Merge pull request #72 from randomir/fix-eid

Less restrictive EnergyImpactDecomposer, Tabu

hybrid/decomposers.py

@@ -49,7 +49,7 @@ class EnergyImpactDecomposer(Runnable, traits.ProblemDecomposer):
     Args:
         max_size (int):
             Maximum number of variables in the subproblem.
-        min_gain (int, optional, default=0):
+        min_gain (int, optional, default=-inf):
             Minimum reduction required to BQM energy, given the current sample. A variable
             is included in the subproblem only if inverting its sample value reduces energy
             by at least this amount.
@@ -65,7 +65,7 @@ class EnergyImpactDecomposer(Runnable, traits.ProblemDecomposer):
         See examples on https://docs.ocean.dwavesys.com/projects/hybrid/en/latest/reference/decomposers.html#examples.
     """
 
-    def __init__(self, max_size, min_gain=0.0, min_diff=1, stride=1):
+    def __init__(self, max_size, min_gain=None, min_diff=1, stride=1):
         super(EnergyImpactDecomposer, self).__init__()
 
         if min_diff > max_size or min_diff < 0:

requirements.txt

@@ -3,7 +3,7 @@ dimod>=0.7.7
 minorminer
 dwave-system
 dwave-neal>=0.4.1
-dwave-tabu>=0.1.2
+dwave-tabu>=0.1.3
 numpy
 networkx
 dwave-networkx>=0.6.6

setup.py

@@ -17,7 +17,7 @@
 # Package requirements, minimal pinning
 install_requires = ['six>=1.10', 'numpy', 'networkx', 'click>5', 'plucky>=0.4.3',
                     'dimod>=0.7.7', 'minorminer', 'dwave-system', 'dwave-networkx>=0.6.6',
-                    'dwave-neal>=0.4.1', 'dwave-tabu>=0.1.2']
+                    'dwave-neal>=0.4.1', 'dwave-tabu>=0.1.3']
 
 # Package extras requirements
 extras_require = {