import warnings
warnings.filterwarnings("ignore", category=UserWarning)

Query Checking with the DECLARE Checker

This tutorial explains how to perform the query checking of a DECLARE constraint in a log and how to browse the results. We first import the classes DeclareQueryChecker and DeclareResultsBrowser and the log.

import os

from Declare4Py.ProcessMiningTasks.QueryChecking.DeclareQueryChecker import DeclareQueryChecker
from Declare4Py.ProcessMiningTasks.QueryChecking.DeclareResultsBrowser import DeclareResultsBrowser
from Declare4Py.D4PyEventLog import D4PyEventLog

log_path = os.path.join("../../../", "tests", "test_logs","Sepsis Cases.xes.gz")
event_log = D4PyEventLog(case_name="case:concept:name")
event_log.parse_xes_log(log_path)

The DeclareQueryChecker class takes as input:

• the event log;

• the boolean parameter consider_vacuity=true that means that vacuously satisfied traces are considered as satisfied, violated otherwise.

• The query/setting parameters template, activation, target, activation_condition, target_condition and time_condition that set the parameters of the DECLARE constraints and the variables to ask. If a parameter is not set then it is treated as a variable to ask.

• The float min_support parameter sets the support to be satisfied in the log by the variable assignments.

• the boolean parameter return_first that if set to false it returns all the variables assignments that satisfy the support in the log. Otherwise, it returns only one variables assignment (the first) that satisfy the support in the log. This saves computational time when one is only interested in the existence of a variable assignment with a given support.

In the following, we compute the assignments for the target by setting template_str='Chain Response', activation='IV Antibiotics', act_cond='A.org:group is A' and min_support=0.2:

After this setting, the method run of the DeclareQueryChecker class will perform the query checking.

query_checker = DeclareQueryChecker(log=event_log, consider_vacuity=False, template='Chain Response', activation='IV Antibiotics', activation_condition='A.org:group is A', min_support=0.2, return_first=False)
query_check_res: DeclareResultsBrowser = query_checker.run()

The query checking results can be filtered with the filter_query_checking method whose parameter queries takes as input a list of variables to ask. In the above example, we set the template and activation and ask for the targets. The output of filter_query_checking is a Pandas dataframe.

query_check_res.filter_query_checking(queries=['template', 'activation', 'target', 'activation_condition'])

	template	activation	target	activation_condition
0	Chain Response	ER Registration	ER Triage	A.org:group is A
1	Chain Response	ER Sepsis Triage	Leucocytes	A.org:group is A
2	Chain Response	ER Sepsis Triage	IV Liquid	A.org:group is A
3	Chain Response	IV Liquid	IV Antibiotics	A.org:group is A
4	Chain Response	IV Antibiotics	Admission NC	A.org:group is A

In the following example we compute the assignments both for the activation and the target by setting template_str='Response' and min_support=0.8:

query_checker = DeclareQueryChecker(log=event_log, consider_vacuity=False, template='Response', min_support=0.8, return_first=False)
query_check_res: DeclareResultsBrowser = query_checker.run()

query_check_res.filter_query_checking(queries=['template', 'activation', 'target'])

	template	activation	target
0	Response	ER Registration	Leucocytes
1	Response	ER Registration	CRP
2	Response	ER Registration	LacticAcid
3	Response	ER Registration	ER Triage
4	Response	ER Registration	ER Sepsis Triage
5	Response	ER Triage	Leucocytes
6	Response	ER Triage	CRP
7	Response	ER Triage	ER Sepsis Triage
8	Response	ER Sepsis Triage	Leucocytes
9	Response	ER Sepsis Triage	CRP

As last example, we set the activation and the target and ask for the template:

query_checker = DeclareQueryChecker(log=event_log, consider_vacuity=False, activation='ER Registration', target='CRP', min_support=0.2, return_first=False)
query_check_res: DeclareResultsBrowser = query_checker.run()

query_check_res.filter_query_checking(queries=['template', 'activation', 'target'])

	template	activation	target
0	Choice	ER Registration	CRP
1	Responded Existence	ER Registration	CRP
2	Response	ER Registration	CRP
3	Alternate Response	ER Registration	CRP
4	Precedence	ER Registration	CRP
5	Alternate Precedence	ER Registration	CRP
6	Not Chain Response	ER Registration	CRP
7	Not Chain Precedence	ER Registration	CRP