jkl-serialization

Python library for (de)serializing jkl data as used by the Bayesian Network Structure Learner GOBNILP and proposed by Jaakkola et al. in Learning Bayesian Network Structure using LP Relaxations.

Example

JKL String

3
0 4
-2.772589 2 1 2
-2.865831 0
-2.963209 1 2
-2.963209 1 1
1 4
-2.772589 2 0 2
-2.865831 0
-2.963209 1 2
-2.963209 1 0
2 4
-2.772589 2 0 1
-2.865831 0
-2.963209 1 1
-2.963209 1 0

Serialized Python Object

{
  '0': [('-2.772589', ['1', '2']),
        ('-2.865831', []),
        ('-2.963209', ['2']),
        ('-2.963209', ['1'])],
 '1': [('-2.772589', ['0', '2']),
       ('-2.865831', []),
       ('-2.963209', ['2']),
       ('-2.963209', ['0'])],
 '2': [('-2.772589', ['0', '1']),
       ('-2.865831', []),
       ('-2.963209', ['1']),
       ('-2.963209', ['0'])]
  }

Build Status