pyAgrum-nightly 2.3.1.9.dev202512261765915415__cp310-abi3-macosx_10_15_x86_64.whl

pyagrum/bnmixture/BNMixture.py

@@ -0,0 +1,464 @@
+from typing import Dict, List, Set
+from tempfile import TemporaryDirectory
+from zipfile import ZipFile
+
+############################################################################
+#   This file is part of the aGrUM/pyAgrum library.                        #
+#                                                                          #
+#   Copyright (c) 2005-2025 by                                             #
+#       - Pierre-Henri WUILLEMIN(_at_LIP6)                                 #
+#       - Christophe GONZALES(_at_AMU)                                     #
+#                                                                          #
+#   The aGrUM/pyAgrum library is free software; you can redistribute it    #
+#   and/or modify it under the terms of either :                           #
+#                                                                          #
+#    - the GNU Lesser General Public License as published by               #
+#      the Free Software Foundation, either version 3 of the License,      #
+#      or (at your option) any later version,                              #
+#    - the MIT license (MIT),                                              #
+#    - or both in dual license, as here.                                   #
+#                                                                          #
+#   (see https://agrum.gitlab.io/articles/dual-licenses-lgplv3mit.html)    #
+#                                                                          #
+#   This aGrUM/pyAgrum library is distributed in the hope that it will be  #
+#   useful, but WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,          #
+#   INCLUDING BUT NOT LIMITED TO THE WARRANTIES MERCHANTABILITY or FITNESS #
+#   FOR A PARTICULAR PURPOSE  AND NONINFRINGEMENT. IN NO EVENT SHALL THE   #
+#   AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER #
+#   LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,        #
+#   ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR  #
+#   OTHER DEALINGS IN THE SOFTWARE.                                        #
+#                                                                          #
+#   See LICENCES for more details.                                         #
+#                                                                          #
+#   SPDX-FileCopyrightText: Copyright 2005-2025                            #
+#       - Pierre-Henri WUILLEMIN(_at_LIP6)                                 #
+#       - Christophe GONZALES(_at_AMU)                                     #
+#   SPDX-License-Identifier: LGPL-3.0-or-later OR MIT                      #
+#                                                                          #
+#   Contact  : info_at_agrum_dot_org                                       #
+#   homepage : http://agrum.gitlab.io                                      #
+#   gitlab   : https://gitlab.com/agrumery/agrum                           #
+#                                                                          #
+############################################################################
+
+import os
+import pyagrum
+import shutil
+
+
+class IMixture:
+  _bns: Dict[str, pyagrum.BayesNet]
+  _weights: Dict[str, float]
+  _refBN: pyagrum.BayesNet
+  _refName: str
+
+  def __init__(self):
+    raise NotImplementedError("class IMixture is not supposed to be initialized !")
+
+  def __repr__(self):
+    nameref = self._refBN.property("name")
+    ret = f"(reference BN : {nameref}), "
+    for name in self.names():
+      ret += "(" + name + ", w=" + str(self._weights[name]) + "), "
+    if len(ret) > 1:
+      ret = ret[:-2]
+    return ret
+
+  def names(self) -> List[str]:
+    """
+    Returns
+    -------
+    List[str]
+        The list of names of the BNs in the model (reference BN not included).
+    """
+    return list(self._bns.keys()).copy()
+
+  def size(self) -> int:
+    """
+    Returns
+    -------
+    int
+        The number of BNs in the model (reference BN not included).
+    """
+    return len(self._bns)
+
+  def setWeight(self, name: str, w: float):
+    """
+    Changes the weight of a BN in the model.
+
+    Parameters
+    ----------
+    name : str
+        Name of the BN to modify.
+    w : float
+        Value of the new weight.
+
+    Raises
+    ------
+    pyagrum.InvalidArgument
+        If the weight is negative.
+    pyagrum.NotFound
+        If the given name doesn't correspond to the name of a BN in the model.
+    """
+    if w < 0:
+      raise pyagrum.InvalidArgument("weight is negative !")
+    if name not in self._bns:
+      raise pyagrum.NotFound(f"'{name}' isn't in the model !")
+    self._weights[name] = w
+
+  def weight(self, name: str) -> float:
+    """
+    Parameters
+    ----------
+    name : str
+        Name of the BN.
+
+    Returns
+    -------
+    float
+        The weight of the BN with name ``name``.
+
+    Raises
+    ------
+    pyagrum.NotFound
+        If the given name doesn't correspond to the name of a BN in the model.
+    """
+    if name not in self._bns:
+      raise pyagrum.NotFound(f"{name} isn't in the model")
+    return self._weights[name]
+
+  def weights(self) -> Dict[str, float]:
+    """
+    Returns
+    -------
+    Dict[str, float]
+        The weights of all the BNs in the model.
+    """
+    return self._weights.copy()
+
+  def add(self, name: str, bn: pyagrum.BayesNet, w: float = 1):
+    """
+    Adds a BN to the model. If the model doesn't have a reference BN when trying to add an element,
+    the BN (before adding new element) with highest weight becomes the new reference.
+
+    Parameters
+    ----------
+    name : str
+        Name of the BN to add.
+    bn : pyagrum.BayesNet
+        BN to add.
+    w : float
+        Weight of the BN.
+
+    Raises
+    ------
+    pyagrum.InvalidArgument
+        If the weight is negative.
+    pyagrum.InvalidArgument
+        If the names of the variables in the BN to add are differents from the one in the reference BN.
+    pyagrum.InvalidArgument
+        If the variables in the BN to add are differents from the one in the reference BN.
+    pyagrum.InvalidArgument
+        If the name in argument is the same as the reference BN's name.
+    pyagrum.ArgumentError
+        If the name in argument already exists for a different BN in the model.
+    """
+    if w < 0:
+      raise pyagrum.InvalidArgument("weight is negative !")
+    if name == self._refName:
+      raise pyagrum.InvalidArgument("Can't give name used by reference BN")
+
+    if self._refBN is None:
+      self._refBN = pyagrum.BayesNet(bn)
+      self._refBN.setProperty("name", "refBN")
+      # self._refTrueName = name
+      self._bns[name] = pyagrum.BayesNet(bn)
+      self._bns[name].setProperty("name", name)
+      self._weights[name] = w
+      return
+
+    if bn.names() != self._refBN.names():
+      raise pyagrum.InvalidArgument("variables names are different from the reference BN variables")
+    if set([bn.variable(name) for name in bn.names()]) != set(
+      [self._refBN.variable(name) for name in self._refBN.names()]
+    ):
+      raise pyagrum.InvalidArgument("variables are different from those in the reference BN")
+
+    if name in self._bns and bn == self._bns[name]:
+      self._weights[name] += w
+    elif name not in self._bns and bn in self._bns.values():
+      # logging.warning("bn already exists with a different name")
+      for tmp in self._bns:
+        if bn == self._bns[tmp]:
+          self._weights[tmp] += w
+          break
+    elif name not in self._bns and bn not in self._bns.values():
+      self._bns[name] = pyagrum.BayesNet(bn)
+      self._bns[name].setProperty("name", name)
+      self._weights[name] = w
+    else:
+      raise pyagrum.ArgumentError("name already exists for different bn")
+
+  def remove(self, name: str):
+    """
+    Removes a BN from the model.
+
+    Parameters
+    ----------
+    name : str
+        Name of the BN to remove.
+
+    Raises
+    ------
+    pyagrum.NotFound
+        If the given name doesn't correspond to the name of a BN in the model.
+    """
+    if name not in self._bns:
+      raise pyagrum.NotFound(f"'{name}' isn't in the model !")
+    self._bns.pop(name)
+    self._weights.pop(name)
+
+  def isNormalized(self) -> bool:
+    """
+    Checks if the model is normalized (the sum of the weights equals 1).
+    """
+    return round(sum(self._weights.values()), 14) == 1
+
+  def normalize(self):
+    """
+    Normalizes the weights.
+    """
+    total = sum(self._weights.values())
+    for name in self._bns.keys():
+      self.setWeight(name, self.weight(name) / total)
+
+  def isValid(self) -> bool:
+    """
+    Checks if all the weights are equal to 0. Valid if sum of the weights is not 0.
+
+    Returns
+    -------
+    bool
+        True if weights are valid. False otherwise.
+    """
+    return sum(self._weights.values()) != 0
+
+  def BN(self, name: str) -> "pyagrum.BayesNet":
+    """
+    Parameters
+    ----------
+    name : str
+        Name of the variable.
+
+    Returns
+    -------
+    pyagrum.BayesNet
+        A copy of the BN with name ``name`` in the model.
+
+    Raises
+    ------
+    pyagrum.NotFound
+        If the given name doesn't correspond to the name of a BN in the model.
+    """
+    if name not in self._bns:
+      raise pyagrum.NotFound(f"{name} isn't in the mixture")
+    res = pyagrum.BayesNet(self._bns[name])
+    res.setProperty("name", name)
+    return res
+
+  def BNs(self) -> List[pyagrum.BayesNet]:
+    """
+    Returns
+    -------
+    List[pyagrum.BayesNet]
+        A list containing a copy of all BNs in the model.
+    """
+    return [self.BN(name) for name in self.names()]
+
+  def zeroBNs(self) -> Set[str]:
+    """
+    Returns
+    -------
+    Set[str]
+        The names of the BNs in the model that have weight with value 0.
+    """
+    return {name for name in self.names() if self.weight(name) == 0}
+
+  def existsArc(self, a, b):
+    """
+    Counts the number of time arc ``a`` -> ``b`` appears among all BNs in the model.
+
+    Parameters
+    ----------
+    a : str | int
+        Tail of the arc.
+    b : str | int
+        Head of the arc.
+
+    Returns
+    -------
+    int
+        The number of time arc ``a`` -> ``b`` appears.
+    """
+    count = 0
+    for bn in self.BNs():
+      if bn.existsArc(a, b):
+        count += 1
+    return count
+
+  def variable(self, name: str):
+    """
+    Parameters
+    ----------
+    name : str
+        Name of the variable.
+
+    Returns
+    -------
+    pyagrum.LabelizedVariable
+        The corresponding variable.
+    """
+    if self._refBN is None:
+      raise pyagrum.NotFound("Reference BN is None")
+    return self._refBN.variable(name)
+
+  def saveBIF(self, fname: str):
+    """
+    Saves a Mixture in BIF format and zip it.
+
+    Parameters
+    ----------
+    fname : str
+        Name of the file (without extenstion).
+    """
+    with TemporaryDirectory("", fname, None) as temp_dir:
+      txt_name = f"{temp_dir}/weights.txt"
+      with open(txt_name, "w+") as f_txt:
+        for e in self.weights():
+          f_txt.write(f"{e}:{self.weight(e)}\n")
+      ref_name = "ref_" + self._refBN.property("name")
+      self._refBN.saveBIF(f"{temp_dir}/{ref_name}.bif")
+      for name in self.names():
+        file_name = f"{temp_dir}/{name}.bif"
+        self.BN(name).saveBIF(file_name)
+      shutil.make_archive(fname, "zip", temp_dir)
+
+
+class BNMixture(IMixture):
+  """
+  Experimental model. The class contains a list of BNs. Each BN has its own weight.
+  """
+
+  def __init__(self):
+    self._bns = {}
+    self._weights = {}
+    self._refBN = None
+    self._refName = "refBN"
+
+  def updateRef(self):
+    """
+    Updates the reference BN. The new reference BN is the one with maximum weight.
+    """
+    highest = max(list(self._weights.keys()), key=lambda x: self._weights[x])
+    newRef = pyagrum.BayesNet(self._bns[highest])
+    newRef.setProperty("name", self._refName)
+    # self._refTrueName = highest
+    self._refBN = newRef
+
+  @staticmethod
+  def loadBIF(filename):
+    """
+    Retrieve a BNMixture from a file.
+
+    Parameters
+    ----------
+    filename : str
+        Zip file containing the mixture.
+
+    Returns
+    -------
+    BNMixture
+        The stored BNMixture
+    """
+    with TemporaryDirectory("", "tmp", None) as temp_dir:
+      # temp_dir = mkdtemp("", "tmp", None)
+      with ZipFile(filename, "r") as zf:
+        zf.extractall(temp_dir)
+      weights = {}
+      with open(f"{temp_dir}/weights.txt", "r") as wf:
+        for line in wf:
+          name, w = line.split(":")
+          weights[name] = float(w)
+      bnm = BNMixture()
+      for path in os.listdir(temp_dir):
+        name, _ = path.split(".")
+        if name in weights:
+          bni = pyagrum.BayesNet()
+          bni.loadBIF(f"{temp_dir}/{path}")
+          bnm.add(name, bni, w=weights[name])
+      bnm.updateRef()
+      return bnm
+
+
+class BootstrapMixture(IMixture):
+  """
+  Experimental model base on bootstraping. The class contains reference BN and a list of BNs. Each BN has its own weight except for the reference BN.
+  The reference BN is used so that every other BN added later contains the same variables as the reference BN.
+
+  Parameters
+  ----------
+  name : str
+      Name of the first BN to add to the model. It is used for reference.
+  bn : pyagrum.BayesNet
+      BN to add. Adding new BNs to the model is allowed only if the variables are the same in the first and new BN.
+  """
+
+  def __init__(self, name: str, bn: pyagrum.BayesNet):
+    self._refBN = pyagrum.BayesNet(bn)
+    self._refBN.setProperty("name", name)
+    self._bns = {}
+    self._weights = {}
+    self._refName = name
+
+  @staticmethod
+  def loadBIF(filename):
+    """
+    Retrieve a BootstrapMixture from a file.
+
+    Parameters
+    ----------
+    filename : str
+        Zip file containing the mixture.
+
+    Returns
+    -------
+    BootstrapMixture
+        The stored BootstrapMixture.
+    """
+    with TemporaryDirectory("", "tmp", None) as temp_dir:
+      with ZipFile(filename, "r") as zf:
+        zf.extractall(temp_dir)
+      weights = {}
+      with open(f"{temp_dir}/weights.txt", "r") as wf:
+        for line in wf:
+          name, w = line.split(":")
+          weights[name] = float(w)
+      bn_dict = {}
+      refBN = None
+      for path in os.listdir(temp_dir):
+        name, form = path.split(".")
+        if form != "bif":
+          continue
+        bni = pyagrum.BayesNet()
+        bni.loadBIF(f"{temp_dir}/{path}")
+        if name in weights:
+          bn_dict[name] = bni
+          bn_dict[name].setProperty("name", name)
+        else:
+          refBN = bni
+          refBN.setProperty("name", name)
+      bnm = BootstrapMixture(refBN.property("name"), refBN)
+      for name in bn_dict:
+        bnm.add(name, bn_dict[name], weights[name])
+      return bnm

pyagrum/bnmixture/__init__.py

@@ -0,0 +1,60 @@
+############################################################################
+#   This file is part of the aGrUM/pyAgrum library.                        #
+#                                                                          #
+#   Copyright (c) 2005-2025 by                                             #
+#       - Pierre-Henri WUILLEMIN(_at_LIP6)                                 #
+#       - Christophe GONZALES(_at_AMU)                                     #
+#                                                                          #
+#   The aGrUM/pyAgrum library is free software; you can redistribute it    #
+#   and/or modify it under the terms of either :                           #
+#                                                                          #
+#    - the GNU Lesser General Public License as published by               #
+#      the Free Software Foundation, either version 3 of the License,      #
+#      or (at your option) any later version,                              #
+#    - the MIT license (MIT),                                              #
+#    - or both in dual license, as here.                                   #
+#                                                                          #
+#   (see https://agrum.gitlab.io/articles/dual-licenses-lgplv3mit.html)    #
+#                                                                          #
+#   This aGrUM/pyAgrum library is distributed in the hope that it will be  #
+#   useful, but WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,          #
+#   INCLUDING BUT NOT LIMITED TO THE WARRANTIES MERCHANTABILITY or FITNESS #
+#   FOR A PARTICULAR PURPOSE  AND NONINFRINGEMENT. IN NO EVENT SHALL THE   #
+#   AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER #
+#   LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,        #
+#   ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR  #
+#   OTHER DEALINGS IN THE SOFTWARE.                                        #
+#                                                                          #
+#   See LICENCES for more details.                                         #
+#                                                                          #
+#   SPDX-FileCopyrightText: Copyright 2005-2025                            #
+#       - Pierre-Henri WUILLEMIN(_at_LIP6)                                 #
+#       - Christophe GONZALES(_at_AMU)                                     #
+#   SPDX-License-Identifier: LGPL-3.0-or-later OR MIT                      #
+#                                                                          #
+#   Contact  : info_at_agrum_dot_org                                       #
+#   homepage : http://agrum.gitlab.io                                      #
+#   gitlab   : https://gitlab.com/agrumery/agrum                           #
+#                                                                          #
+############################################################################
+
+""" """
+
+__author__ = "Pierre-Henri Wuillemin, Rodolphe Tavernier"
+__copyright__ = "(c) 2022-2023 PARIS"
+
+from .BNMixture import BNMixture, BootstrapMixture, IMixture
+from .BNMInference import BNMixtureInference, BootstrapMixtureInference, IMixtureInference
+from .BNMLearning import BNMLearner, BNMBootstrapLearner, IMixtureLearner
+
+__all__ = [
+  "BNMixture",
+  "BootstrapMixture",
+  "IMixture",
+  "BNMixtureInference",
+  "BootstrapMixtureInference",
+  "IMixtureInference",
+  "BNMLearner",
+  "BNMBootstrapLearner",
+  "IMixtureLearner",
+]