Project repository for a Python tool for checking and computing causes in concrete runs of real-time systems modeled in Uppaal.
For executing the tool, you must have installed
- a current Python version (3.10.0 or newer)
- Uppaal: https://uppaal.org/ (we run the tool with Uppaal version 4.1, but also other versions should be fine)
- the Python library pyuppaal: https://pypi.org/project/pyuppaal/1.0.0/
After downloading our project files, you have to set at first the verifyta-path (that is the system path for executing Uppaal's model checker) . Include the absolute path to verfifyta in the verifyta_path.txt file in the top folder of our project.
By default, this should be the following:
Windows: absolute_path_to_uppaal\bin-Windows\verifyta.exe
Linux : absolute_path_to_uppaal/bin-Linux/verifyta
macOS : absolute_path_to_uppaal/bin-Darwin/verifyta
Now, you should be able to execute the Python files from this project.
For running the main functionalities of the tool, execute the Python file causality_tool.py together with the following program arguments:
"-d" for checking a cause or "-f" for computing causes
"-c <cause_kind>", whereby <cause_kind> can be instantiated with
<cause_kind> = "b" for but-for causality,
<cause_kind> = "m" for minimal but-for causality, or
<cause_kind> = "a" for actual causality
"-s ", with the xml-file of a Uppaal system containing the automaton and effect query
"-t ", with the txt-file containing considered trace
"-e ", with the txt-file containing the considered set of events (only necessary to give as argument for the computation of causes)
For instance, the following commands performs minimal but-for cause testing for event set Example_3_1_1\Cause1:
python .\causality_tool.py -d -c m -s .\Thesis_Examples\Example_3_1_1\TA.xml -t .\Thesis_Examples\Example_3_1_1\Trace.txt -e .\Thesis_Examples\Example_3_1_1\Cause1.txt
Converseley, the following command performs actual cause computations for system .\Thesis_Examples\Example_3_1_1\TA.xml and trace .\Thesis_Examples\Example_3_1_1\Trace.txt.
python .\causality_tool.py -f -c a -s .\Thesis_Examples\Example_3_1_1\TA.xml -t .\Thesis_Examples\Example_3_1_1\Trace.txt
The experiment scripts used for measurements and literature examples should be executable without any arguments.
We support Uppaal systems only with one agent, enabling non-internal actions to be performed by dummy handshaking. The dummy handshaker might either be given directly as template already in the input file (the template must then have the name "Dummy_Handshaker") or is added by the tool itself. Besides this restriction, we do also not support parameters, shared global variables, further chaneltypes. Also, we only allow fixed clock assignments.
We refer also to the multiple examples for more details on the possible input forms for the Uppaal System.
We represent only the concrete actions and delays of a run (i.e., its trace). We can handle delay traces, delay lasso traces, and timestamp traces as well as the corresponding sets of events. See the traces and causes in Example_3_1_5 and Example_3_2_3 for more details on the input format.
"Counterfactual Explanations for MITL Violations". Bernd Finkbeiner, Felix Jahn, and Julian Siber. FSTTCS 2024.
The tool was orignally developed in the course of the master's project on "Counterfactual Causality in Real-Time Systems" by Felix Jahn, advised by Julian Siber, and supervised by Prof. Bernd Finkbeiner, Ph.D. at Saarland University. Causes are checked and computed relying on the formal notions of causality in real-time systems developed in this thesis.