COOM Suite

Python package to parse and solve product configuration problems specified in COOM using ASP.

Installation

pip install .

Usage

The COOM language is a domain-specific language for modelling product configuration problems. While currently geared towards ASP, the COOM Suite is intended to serve as a general workbench for experimentation with industrial-scale product configuration problems. It includes a (customizable) ANTLR v4 parser to convert COOM specifications into facts, and currently contains two ASP encodings for solving: one for clingo and one for hybrid solver fclingo.

In addition, a range of examples and a benchmark collection with four scalable benchmark sets is provided.

COOM language fragments

We define the following three COOM language fragments:

• COOM Core corresponds to a simple Constraint Satisfaction Problem (CSP). It mainly consists of table and Boolean constraints over discrete attributes and all variables are defined.
• COOM[P] adds partonomies and cardinalities on top of COOM Core.
• COOM[X] adds numeric variables and calculations (arithmetic expressions and aggregate functions) on top of COOM[P].

Examples and Benchmarks

Examples

The COOM Suite contains a range of (product) configuration examples encoded in the COOM language. They can be found in the examples/coom directory. We highlight here only the Bike collection, containing three examples in increasing complexity that correspond to the three COOM language fragments defined above: the Kids, City, and Travel Bike.

We also provide the corresponding (serialized and refined) facts to these examples under examples/asp.

Benchmarks

The COOM suite includes four scalable benchmark sets. More information can be found in the benchmarks directory.

Convert COOM to facts

To convert a COOM instance into a set of (serialized) facts run

coomsuite convert examples/coom/kids-bike.coom

By default, the facts are printed to the console.

Optionally, an output directory can be provided using option --output dir in which case the facts will be saved to the given directory.

Solving

To solve a COOM model using ASP run

coomsuite solve examples/coom/kids-bike.coom

This will convert the given COOM file to a set of (serialized) facts and solve them with the clingo encoding.

Possible options include:

• Using --output coom, the output facts will be converted into a (more readable) COOM format.

• For solving with solver fclingo, specify --solver fclingo.

• The usual clingo command-line arguments can be used, eg. to calculate all stable models and suppress printing models.

coomsuite solve examples/coom/kids-bike.coom 0 -q

• Note that the direct conversion of COOM into facts is a mere serialization of the COOM model. We make use of a preprocessing encoding to translate the "serialized" facts into a set of "refined" facts, capturing the essence of a configuration problem. To show the "refined" facts, run with option --show. This will print the facts to the console.

Extending the workbench

Customizing the parser

To get started have a look at the COOM grammar.

You can customize the conversion by modifying the ASP Visitor.

More information on the Python target of ANTLR v4 can be found here.

Add encodings

The encodings are stored in src/coomsuite/encodings/.

Loading of the encodings is handled with clingo's Application class. Modify this line to insert your own encoding. Note that you might also have to disable/modify the preprocessing encoding here.

Development

To improve code quality, we use nox to run linters, type checkers, unit tests, documentation and more. We recommend installing nox using pipx to have it available globally.

# install
python -m pip install pipx
python -m pipx install nox

# run all sessions
nox

# list all sessions
nox -l

# run individual session
nox -s session_name

# run individual session (reuse install)
nox -Rs session_name

Note that the nox sessions create editable installs. In case there are issues, try recreating environments by dropping the -R option. If your project is incompatible with editable installs, adjust the noxfile.py to disable them.

We also provide a pre-commit config to autoformat code upon commits. It can be set up using the following commands:

python -m pipx install pre-commit
pre-commit install