TEST_F(AnalysisDriverFixture, DDACE_MonteCarlo_Continuous) {
// RETRIEVE PROBLEM
Problem problem = retrieveProblem("Continuous",true,false);
// DEFINE SEED
Model model = model::exampleModel();
openstudio::path p = toPath("./example.osm");
model.save(p,true);
FileReference seedModel(p);
// CREATE ANALYSIS
DDACEAlgorithmOptions algOptions(DDACEAlgorithmType::random);
Analysis analysis("DDACE Monte Carlo Sampling",
problem,
DDACEAlgorithm(algOptions),
seedModel);
// RUN ANALYSIS
{
ProjectDatabase database = getCleanDatabase("DDACEMonteCarlo_Continuous_NoSamples");
AnalysisDriver analysisDriver(database);
AnalysisRunOptions runOptions = standardRunOptions(analysisDriver.database().path().parent_path());
CurrentAnalysis currentAnalysis = analysisDriver.run(analysis,runOptions);
EXPECT_TRUE(analysisDriver.waitForFinished());
boost::optional<runmanager::JobErrors> jobErrors = currentAnalysis.dakotaJobErrors();
ASSERT_TRUE(jobErrors);
EXPECT_FALSE(jobErrors->errors().empty()); // require specification of number of samples
EXPECT_TRUE(analysisDriver.currentAnalyses().empty());
Table summary = currentAnalysis.analysis().summaryTable();
EXPECT_EQ(1u,summary.nRows()); // no points
}
{
algOptions.setSamples(6);
analysis = Analysis("DDACE MonteCarlo Sampling",
problem,
DDACEAlgorithm(algOptions),
seedModel);
ProjectDatabase database = getCleanDatabase("DDACEMonteCarlo_Continuous");
AnalysisDriver analysisDriver = AnalysisDriver(database);
AnalysisRunOptions runOptions = standardRunOptions(analysisDriver.database().path().parent_path());
runOptions.setQueueSize(4);
CurrentAnalysis currentAnalysis = analysisDriver.run(analysis,runOptions);
EXPECT_TRUE(analysisDriver.waitForFinished());
boost::optional<runmanager::JobErrors> jobErrors = currentAnalysis.dakotaJobErrors();
ASSERT_TRUE(jobErrors);
EXPECT_TRUE(jobErrors->errors().empty());
EXPECT_TRUE(analysisDriver.currentAnalyses().empty());
Table summary = currentAnalysis.analysis().summaryTable();
EXPECT_EQ(7u,summary.nRows()); // 6 points
summary.save(analysisDriver.database().path().parent_path() / toPath("summary.csv"));
BOOST_FOREACH(const DataPoint& dataPoint,analysis.dataPoints()) {
EXPECT_TRUE(dataPoint.isComplete());
EXPECT_FALSE(dataPoint.failed());
// EXPECT_FALSE(dataPoint.responseValues().empty());
}
}
}
TEST_F(AnalysisDriverFixture, DDACE_MonteCarlo_MixedOsmIdf_ProjectDatabaseOpen) {
// RETRIEVE PROBLEM
Problem problem = retrieveProblem("MixedOsmIdf",false,false);
// DEFINE SEED
Model model = model::exampleModel();
openstudio::path p = toPath("./example.osm");
model.save(p,true);
FileReference seedModel(p);
// CREATE ANALYSIS
DDACEAlgorithmOptions algOptions(DDACEAlgorithmType::oas);
algOptions.setSamples(4);
Analysis analysis("DDACE Monte Carlo Sampling",
problem,
DDACEAlgorithm(algOptions),
seedModel);
// RUN ANALYSIS
{
analysis = Analysis("DDACE Monte Carlo Sampling - MixedOsmIdf",
problem,
DDACEAlgorithm(algOptions),
seedModel);
ProjectDatabase database = getCleanDatabase("DDACEMonteCarlo_MixedOsmIdf");
AnalysisDriver analysisDriver = AnalysisDriver(database);
AnalysisRunOptions runOptions = standardRunOptions(analysisDriver.database().path().parent_path());
CurrentAnalysis currentAnalysis = analysisDriver.run(analysis,runOptions);
EXPECT_TRUE(analysisDriver.waitForFinished());
boost::optional<runmanager::JobErrors> jobErrors = currentAnalysis.dakotaJobErrors();
ASSERT_TRUE(jobErrors);
EXPECT_TRUE(jobErrors->errors().empty());
EXPECT_TRUE(analysisDriver.currentAnalyses().empty());
Table summary = currentAnalysis.analysis().summaryTable();
EXPECT_EQ(5u,summary.nRows()); // 4 points
summary.save(analysisDriver.database().path().parent_path() / toPath("summary.csv"));
BOOST_FOREACH(const DataPoint& dataPoint,analysis.dataPoints()) {
EXPECT_TRUE(dataPoint.isComplete());
EXPECT_FALSE(dataPoint.failed());
}
}
{
project::OptionalProjectDatabase oDatabase = project::ProjectDatabase::open(toPath("AnalysisDriverFixtureData/DDACEMonteCarloMixedOsmIdf/DDACEMonteCarlo_MixedOsmIdf.osp"));
ASSERT_TRUE(oDatabase);
project::AnalysisRecordVector analysisRecords = project::AnalysisRecord::getAnalysisRecords(*oDatabase);
EXPECT_EQ(1u,analysisRecords.size());
if (!analysisRecords.empty()) {
EXPECT_NO_THROW(analysisRecords[0].analysis());
}
}
}
TEST_F(AnalysisDriverFixture, FSUDace_CVT_MixedOsmIdf) {
// RETRIEVE PROBLEM
Problem problem = retrieveProblem("MixedOsmIdf",false,false);
// DEFINE SEED
Model model = model::exampleModel();
openstudio::path p = toPath("./example.osm");
model.save(p,true);
FileReference seedModel(p);
// CREATE ANALYSIS
FSUDaceAlgorithmOptions algOptions(FSUDaceAlgorithmType::cvt);
algOptions.setSamples(3);
Analysis analysis("FSUDace CVT",
problem,
FSUDaceAlgorithm(algOptions),
seedModel);
// RUN ANALYSIS
analysis = Analysis("FSUDace CVT - MixedOsmIdf",
problem,
FSUDaceAlgorithm(algOptions),
seedModel);
ProjectDatabase database = getCleanDatabase("FSUDaceCVT_MixedOsmIdf");
AnalysisDriver analysisDriver = AnalysisDriver(database);
AnalysisRunOptions runOptions = standardRunOptions(analysisDriver.database().path().parent_path());
CurrentAnalysis currentAnalysis = analysisDriver.run(analysis,runOptions);
EXPECT_TRUE(analysisDriver.waitForFinished());
boost::optional<runmanager::JobErrors> jobErrors = currentAnalysis.dakotaJobErrors();
ASSERT_TRUE(jobErrors);
EXPECT_TRUE(jobErrors->errors().empty());
EXPECT_TRUE(analysisDriver.currentAnalyses().empty());
Table summary = currentAnalysis.analysis().summaryTable();
// EXPECT_EQ(10u,summary.nRows()); // 9 points
summary.save(analysisDriver.database().path().parent_path() / toPath("summary.csv"));
BOOST_FOREACH(const DataPoint& dataPoint,analysis.dataPoints()) {
EXPECT_TRUE(dataPoint.isComplete());
EXPECT_FALSE(dataPoint.failed());
}
}
TEST_F(AnalysisDriverFixture, DDACE_LatinHypercube_UserScriptContinuous) {
// RETRIEVE PROBLEM
Problem problem = retrieveProblem("UserScriptContinuous",true,false);
// DEFINE SEED
Model model = model::exampleModel();
openstudio::path p = toPath("./example.osm");
model.save(p,true);
FileReference seedModel(p);
// CREATE ANALYSIS
DDACEAlgorithmOptions algOptions(DDACEAlgorithmType::lhs);
algOptions.setSamples(10);
// RUN ANALYSIS
Analysis analysis("DDACE Latin Hypercube Sampling - UserScriptContinuous",
problem,
DDACEAlgorithm(algOptions),
seedModel);
ProjectDatabase database = getCleanDatabase("DDACELatinHypercube_UserScriptContinuous");
AnalysisDriver analysisDriver = AnalysisDriver(database);
AnalysisRunOptions runOptions = standardRunOptions(analysisDriver.database().path().parent_path());
CurrentAnalysis currentAnalysis = analysisDriver.run(analysis,runOptions);
EXPECT_TRUE(analysisDriver.waitForFinished());
boost::optional<runmanager::JobErrors> jobErrors = currentAnalysis.dakotaJobErrors();
ASSERT_TRUE(jobErrors);
EXPECT_TRUE(jobErrors->errors().empty());
EXPECT_TRUE(analysisDriver.currentAnalyses().empty());
Table summary = currentAnalysis.analysis().summaryTable();
EXPECT_EQ(11u,summary.nRows());
summary.save(analysisDriver.database().path().parent_path() / toPath("summary.csv"));
BOOST_FOREACH(const DataPoint& dataPoint,analysis.dataPoints()) {
EXPECT_TRUE(dataPoint.isComplete());
EXPECT_FALSE(dataPoint.failed());
}
}
TEST_F(AnalysisDriverFixture, DDACE_Grid_Continuous) {
// RETRIEVE PROBLEM
Problem problem = retrieveProblem("Continuous",true,false);
// DEFINE SEED
Model model = model::exampleModel();
openstudio::path p = toPath("./example.osm");
model.save(p,true);
FileReference seedModel(p);
// CREATE ANALYSIS
DDACEAlgorithmOptions algOptions(DDACEAlgorithmType::grid);
Analysis analysis("DDACE Grid Sampling",
problem,
DDACEAlgorithm(algOptions),
seedModel);
// RUN ANALYSIS
{
ProjectDatabase database = getCleanDatabase("DDACEGridSampling_NoSamples");
AnalysisDriver analysisDriver(database);
AnalysisRunOptions runOptions = standardRunOptions(analysisDriver.database().path().parent_path());
runOptions.setQueueSize(4);
CurrentAnalysis currentAnalysis = analysisDriver.run(analysis,runOptions);
EXPECT_TRUE(analysisDriver.waitForFinished());
boost::optional<runmanager::JobErrors> jobErrors = currentAnalysis.dakotaJobErrors();
ASSERT_TRUE(jobErrors);
EXPECT_FALSE(jobErrors->errors().empty()); // require specification of number of samples
EXPECT_TRUE(analysisDriver.currentAnalyses().empty());
Table summary = currentAnalysis.analysis().summaryTable();
EXPECT_EQ(1u,summary.nRows()); // no points
}
// ETH@20120120 At first tried just using the same database, analysisdriver, etc., but
// it did not go well. Restarting an initially failed Dakota analysis should be part of
// enabling Dakota restart more generally.
{
// algorithm rounds samples up to next one that fits n**(problem.numVariables())
algOptions.setSamples(6);
analysis = Analysis("DDACE Grid Sampling - Wrong Samples",
problem,
DDACEAlgorithm(algOptions),
seedModel);
ProjectDatabase database = getCleanDatabase("DDACEGridSampling_WrongSamples");
AnalysisDriver analysisDriver = AnalysisDriver(database);
AnalysisRunOptions runOptions = standardRunOptions(analysisDriver.database().path().parent_path());
CurrentAnalysis currentAnalysis = analysisDriver.run(analysis,runOptions);
EXPECT_TRUE(analysisDriver.waitForFinished());
boost::optional<runmanager::JobErrors> jobErrors = currentAnalysis.dakotaJobErrors();
ASSERT_TRUE(jobErrors);
EXPECT_TRUE(jobErrors->errors().empty());
EXPECT_TRUE(analysisDriver.currentAnalyses().empty());
Table summary = currentAnalysis.analysis().summaryTable();
EXPECT_EQ(10u,summary.nRows()); // 9 points
summary.save(analysisDriver.database().path().parent_path() / toPath("summary.csv"));
BOOST_FOREACH(const DataPoint& dataPoint,analysis.dataPoints()) {
EXPECT_TRUE(dataPoint.isComplete());
EXPECT_FALSE(dataPoint.failed());
// EXPECT_FALSE(dataPoint.responseValues().empty());
}
}
{
// algorithm rounds samples up to next one that fits n**(problem.numVariables())
algOptions.setSamplesForGrid(2,problem);
analysis = Analysis("DDACE Grid Sampling - Correct Samples",
problem,
DDACEAlgorithm(algOptions),
seedModel);
ProjectDatabase database = getCleanDatabase("DDACEGridSampling_CorrectSamples");
AnalysisDriver analysisDriver = AnalysisDriver(database);
AnalysisRunOptions runOptions = standardRunOptions(analysisDriver.database().path().parent_path());
CurrentAnalysis currentAnalysis = analysisDriver.run(analysis,runOptions);
EXPECT_TRUE(analysisDriver.waitForFinished());
boost::optional<runmanager::JobErrors> jobErrors = currentAnalysis.dakotaJobErrors();
ASSERT_TRUE(jobErrors);
EXPECT_TRUE(jobErrors->errors().empty());
EXPECT_TRUE(analysisDriver.currentAnalyses().empty());
Table summary = currentAnalysis.analysis().summaryTable();
EXPECT_EQ(5u,summary.nRows()); // 4 points
summary.save(analysisDriver.database().path().parent_path() / toPath("summary.csv"));
BOOST_FOREACH(const DataPoint& dataPoint,analysis.dataPoints()) {
EXPECT_TRUE(dataPoint.isComplete());
EXPECT_FALSE(dataPoint.failed());
// EXPECT_FALSE(dataPoint.responseValues().empty());
}
}
}
TEST_F(AnalysisDriverFixture, DDACE_OrthogonalArray_Continuous) {
// RETRIEVE PROBLEM
Problem problem = retrieveProblem("Continuous",true,false);
// DEFINE SEED
Model model = model::exampleModel();
openstudio::path p = toPath("./example.osm");
model.save(p,true);
FileReference seedModel(p);
// CREATE ANALYSIS
DDACEAlgorithmOptions algOptions(DDACEAlgorithmType::oas);
Analysis analysis("DDACE Orthogonal Array Sampling",
problem,
DDACEAlgorithm(algOptions),
seedModel);
// RUN ANALYSIS
{
ProjectDatabase database = getCleanDatabase("DDACEOrthogonalArray_NoSamples");
AnalysisDriver analysisDriver(database);
AnalysisRunOptions runOptions = standardRunOptions(analysisDriver.database().path().parent_path());
CurrentAnalysis currentAnalysis = analysisDriver.run(analysis,runOptions);
EXPECT_TRUE(analysisDriver.waitForFinished());
boost::optional<runmanager::JobErrors> jobErrors = currentAnalysis.dakotaJobErrors();
ASSERT_TRUE(jobErrors);
EXPECT_FALSE(jobErrors->errors().empty()); // require specification of number of samples
EXPECT_TRUE(analysisDriver.currentAnalyses().empty());
Table summary = currentAnalysis.analysis().summaryTable();
EXPECT_EQ(1u,summary.nRows()); // no points
}
{
algOptions.setSamples(6); // symbols = 0
analysis = Analysis("DDACE Orthogonal Array Sampling - Wrong Samples",
problem,
DDACEAlgorithm(algOptions),
seedModel);
ProjectDatabase database = getCleanDatabase("DDACEOrthogonalArray_WrongSamples");
AnalysisDriver analysisDriver = AnalysisDriver(database);
AnalysisRunOptions runOptions = standardRunOptions(analysisDriver.database().path().parent_path());
CurrentAnalysis currentAnalysis = analysisDriver.run(analysis,runOptions);
EXPECT_TRUE(analysisDriver.waitForFinished());
boost::optional<runmanager::JobErrors> jobErrors = currentAnalysis.dakotaJobErrors();
ASSERT_TRUE(jobErrors);
EXPECT_TRUE(jobErrors->errors().empty());
EXPECT_TRUE(analysisDriver.currentAnalyses().empty());
Table summary = currentAnalysis.analysis().summaryTable();
EXPECT_EQ(10u,summary.nRows()); // 9 points
summary.save(analysisDriver.database().path().parent_path() / toPath("summary.csv"));
BOOST_FOREACH(const DataPoint& dataPoint,analysis.dataPoints()) {
EXPECT_TRUE(dataPoint.isComplete());
EXPECT_FALSE(dataPoint.failed());
// EXPECT_FALSE(dataPoint.responseValues().empty());
}
}
{
bool ok = algOptions.setSamplesAndSymbolsForOrthogonalArray(6,1);
EXPECT_FALSE(ok);
ok = algOptions.setSamplesAndSymbolsForOrthogonalArray(2,2); // 2,3 not ok for no apparent reason
EXPECT_TRUE(ok);
ASSERT_TRUE(algOptions.symbols());
EXPECT_EQ(2,algOptions.symbols().get());
ASSERT_TRUE(algOptions.samples());
EXPECT_EQ(8,algOptions.samples().get());
analysis = Analysis("DDACE Orthogonal Array Sampling - Correct Samples",
problem,
DDACEAlgorithm(algOptions),
seedModel);
ProjectDatabase database = getCleanDatabase("DDACEOrthogonalArray_CorrectSamples");
AnalysisDriver analysisDriver = AnalysisDriver(database);
AnalysisRunOptions runOptions = standardRunOptions(analysisDriver.database().path().parent_path());
CurrentAnalysis currentAnalysis = analysisDriver.run(analysis,runOptions);
EXPECT_TRUE(analysisDriver.waitForFinished());
boost::optional<runmanager::JobErrors> jobErrors = currentAnalysis.dakotaJobErrors();
ASSERT_TRUE(jobErrors);
EXPECT_TRUE(jobErrors->errors().empty());
EXPECT_TRUE(analysisDriver.currentAnalyses().empty());
Table summary = currentAnalysis.analysis().summaryTable();
EXPECT_EQ(9u,summary.nRows());
summary.save(analysisDriver.database().path().parent_path() / toPath("summary.csv"));
BOOST_FOREACH(const DataPoint& dataPoint,analysis.dataPoints()) {
EXPECT_TRUE(dataPoint.isComplete());
EXPECT_FALSE(dataPoint.failed());
}
}
}