58 page with prediction

bn_modeller/app.py

@@ -10,6 +10,15 @@
 
 
 def create_application(argv: list[str]) -> QApplication:
+    """
+    Create a QApplication instance.
+
+    Args:
+        argv (list[str]): List of command-line arguments.
+
+    Returns:
+        QApplication: A new QApplication instance.
+    """
     QCoreApplication.setOrganizationName("Digiratory")
     QCoreApplication.setOrganizationDomain("digiratory.ru")
     QCoreApplication.setApplicationName("bn_modeller")
@@ -25,6 +34,15 @@ def create_application(argv: list[str]) -> QApplication:
 
 
 def manage_cli_args(app: QCoreApplication) -> QCommandLineParser:
+    """
+    Manage command-line arguments.
+
+    Args:
+        app (QCoreApplication): QCoreApplication instance.
+
+    Returns:
+        QCommandLineParser: A new QCommandLineParser instance.
+    """
     cl_parser: QCommandLineParser = QCommandLineParser()
     cl_parser.setApplicationDescription("Bayesian Network Modeller")
     cl_parser.addHelpOption()
@@ -37,6 +55,16 @@ def manage_cli_args(app: QCoreApplication) -> QCommandLineParser:
 
 
 if __name__ == "__main__":
+    """
+    Entry point of the application.
+    Args:
+        app (QCoreApplication): QCoreApplication instance.
+        cl_parser (QCommandLineParser): A new QCommandLineParser instance.
+
+    Raises:
+        ValueError: If headless mode is not supported.
+
+    """
     app = create_application(sys.argv)
     cl_parser = manage_cli_args(app)
 

bn_modeller/bayesian_nets/graph_preparation.py

@@ -1,3 +1,5 @@
+from typing import Callable, Tuple
+
 import networkx as nx
 
 
@@ -40,7 +42,7 @@ def __init__(self, corr_matrix, table_connection, threshold=0):
 
         self.G.add_weighted_edges_from(e)
 
-    def drop_cycle(self):
+    def drop_cycle(self, onRemoveEdgeHandler: Callable[[Tuple[int, int]], None] = None):
         while True:
             try:
                 cycle_list = nx.find_cycle(self.G)
@@ -57,17 +59,18 @@ def drop_cycle(self):
                     min_nodes = i
 
             self.G.remove_edge(*min_nodes)
+            if onRemoveEdgeHandler is not None:
+                onRemoveEdgeHandler(min_nodes)
 
     def getNodeList(self):
         return self.G.nodes
 
     def getEdgeList(self):
         return self.G.edges
 
-    def getGraph(self):
+    def getInternalGraph(self):
         return self.G
 
-    def renaming(self):
+    def getGraph(self) -> nx.Graph:
         k = {val: key for key, val in self.code_columns.items()}
-
         return nx.relabel_nodes(self.G, k)

bn_modeller/bayesian_nets/pyBansheeCalculation.py

@@ -47,6 +47,8 @@ def getRankCorr(self):
         return self.R
 
     def getInference(self, len_input_list):
+        # TODO: Rewrite this function to handle set of features and features to predoct as input.
+
         nodes = list(
             range(len_input_list)
         )  # all variables except for value of interest

bn_modeller/button_page.py

@@ -411,17 +411,17 @@ def on_acycle_graph(self):
         )
 
         # удалить циклы в графе
-        G_before = copy.deepcopy(self.graph.renaming())
+        G_before = copy.deepcopy(self.graph.getGraph())
 
         self.graph.drop_cycle()
         self.changeLinkTable()
 
-        d = {"Before": G_before, "After": self.graph.renaming()}
+        d = {"Before": G_before, "After": self.graph.getGraph()}
         # import pickle
         # pickle.dump(self.graph, open('graph.txt', 'w'))
         # print(self.graph.G.nodes())
 
-        nx.write_adjlist(self.graph.renaming(), "graph.txt")
+        nx.write_adjlist(self.graph.getGraph(), "graph.txt")
         dialog = SubplotGraph(data=d)
         self.dialogs.append(dialog)
         dialog.show()
@@ -508,7 +508,7 @@ def changeLinkTable(self):
             columns=self.input_df.columns, index=self.input_df.columns
         )
 
-        for i in self.graph.renaming().edges(data=True):
+        for i in self.graph.getGraph().edges(data=True):
             newLinkTab.loc[i[0], i[1]] = 1
         newLinkTab = newLinkTab.fillna(0)
         self.updLinkTable = newLinkTab

bn_modeller/models/bayesian_inference_model.py

@@ -0,0 +1,51 @@
+from PySide6.QtCore import (
+    QAbstractItemModel,
+    QModelIndex,
+    QObject,
+    QSortFilterProxyModel,
+    Qt,
+    Signal,
+    Slot,
+)
+
+
+class BayesianInferenceModel(QAbstractItemModel):
+    def __init__(self, parent: QObject = None):
+        super().__init__(parent)
+
+    def headerData(
+        self,
+        section: int,
+        orientation: Qt.Orientation,
+        role: Qt.ItemDataRole = Qt.ItemDataRole,
+    ):
+        if role == Qt.DisplayRole:
+            if orientation == Qt.Horizontal:
+                return f"Column {section}"
+            elif orientation == Qt.Vertical:
+                return f"Row {section}"
+        return super().headerData(section, orientation, role)
+
+    def rowCount(self, parent: QModelIndex = QModelIndex()) -> int:
+        if not parent.isValid():
+            return 10
+        return 5
+
+    def columnCount(self, parent: QModelIndex = QModelIndex()) -> int:
+        if not parent.isValid():
+            return None
+        return 2
+
+    def data(self, index: QModelIndex, role: Qt.ItemDataRole = Qt.ItemDataRole):
+        if not index.isValid():
+            return None
+        row = index.row()
+        column = index.column()
+        if role == Qt.DisplayRole:
+            return f"Row {row}, Column {column}"
+        return super().data(index, role)
+
+
+class BayesianInferenceIOOnlyModel(QSortFilterProxyModel):
+    def __init__(self, parent: QObject = None):
+        super().__init__(parent)

bn_modeller/widgets/bn_visualization_view.py

@@ -16,7 +16,7 @@ def __init__(self, parent=None, width=12, height=12, dpi=100):
         super().__init__(self.fig)
         self.bn_ax = self.fig.add_subplot(1, 1, 1)
 
-    def update_plot(self, graph):
+    def update_plot(self, graph: nx.Graph):
         self.bn_ax.clear()
         # self.bn_ax = self.fig.add_subplot(1, 1, 1)
 
@@ -116,7 +116,7 @@ def drawBN(self):
         graph.drop_cycle()
         # self.changeLinkTable()
 
-        self.bn_canvas.update_plot(graph.renaming())
+        self.bn_canvas.update_plot(graph.getGraph())
         # import pickle
         # pickle.dump(self.graph, open('graph.txt', 'w'))
         # print(self.graph.G.nodes())

bn_modeller/widgets/page/bayesian_inference_page.py

@@ -0,0 +1,33 @@
+from PySide6.QtCore import Qt
+from PySide6.QtWidgets import (
+    QGridLayout,
+    QHBoxLayout,
+    QLabel,
+    QSplitter,
+    QTableView,
+    QTabWidget,
+    QVBoxLayout,
+    QWidget,
+)
+
+
+class BayesianInferencePageWidget(QWidget):
+    def __init__(self, parent: QWidget | None = None, f=Qt.WindowType()):
+        super().__init__(parent, f)
+        self._init_ui()
+
+    def _init_ui(self):
+
+        self.mainLayout = QVBoxLayout(self)
+
+        # Top Layout with tables
+
+        topLayout = QHBoxLayout()
+        self.tableSource = QTableView()
+        self.tableTarget = QTableView()
+
+        topLayout.addWidget(self.tableSource)
+        topLayout.addWidget(self.tableTarget)
+
+        # Finalization
+        self.setLayout(self.mainLayout)

bn_modeller/widgets/page/bayesian_network_page.py

@@ -15,6 +15,7 @@
 )
 from bn_modeller.widgets import DependencySetupTableView, SelectableListView
 from bn_modeller.widgets.bn_visualization_view import BayesianNetView
+from bn_modeller.widgets.page.bayesian_inference_page import BayesianInferencePageWidget
 from bn_modeller.widgets.vertical_label import QVertivalLabel
 
 
@@ -30,13 +31,13 @@ def _init_ui(self):
 
         # Dependency tab
 
-        # Feature selection
+        ## Feature selection
         self.dependencyTabWidget = QSplitter()
         self.featureSelectorView = SelectableListView()
         self.dependencyTabWidget.addWidget(self.featureSelectorView)
         self.dependencyTabWidget.setStretchFactor(0, 1)
 
-        # Dependency Table
+        ## Dependency Table
         depTableWidget = QWidget()
         depTableLayout = QGridLayout()
 
@@ -60,6 +61,10 @@ def _init_ui(self):
 
         self.tabWidget.addTab(self.visualizationTabWidget, self.tr("Visulization"))
 
+        # Inference Tab
+        self.inferenceTabWidget = BayesianInferencePageWidget()
+        self.tabWidget.addTab(self.inferenceTabWidget, self.tr("Inference"))
+
         # Finalization
         self.mainLayout.addWidget(self.tabWidget)
         self.setLayout(self.mainLayout)