
trajNames = ["tests/data/protein_protein_data/traj*"]
# function to test
clusteringInstance.cluster(trajNames)
# assertion
allClusters = clusteringInstance.clusters.clusters
goldenNumberOfClusters = 2
goldenEnergyCluster1 = -18906.7
goldenEnergyCluster2 = -18905
goldenElementsCluster1 = 8
goldenElementsCluster2 = 3
self.assertEqual(len(allClusters), goldenNumberOfClusters)
self.assertAlmostEqual(allClusters[0].getMetric(), goldenEnergyCluster1, 2)
self.assertAlmostEqual(allClusters[1].getMetric(), goldenEnergyCluster2, 2)
self.assertEqual(allClusters[0].elements, goldenElementsCluster1)
self.assertEqual(allClusters[1].elements, goldenElementsCluster2)
def getWorkingClusteringObjectAndReclusterIfNecessary(
firstRun, outputPathConstants, clusteringBlock, spawningParams,
simulationRunner, topologies, processManager):
"""
It reads the previous clustering method, and, if there are changes,
it reclusters the previous trajectories. Returns the clustering object to use
:param firstRun: New epoch to run
:type firstRun: int
:param outputPathConstants: Contains outputPath-related constants
:type outputPathConstants: :py:class:`.OutputPathConstants`
:param clusteringBlock: Contains the new clustering block
:type clusteringBlock: json
:param spawningParams: Spawning params, to know what reportFile and column to read
:type spawningParams: :py:class:`.SpawningParams`
:param topologies: Topology object containing the set of topologies needed for the simulation
:type topologies: :py:class:`.Topology`
:param processManager: Object to synchronize the possibly multiple processes
:type processManager: :py:class:`.ProcessesManager`
:returns: :py:class:`.Clustering` -- The clustering method to use in the
adaptive sampling simulation
"""
if not processManager.isMaster():
for ij in range(firstRun):
topologies.readMappingFromDisk(
outputPathConstants.epochOutputPathTempletized % ij, ij)
return
lastClusteringEpoch = firstRun - 1
clusteringObjectPath = outputPathConstants.clusteringOutputObject % (
lastClusteringEpoch)
oldClusteringMethod = utilities.readClusteringObject(clusteringObjectPath)
clusteringBuilder = clustering.ClusteringBuilder()
clusteringMethod = clusteringBuilder.buildClustering(
clusteringBlock, spawningParams.reportFilename,
spawningParams.reportCol)
clusteringMethod.setProcessors(simulationRunner.getWorkingProcessors())
if needToRecluster(oldClusteringMethod, clusteringMethod):
utilities.print_unbuffered("Reclustering!")
startTime = time.time()
clusterPreviousEpochs(clusteringMethod, firstRun,
outputPathConstants.epochOutputPathTempletized,
simulationRunner, topologies,
outputPathConstants.allTrajsPath)
endTime = time.time()
utilities.print_unbuffered("Reclustering took %s sec" %
(endTime - startTime))
else:
clusteringMethod = oldClusteringMethod
clusteringMethod.setCol(spawningParams.reportCol)
for ij in range(firstRun):
topologies.readMappingFromDisk(
outputPathConstants.epochOutputPathTempletized % ij, ij)
return clusteringMethod
def buildNewClusteringAndWriteInitialStructuresInNewSimulation(
debug, controlFile, outputPathConstants, clusteringBlock,
spawningParams, initialStructures, simulationRunner, processManager):
"""
Build the clustering object and copies initial structures from control file.
Returns the clustering object to use and the initial structures filenames as string
:param debug: In debug, it will not remove the simulations
:type debug: bool
:param controlFile: Adaptive sampling control file
:type controlFile: str
:param outputPathConstants: Contains outputPath-related constants
:type outputPathConstants: :py:class:`.OutputPathConstants`
:param clusteringBlock: Contains the new clustering block
:type clusteringBlock: json
:param spawningParams: Spawning params
:type spawningParams: :py:class:`.SpawningParams`
:param initialStructures: Control file initial structures
:type initialStructures: list
:param simulationRunner: :py:class:`.SimulationRunner` Simulation runner object
:type simulationRunner: :py:class:`.SimulationRunner`
:param processManager: Object to synchronize the possibly multiple processes
:type processManager: :py:class:`.ProcessesManager`
:returns: :py:class:`.Clustering`, str -- The clustering method to use in the adaptive sampling simulation and the initial structures filenames
"""
firstRun = 0
if processManager.isMaster():
saveInitialControlFile(controlFile,
outputPathConstants.originalControlFile)
copyInitialStructures(initialStructures,
outputPathConstants.tmpInitialStructuresTemplate,
firstRun)
processManager.barrier()
initialStructuresAsString = simulationRunner.createMultipleComplexesFilenames(
len(initialStructures),
outputPathConstants.tmpInitialStructuresTemplate, firstRun)
if processManager.isMaster():
clusteringBuilder = clustering.ClusteringBuilder()
clusteringMethod = clusteringBuilder.buildClustering(
clusteringBlock, spawningParams.reportFilename,
spawningParams.reportCol)
else:
clusteringMethod = None
return clusteringMethod, initialStructuresAsString
def test_cluster_accumulative(self):
# preparation
clusteringParams = {"type": "contactMap",
"params": {"ligandResname": "AIN",
"contactThresholdDistance": 8,
"similarityEvaluator": "correlation"},
"thresholdCalculator": {
"type": "constant",
"params": {
"value": 0.15
}
}}
clusteringBuilder = clustering.ClusteringBuilder()
clusteringInstance = clusteringBuilder.buildClustering(clusteringParams,
"ain_report", 3)
trajNames = ["tests/data/aspirin_data/traj*"]
# function to test
clusteringInstance.cluster(trajNames)
# assertion
allClusters = clusteringInstance.clusters.clusters
# goldenNumberOfClusters = 3
# goldenEnergyCluster1 = -8421.5
# goldenEnergyCluster2 = -8424.8
# goldenEnergyCluster3 = -8453.29
# goldenElementsCluster1 = 1
# goldenElementsCluster2 = 1
# goldenElementsCluster3 = 1
goldenNumberOfClusters = 2
goldenEnergyCluster1 = -8424.8
goldenEnergyCluster2 = -8453.29
goldenElementsCluster1 = 2
goldenElementsCluster2 = 1
self.assertEqual(len(allClusters), goldenNumberOfClusters)
self.assertAlmostEqual(allClusters[0].getMetric(), goldenEnergyCluster1, 2)
self.assertAlmostEqual(allClusters[1].getMetric(), goldenEnergyCluster2, 2)
# self.assertAlmostEqual(allClusters[2].getMetric(), goldenEnergyCluster3, 2)
self.assertEqual(allClusters[0].elements, goldenElementsCluster1)
self.assertEqual(allClusters[1].elements, goldenElementsCluster2)
