We can use event calculus (EC) to model events and expectations in product tracking and veracity scenarios. This demonstration makes use of discrete event calculus (we have discrete periods of time) to exhibit some helpful capabilities of EC in an organ donation scenario.
See Report.md in the FurtherDocumentation folder for a summary of the Bursary project, and Notes.md in the same folder for notes made in Weeks 1 - 6 of the project.
- We can't evaluate multiple expectations of the same type which occur for the same actor at the same time - we will only get output for one of the expectations. This feature can potentially be turned off if necessary.
You can launch this repository as a Deepnote project by clicking here.
If you wish to run any of the notebooks, you will need to set the environment to the local Dockerfile, build the Dockerfile and run the initialization notebook (init.ipynb). You may need to recomment the initialisation notebook line sed -i '/jedi/d;/jupyter/d;' ./requirements.txt.
Note that I am currently experiencing very slow loading times for notebooks to be set up. There may be a missing installation step for Graphviz, however I haven't been able to check this with my account as a working environment is cached.
Note that while individual "View in Deepnote" buttons have been added to the notebooks listed below, the Deepnote Viewer appears to have problems displaying all of the notebooks in full in this format.
- AgentsAndEvents.ipynb should be read first. It contains the agent and event declarations for the organ transplant scenario, and provides some context. The prolog code from this notebook has been automatically saved to AE.pl in Deepnote.
- Expectations.ipynb provides an overview of the expectation rule examples covered.
- Individual expectation examples can be found in the Visualisation folder, which also includes images produced by using Graphviz to display text from query output.
- For the current clauses which support discrete event calculus with Agent:Event syntax, see dec:notation.pl.
- DEC_Changes.pl records the changes made to dec.pl during the 10 Week Bursary period.
- date_time.pl contains code from the Date time package by Falco Nogatz as using pack_install was not possible in Deepnote. Some Date time predicates are used by a notebook in the ExpectationExamples](ExpectationExamples) folder.
- The Dockerfile used to set up the Deepnote environment. As above, note that this may have a Graphviz dependency missing.
- For the Jswipl kernel, see jupyter-swi-prolog. This kernel uses code from Veracity Lab, but has additional support for running Python cells, saving prolog output and combining code from multiple cells into a single prolog file.
- The installation notebook.
- The requirements file which is used by the installation notebook.
- For a guide on how to use the JSwipl kernel in Deepnote, see using_jswipl.ipynb.