new dt implementations

Author: Jorge Ejarque

rom_pillar_I/reduce_order_model/env.sh

@@ -2,3 +2,7 @@ export PYTHONPATH=/reduce_order_model/src:/opt/view/lib/python3/dist-packages/:$
 export LD_LIBRARY_PATH=/opt/view/libs:$LD_LIBRARY_PATH
 export PYTHONPATH=/opt/view/:$PYTHONPATH
 export ComputingUnits=8
+export KRATOS_CUS=8
+export QR_CUS=16
+export SVD_CUS=16
+export SW_CATALOG=/software-catalog/packages/

rom_pillar_I/reduce_order_model/src/ROM_workflow.py

@@ -0,0 +1,152 @@
+# Importing the Kratos Library
+import sys
+import time
+import os
+import KratosMultiphysics
+
+# Import packages
+import numpy as np
+
+# Import pickle for serialization
+import pickle
+
+# Import pycompss
+from pycompss.api.task import task
+from pycompss.api.constraint import constraint
+from pycompss.api.api import compss_wait_on, compss_barrier
+from pycompss.api.parameter import *
+from pycompss.api.api import compss_barrier
+from pycompss.api.software import software
+from pycompss.api.data_transformation import *
+
+from dts import *
+
+SW_CATALOG = os.environ.get("SW_CATALOG","/software-catalog/packages")
+
+# Workflows constants
+TotalNumberOFCases = 5
+number_of_dofs = 604264
+snapshots_per_simulation = 11
+number_of_columns = TotalNumberOFCases * snapshots_per_simulation
+expected_shape = (number_of_dofs,number_of_columns)  # We will know the size of the array!
+row_splits = 10
+column_splits = 1
+A_row_chunk_size = int(number_of_dofs / row_splits)
+A_column_chunk_size = int(number_of_columns / column_splits)
+desired_block_size = (A_row_chunk_size, A_column_chunk_size)
+simulation_block_size = (number_of_dofs, snapshots_per_simulation)
+desired_rank=30
+
+@software(config_file = SW_CATALOG+"/kratos/fom.json")
+def execute_FOM_instance(model,parameters, sample):
+    import KratosMultiphysics
+    from kratos_simulations import GetTrainingData
+    current_model = KratosMultiphysics.Model()
+    model.Load("ModelSerialization",current_model)
+    del(model)
+    current_parameters = KratosMultiphysics.Parameters()
+    parameters.Load("ParametersSerialization",current_parameters)
+    del(parameters)
+    # get sample
+    simulation = GetTrainingData(current_model,current_parameters,sample)
+    simulation.Run()
+    return simulation.GetSnapshotsMatrix()
+
+
+@dt(target="rom", function=ROM_file_generation, type=OBJECT_TO_FILE, destination=sys.argv[3])
+@software(config_file = SW_CATALOG + "/kratos/rom.json")
+def execute_ROM_instance(model,parameters,sample,rom):
+    import KratosMultiphysics 
+    from kratos_simulations import RunROM_SavingData
+    load_ROM(rom)
+    current_model = KratosMultiphysics.Model()
+    model.Load("ModelSerialization",current_model)
+    del(model)
+    current_parameters = KratosMultiphysics.Parameters()
+    parameters.Load("ParametersSerialization",current_parameters)
+    del(parameters)
+    # get sample
+    simulation = RunROM_SavingData(current_model,current_parameters,sample)
+    simulation.Run()
+    return simulation.GetSnapshotsMatrix()
+
+
+@software(config_file = SW_CATALOG+"/kratos/model.json")
+def load_model_parameters(model_file):
+    import KratosMultiphysics
+    from kratos_simulations import GetTrainingData
+    with open(model_file,'r') as parameter_file:
+        parameters = KratosMultiphysics.Parameters(parameter_file.read())
+    model = KratosMultiphysics.Model()
+    fake_sample = [5]
+    simulation = GetTrainingData(model,parameters,fake_sample)
+    serialized_model = KratosMultiphysics.StreamSerializer()
+    serialized_model.Save("ModelSerialization",simulation.model)
+    serialized_parameters = KratosMultiphysics.StreamSerializer()
+    serialized_parameters.Save("ParametersSerialization",simulation.project_parameters)
+    return serialized_model,serialized_parameters
+
+@dt("blocks", load_blocks_rechunk, shape=expected_shape, block_size=simulation_block_size,
+    new_block_size=desired_block_size, is_workflow=True)
+@software(config_file = SW_CATALOG + "/dislib/dislib.json")
+def rSVD(blocks, desired_rank=30):
+    from dislib_randomized_svd import rsvd
+    u,s = rsvd(blocks, desired_rank, A_row_chunk_size, A_column_chunk_size)
+    return u
+
+
+@dt("SnapshotsMatrixROM", load_blocks_rechunk, shape=expected_shape, block_size=simulation_block_size,
+    new_block_size=desired_block_size, is_workflow=True)
+@dt("SnapshotsMatrixFOM", load_blocks_rechunk, shape=expected_shape, block_size=simulation_block_size,
+    new_block_size=desired_block_size, is_workflow=True)
+@software(config_file = SW_CATALOG + "/dislib/dislib.json")
+def compare_ROM_vs_FOM(SnapshotsMatrixROM, SnapshotsMatrixFOM):
+    import dislib as ds
+    import numpy as np
+    #using the Frobenious norm of the snapshots of the solution
+    original_norm= np.linalg.norm((SnapshotsMatrixFOM.norm().collect()))
+    intermediate = ds.data.matsubtract(SnapshotsMatrixROM, SnapshotsMatrixFOM) #(available on latest release)
+    intermediate = np.linalg.norm((intermediate.norm().collect()))
+    final = intermediate/original_norm
+    np.save('relative_error_rom.npy', final)
+
+if __name__ == '__main__':
+
+    data_path = sys.argv[1]
+    parameters_template = sys.argv[2]
+    rom_file = sys.argv[3]
+    model_file="ProjectParameters_run.json"
+    replace_template(parameters_template, model_file, '%MODEL_PATH%', data_path)    
+
+    """
+    Here we define the parameters for the simulation.
+    In this case a sinlge parameter is defined.
+    More parameters are possible.
+    """
+    sim_cfgs = range(5,10)
+    model, parameters = load_model_parameters(model_file)
+    """
+    Stage 1
+    - launches in parallel a Full Order Model (FOM) simulation for each simulation parameter.
+    """
+    sim_results=[]
+    for cfg in sim_cfgs:
+        sim_results.append(execute_FOM_instance(model,parameters,[cfg]))
+    """
+    Stage 2
+    - computes the "fixed rank" randomized SVD in parallel using the dislib library  #TODO implement the fixed presicion RSVD
+    """
+    rom = rSVD(sim_results, desired_rank)
+    """
+    Stage 3
+    - launches the Reduced Order Model simulations for the same simulation parameters used for the FOM
+    """
+    rom_results=[]
+    for cfg in sim_cfgs:
+        sim_results.append(execute_ROM_instance(model,parameters,[cfg],rom))
+    
+    #compare_ROM_vs_FOM(rom_results, sim_results)
+
+
+
+