Explainability signal

.gitignore

@@ -135,3 +135,8 @@ dmypy.json
 .idea/
 
 .vscode/
+
+# explainer Stuff
+explainer/test.py
+explainer/old_code.py
+tutorial/conf.py

bpmnconstraints/script.py

@@ -1,4 +1,5 @@
 """Entry point for bpmnsignal command. Verifies argument and runs parser."""
+
 # pylint: disable=import-error
 import argparse
 import logging

bpmnconstraints/utils/plot.py

@@ -1,6 +1,7 @@
 """
 Module for plotting functions.
 """
+
 # pylint: disable=import-error
 
 import matplotlib.pyplot as plt

explainer/README.md

@@ -0,0 +1,15 @@
+# Symbolic Explanations of Process Conformance Violations
+This module is made by Marcus Rost, for his thesis subject, **Symbolic Explanations for Control-Flow Objects**. 
+
+The code should work properly, with some hard-coded parts that requires the data set for the signal code to function. This module is more seen as a proof of concept, where some core functions from explainability theory is utilized.
+
+## Overview
+The module mainly consists of 2 classes, ExplainerRegex and ExplainerSignal.
+
+**ExplainerRegex** is meant to simulate a real event log and system, with traces and constraints, using regex patterns.
+
+**ExplainerSignal** uses real event logs to calculate the explanations. This uses SAP Signavio's API to do so, but should in reality have it's own interpreter for the Signal queries, instead of having to make API calls. 
+
+Currently, the Signal code is sort of bad and specialized for the specific data set used. Hopefully, someone can see the value of what it done, and generalize it further.
+
+There is also 2 Notebook's located in `explainer/tutorial`, but I recommend doing `bpmn2constraints/tutorial/tutorial.ipynb` first, to see how to do the configuration for the API works.