First working version for marginals and MPEs.

.github/workflows/publish-to-pypi.yml

@@ -1,4 +1,4 @@
-name: Publish Python 🐍 distribution 📦 to PyPI and TestPyPI
+name: Publish Python 🐍 distribution 📦 to PyPI
 
 on: push
 

requirements-dev.txt

@@ -1,3 +1 @@
-fglib==0.2.3
-portion>=2.4.1
-pydantic>=2.4.2
+fglib==0.2.3

requirements.txt

@@ -1,4 +1,5 @@
 networkx>=3.0
 numpy>=1.24.4
-random_events>=1.1.1
+random_events>=1.1.3
 tabulate>=0.9.0
+probabilistic-model>=1.1.0

src/fglib2/__init__.py

@@ -1 +1 @@
-__version__ = '1.1.2'
+__version__ = '1.1.3'

src/fglib2/distributions.py

@@ -7,9 +7,10 @@
 from random_events.events import Event, EncodedEvent
 
 import tabulate
+from probabilistic_model.probabilistic_model import ProbabilisticModel
 
 
-class Multinomial:
+class Multinomial(ProbabilisticModel):
     """
     A multinomial distribution over discrete random variables.
     """
@@ -27,7 +28,7 @@ class Multinomial:
     """
 
     def __init__(self, variables: Iterable[Discrete], probabilities: Optional[np.ndarray] = None):
-        self.variables = tuple(sorted(variables))
+        super().__init__(variables)
 
         shape = tuple(len(variable.domain) for variable in self.variables)
 
@@ -67,14 +68,6 @@ def _mode(self) -> Tuple[List[EncodedEvent], float]:
         mode = [EncodedEvent(zip(self.variables, event)) for event in events.tolist()]
         return mode, likelihood
 
-    def mode(self) -> Tuple[List, float]:
-        """
-        Calculate the most likely event.
-        :return: The mode of the distribution as Event and its likelihood.
-        """
-        mode, likelihood = self._mode()
-        return [mode_.decode() for mode_ in mode], likelihood
-
     def __copy__(self) -> 'Multinomial':
         """
         :return: a shallow copy of the distribution.
@@ -196,15 +189,6 @@ def _likelihood(self, event: List[int]) -> float:
         """
         return float(self.probabilities[tuple(event)])
 
-    def likelihood(self, event: List) -> float:
-        """
-        Calculate the likelihood of a full evidence query.
-        The event is a list of values for the variables in the same order
-        :param event:
-        :return: P(event)
-        """
-        return self._likelihood(self.encode(event))
-
     def _conditional(self, event: EncodedEvent) -> 'Multinomial':
         """
         Calculate the conditional distribution given an event encoded.