Shop2ASS

Semi-automatic conversion of lisp code from the Shop system for Answer Set Solving with Gringo and Clasp.

Clone this git repository by

$ git clone https://github.com/arksch/Shop2ASS.git

You can then see the options

$ cd Shop2ASS
$ python conversion.py --help

For usual input and output type

$ python conversion.py input_file output_file

The script will then ask you to give the right variable domains for :operators from the given Shop file. You can also add prerequisites that are implicitly used by the :methods that call these operators.

Note that this script does not translate :methods, yet.

For translating problem descriptions try to have all atoms, initial states and goals in one block each, with no leading and trailing paranthesis in the same lines. Then call

$ python conversion.py input_file output_file --atoms startline:endline --initials startline:endline --goals startline:endline

Note that counting starts with zero and each last line is not included.

Feel free to change the script to include more edge cases of the SHOP system. You can test that you did not break anything

$ py.test test_conversion.py

How to start the ASS?

This script only runs semi-automatically. You will have to alter some of the output to apply to Gringo/Clasp logic and syntax. Once you have tuned the output of the main problem and the problem's instance you can run it

$ ./path/to/gringo problem instance | ./path/to/clasp 0

Note that this will not run fast, since the included stub only performs a very dumb guessing. To test the bad performance you can run the example file

$ ./path/to/gringo Shop_solutions/blocks_test_runtime.lp | ./path/to/clasp 0

To improve the solving speed you should look at the Shop methods and probably include some additional goals and states.

For questions write me a mail.