gbdc

GBDC: Global Benchmark Database, C++ Extension Module

GBDC provides efficient implementations of functions for benchmark instance identification, instance feature extraction and instance transformation. GBDC provides a command-line tool as well as the Python package gbdc. The Python package gbdc is used by Global Benchmark Database.

Documentation

GBDC provides benchmark instance identifiers, feature extractors, and instance transformers for several problem domains, including propositional satisfiability (SAT) and optimization (MaxSAT), as well as Pseudo-Boolean Optimization (PBO). A description of the supported domains, feature extractors, and instance transformers can be found in the documentation.

Installation from PyPI

• Pre-built distributions for Linux and MacOS.
• Requires at least Python 3.8.0 (3.10.0 for Apple Silicon).
• Installation via pip install gbdc

Installation from Source

• GBDC uses libarchive for reading from a large variety of compressed formats (in some systems provided by the package libarchive-dev).
• Some GBDC functions use an IPASIR SAT Solver. GBDC’s build-system pulls the external SAT Solver CaDiCaL by A. Biere (MIT licensed).

Steps:

1. Install Dependencies (libarchive, pybind, ninja)
   • For Ubuntu: apt install libarchive-dev pybind11-dev ninja-build
   • For macOS: brew install libarchive pybind11 ninja
2. Run pip install . --user in the repository directory. <!– #### Shipped Dependencies

• A copy of the command-line argument parser by P. S. Kumar argparse.h (MIT licensed) resides in the lib folder.

• A copy of the MD5 hash implementation by M. Lloyd (MIT licensed) resides in the lib folder. –>

This site is open source. Improve this page.