Coral updates

commands.txt

@@ -0,0 +1,3 @@
+cd C:\DanielBaron\GenerativeLSTM
+conda activate deep_generator
+python dg_prediction.py

dg_prediction.py

@@ -0,0 +1,166 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Tue Feb 23 19:08:25 2021
+
+@author: Manuel Camargo
+"""
+import os
+import subprocess
+
+os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'
+import sys
+import getopt
+import shutil
+
+from model_prediction import model_predictor as pr
+import support_functions as sf
+from support_modules import stochastic_model as sm
+from support_modules import models_merger as mm
+
+# =============================================================================
+# Main function
+# =============================================================================
+def catch_parameter(opt):
+    """Change the captured parameters names"""
+    switch = {'-h': 'help', '-a': 'activity', '-c': 'folder',
+              '-b': 'model_file', '-v': 'variant', '-r': 'rep'}
+    return switch.get(opt)
+
+def call_simod(file_name):
+    print('----------------------------------------------------------------------')
+    print('-------------------     RUNNING SIMOD    -----------------------------')
+    print('----------------------------------------------------------------------')
+
+    simod_files = os.listdir(os.path.join('input_files', 'simod'))
+
+    #Copy event log file to Simod
+    source_file = os.path.join('input_files', file_name)
+    destination_file = os.path.join('..', 'Simod-2.3.1','inputs', file_name)
+    shutil.copy(source_file, destination_file)
+
+    if file_name not in simod_files:
+
+        os.chdir('../Simod-2.3.1/')
+        bash_command = 'bash.sh'
+        subprocess.run([bash_command, file_name], shell=True)
+        os.chdir('../GenerativeLSTM/')
+
+    #Delete event log file from Simod
+    os.remove(destination_file)
+
+def call_spmd(parameters):
+    print('----------------------------------------------------------------------')
+    print('--------------  RUNNING Stochastic Process Model  --------------------')
+    print('----------------------------------------------------------------------')
+    settings = dict()
+    settings['timeformat'] = parameters['read_options']['timeformat']
+    settings['column_names'] = parameters['read_options']['column_names']
+    settings['one_timestamp'] = parameters['read_options']['one_timestamp']
+    settings['filter_d_attrib'] = parameters['read_options']['filter_d_attrib']
+
+    settings['file'] = parameters['filename'].split('.')[0]
+    settings['sm3_path'] = parameters['sm3_path']
+
+    settings['bimp_path'] = parameters['bimp_path']
+    settings['concurrency'] = parameters['concurrency']
+    settings['epsilon'] = parameters['epsilon']
+    settings['eta'] = parameters['eta']
+
+    settings['log_path'] = os.path.join('input_files', settings['file'] + '.xes')
+    settings['tobe_bpmn_path'] = os.path.join('input_files', 'spmd', settings['file'] + '.bpmn')
+
+    spmd = sm.StochasticModel(settings)
+    return spmd
+
+def call_merger(parameters, spmd):
+    print('----------------------------------------------------------------------')
+    print('---------------------------  RUNNING MERGER --------------------------')
+    print('----------------------------------------------------------------------')
+    settings = dict()
+    settings['file'] = parameters['filename'].split('.')[0]
+    settings['bimp_path'] = parameters['bimp_path']
+    settings['tobe_bpmn_path'] = os.path.join('input_files', 'spmd', settings['file'] + '.bpmn')
+    settings['asis_bpmn_path'] = os.path.join('input_files', 'simod', settings['file'] + '.bpmn')
+    settings['csv_output_path'] = os.path.join('output_files', 'simulation_stats', settings['file'] + '.csv')
+    settings['output_path'] = os.path.join('output_files', 'simulation_files', settings['file'] + '.bpmn')
+    settings['lrs'] = spmd.lrs
+
+    mod_mer = mm.MergeModels(settings)
+
+def main(argv):
+    parameters = dict()
+    column_names = {'Case ID': 'caseid',
+                    'Activity': 'task',
+                    'lifecycle:transition': 'event_type',
+                    'Resource': 'user'}
+    parameters['one_timestamp'] = False  # Only one timestamp in the log
+    parameters['include_org_log'] = False
+    parameters['read_options'] = {
+
+        #Production and Purchasing: "%Y-%m-%d %H:%M:%S%z"
+        #RunningExample: "%Y-%m-%d %H:%M:%S%z"
+        #ConsultaDataMining201618: "%Y-%m-%d %H:%M:%S%z"
+
+        'timeformat': "%Y-%m-%d %H:%M:%S%z",
+        'column_names': column_names,
+        'one_timestamp': parameters['one_timestamp'],
+        'filter_d_attrib': False}
+
+    parameters['filename'] = 'Production.xes'
+    parameters['input_path'] = 'input_files'
+
+    parameters['sm3_path'] = os.path.join('external_tools', 'splitminer3', 'bpmtk.jar')
+    parameters['bimp_path'] = os.path.join('external_tools', 'bimp', 'qbp-simulator-engine_with_csv_statistics.jar')
+    parameters['concurrency'] = 0.0
+    parameters['epsilon'] = 0.5
+    parameters['eta'] = 0.7
+
+    # Parameters settled manually or catched by console for batch operations
+    if not argv:
+        # predict_next, pred_sfx
+        parameters['activity'] = 'pred_log'
+
+        #PurchasingExample:'20230615_13B0567C_E061_48DD_834E_459450B6076F' 
+        #Production: '20230615_FF2C5479_8FD9_4E8A_9473_F298F8D2618D'
+        #RunningExample: '20230615_337FD5C6_4BEA_4D3C_A7B4_57E09596867E'
+        #ConsultaDataMining201618: '20230615_10C7EA49_AF7D_48B6_8039_E007B2F61885'
+
+        parameters['folder'] = '20230615_FF2C5479_8FD9_4E8A_9473_F298F8D2618D' 
+        parameters['model_file'] = parameters['filename'].split('.')[0] + '.h5'
+        parameters['log_name'] = parameters['model_file'].split('.')[0]
+        parameters['is_single_exec'] = False  # single or batch execution
+        # variants and repetitions to be tested Random Choice, Arg Max, Rules Based Random Choice, Rules Based Arg Max
+        parameters['variant'] = 'Rules Based Random Choice'
+        parameters['rep'] = 1
+    else:
+        # Catch parms by console
+        try:
+            opts, _ = getopt.getopt(argv, "ho:a:f:c:b:v:r:",
+                                    ['one_timestamp=', 'activity=', 'folder=',
+                                     'model_file=', 'variant=', 'rep='])
+            for opt, arg in opts:
+                key = catch_parameter(opt)
+                if key in ['rep']:
+                    parameters[key] = int(arg)
+                else:
+                    parameters[key] = arg
+        except getopt.GetoptError:
+            print('Invalid option')
+            sys.exit(2)
+    print(parameters['folder'])
+    print(parameters['model_file'])
+
+    # Call Simod
+    call_simod(parameters['filename'])
+
+    #Generative model prediction
+    #pr.ModelPredictor(parameters)
+
+    #Call SPMD
+    #spmd = call_spmd(parameters)
+
+    #Call Merger
+    #call_merger(parameters, spmd)
+
+if __name__ == "__main__":
+    main(sys.argv[1:])

dg_training.py

@@ -31,16 +31,18 @@ def main(argv):
                     'Resource': 'user'}
     parameters['one_timestamp'] = False  # Only one timestamp in the log
     parameters['read_options'] = {
-        'timeformat': '%Y-%m-%dT%H:%M:%S.%f',
+        #RunningExample: "%Y-%m-%dT%H:%M:%S.%fZ"
+        #PurchasingExample y Production: '%Y/%m/%d %H:%M:%S'
+        'timeformat': '%Y/%m/%d %H:%M:%S',
         'column_names': column_names,
         'one_timestamp': parameters['one_timestamp']}
     # Parameters settled manually or catched by console for batch operations
     if not argv:
         # Event-log filename
-        parameters['file_name'] = 'Production.csv'
+        parameters['file_name'] = 'ConsultaDataMining201618.xes'
         parameters['model_family'] = 'lstm'
         parameters['opt_method'] = 'bayesian'  # 'rand_hpc', 'bayesian'
-        parameters['max_eval'] = 1
+        parameters['max_eval'] = 10
     else:
         # Catch parms by console
         try:

environment.yml

@@ -9,11 +9,13 @@ dependencies:
   - nltk
   - scikit-learn
   - ipywidgets
+  - networkx
+  - xmltodict
   - pip:
     - hyperopt
     - jellyfish
     - keras
     - pm4py
     - tensorflow
     - opyenxes
-    - git+https://github.com/Mcamargo85/support_modules.git
+    - git+https://github.com/Mcamargo85/support_modules.git

external_tools/proconformance/readMe

@@ -0,0 +1,67 @@
+------------------------
+ProConformance v2.0
+------------------------
+
+REQUIREMENTS: 
+Java 8 or above
+
+WHAT IT DOES:
+	ProConformance v2.0 is a business process conformance checking tool for identifying and verbalising differences between a process model and an event log in a scalable way.
+
+HOW:
+	ProConformance v2.0 implements the conformance checking technique described in the paper 
+      "Scalable Conformance Checking of Business Processes" by D. Reißner, R. Conforti, M. Dumas, M. La Rosa 
+	and A. Armas-Cervantes.
+
+USAGE:
+	java -jar ProConformance2.jar [folder] [logfile] [modelfile] [doOneOptimal] [stateLimit] [materlialiseDataStructures]
+
+	Input : A process model in bpmn or pnml format and an event log in either mxml or xes format.
+	Output: The difference between the model and the event log in terms of control-flow.
+
+PARAMETERS:
+
+[folder]      						: Folder where all input files are located and where the output files will be created.
+[logFile]     					: Event log including extension (either .mxml or .xes).
+[modelFile]    					: Process model including extension (.pnml or bpmn).
+[doOneOptimal]  				: (optional) type “true” for applying the one optimal alignment variant.
+[stateLimit]						: (optional) specify a state limit for the algorithm, i.e. “10000”.
+[materialiseDataStructures]	: (optional) type “true” for materialising the underlying DAFSA, tales reachability graph and PSP Data structures.
+
+Parameters [folder], [logfile] and [modelfile] are compulsory.
+Parameters [doOneOptimal], [stateLimit] and [materialiseDataStructures] are optional, but expected in this order.
+If the algorithm should materialise data structures for the all optimal variant with no state limit, it can be specified by adding: “” “no” “true”. 
+
+Using the tool without any parameters will provide release information of the tool.
+
+OUTPUT:
+	The tool will create three output files in the specified [folder] with the following names and contents:
+		AlignmentStatistics.csv : Statistics of the alignment process for each case present in the event log. Averages are also reported.
+		CaseTypeAlignmentResults.csv : Alignment Results for each case type. Can be cross referenced with file AlignmentStatistics.csv to obtain all optimal alignments for each case.
+		BehavioralStatements.txt: List of natural language statements about higher level behavioural issues between the event log and the process model.
+
+	In Case the parameter [materialiseDataStructures] is set to “true”, the tool will in addition output three files representing the applied data structures:
+		[logfile]_dafsa.dot : DAFSA representation of the event log in the .dot format.
+		[modelfile]_taulessRG.dot : Tauless reachability graph representation of the specified process model in the .dot format.
+		[psp.dot] : The synchronised product automaton between the DAFSA and the tales reachability graph in the .dot format.
+
+	The .dot format (https://en.wikipedia.org/wiki/DOT_(graph_description_language)) represents graph structures.
+		Files in this format can be opened with the application graphviz (http:http://www.graphviz.org) for example.
+
+ASSUMPTIONS:
+	The input model should be formatted according to the pnml (http://www.pnml.org/) format.
+      	To check that the input model is correctly formatted, you can open the model using WoPeD.
+
+	The tool accepts models formatted according to the bpmn (http://http://www.bpmn.org) format.
+		The tool will internally convert the bpmn diagram to its equivalent Petri net representation.
+
+	The input log should be formatted according to the mxml or xes (OpenXES, http://www.xes-standard.org/) format.
+      	To check that the input log is correctly formatted, you can open the log using the ProM toolkit.
+
+	If the parameter [doOneOptimal] is anything else than “true”, the tool applies the all optimal variant.
+	The parameter [stateLimit] expects an integer value or the value “no”, if no state limit should be specified.
+	If the parameter [materialiseDataStructures] is anything else than “true”, the tool will not output the data structures.
+
+RELEASE NOTES:
+
+	1.0: initial version