Behavior GUI

src/gui/rawdata_visualizer.py

@@ -18,6 +18,7 @@
 
 import iblphotometry.preprocessing as ffpr
 import numpy as np
+import pandas as pd
 
 
 class DataFrameVisualizerApp(QWidget):
@@ -30,6 +31,8 @@ def __init__(self):
 
         self.plot_time_index = None
 
+        self.behavior_gui = None
+
         self.filtered_df = None  # Filtered DataFrame used for plotting only
         self.init_ui()
 
@@ -40,7 +43,7 @@ def init_ui(self):
         # Layout for file loading and selection
         file_layout = QHBoxLayout()
         self.load_button = QPushButton('Load File', self)
-        self.load_button.clicked.connect(self.load_file)
+        self.load_button.clicked.connect(self.open_dialog)
         file_layout.addWidget(self.load_button)
 
         self.column_selector = QComboBox(self)
@@ -56,6 +59,10 @@ def init_ui(self):
         self.filter_selector.currentIndexChanged.connect(self.apply_filter)
         file_layout.addWidget(self.filter_selector)
 
+        self.behavior_button = QPushButton('Open behavior GUI', self)
+        self.behavior_button.clicked.connect(self.open_behavior_gui)
+        file_layout.addWidget(self.behavior_button)
+
         # # Table widget to display DataFrame
         # self.table = QTableWidget(self)
         # self.table.setSelectionMode(QTableWidget.SingleSelection)
@@ -95,47 +102,50 @@ def init_ui(self):
         self.setWindowTitle('DataFrame Plotter')
         self.setGeometry(300, 100, 800, 600)
 
-    def load_file(self):
+    def load_file(self, file_path):
+        try:
+            if (
+                file_path.endswith('.csv')
+                or file_path.endswith('.pqt')
+                or file_path.endswith('.parquet')
+            ):
+                self.dfs = from_raw_neurophotometrics_file(file_path)
+            else:
+                raise ValueError('Unsupported file format')
+
+            if 'GCaMP' in self.dfs.keys():
+                self.df = self.dfs['GCaMP']
+                self.times = self.dfs['GCaMP'].index.values
+                self.plot_time_index = np.arange(0, len(self.times))
+                self.filtered_df = None
+            else:
+                raise ValueError('No GCaMP found')
+
+            if 'Isosbestic' in self.dfs.keys():
+                self.dfiso = self.dfs['Isosbestic']
+
+            # Display the dataframe in the table
+            # self.display_dataframe()
+            # Update the column selector
+            self.update_column_selector()
+
+            # Load into Pynapple dataframe
+            self.dfs = from_raw_neurophotometrics_file(file_path)
+
+            # Set filter combo box
+            self.filter_selector.setCurrentIndex(0)  # Reset to "Select Filter"
+
+        except Exception as e:
+            print(f'Error loading file: {e}')
+
+    def open_dialog(self):
         # Open a file dialog to choose the CSV or PQT file
         file_path, _ = QFileDialog.getOpenFileName(
             self, 'Open File', '', 'CSV and PQT Files (*.csv *.pqt);;All Files (*)'
         )
         if file_path:
             # Load the file into a DataFrame based on its extension
-            try:
-                if (
-                    file_path.endswith('.csv')
-                    or file_path.endswith('.pqt')
-                    or file_path.endswith('.parquet')
-                ):
-                    self.dfs = from_raw_neurophotometrics_file(file_path)
-                else:
-                    raise ValueError('Unsupported file format')
-
-                if 'GCaMP' in self.dfs.keys():
-                    self.df = self.dfs['GCaMP']
-                    self.times = self.dfs['GCaMP'].index.values
-                    self.plot_time_index = np.arange(0, len(self.times))
-                    self.filtered_df = None
-                else:
-                    raise ValueError('No GCaMP found')
-
-                if 'Isosbestic' in self.dfs.keys():
-                    self.dfiso = self.dfs['Isosbestic']
-
-                # Display the dataframe in the table
-                # self.display_dataframe()
-                # Update the column selector
-                self.update_column_selector()
-
-                # Load into Pynapple dataframe
-                self.dfs = from_raw_neurophotometrics_file(file_path)
-
-                # Set filter combo box
-                self.filter_selector.setCurrentIndex(0)  # Reset to "Select Filter"
-
-            except Exception as e:
-                print(f'Error loading file: {e}')
+            self.load_file(file_path)
 
     # TODO this does not work with pynapple as format, convert back to pandas DF
     # def display_dataframe(self):
@@ -235,9 +245,9 @@ def on_column_header_clicked(self, logical_index):
         # Update the plots based on the selected column
         self.update_plots()
 
-    def apply_filter(self):
+    def apply_filter(self, filter_idx, filter_option=None):
         # Get the selected filter option from the filter dropdown
-        filter_option = self.filter_selector.currentText()
+        filter_option = filter_option or self.filter_selector.currentText()
 
         if filter_option == 'Select Filter':
             self.filtered_df = None
@@ -282,9 +292,104 @@ def filter_mad(self, df):
     #         filtered_df[col] = filtered_df[col].apply(lambda x: x if x <= 100 else None)
     #     return filtered_df
 
+    def open_behavior_gui(self):
+        signal = self.plotobj.processed_signal
+
+        if self.behavior_gui is None:
+            self.behavior_gui = BehaviorVisualizerGUI()
+        assert self.behavior_gui is not None
+
+        if signal is None:
+            print('Apply a filter before opening the Behavior GUI')
+        else:
+            print('Opening Behavior GUI')
+            self.behavior_gui.set_data(signal, self.times)
+            self.behavior_gui.show()
+
+
+class BehaviorVisualizerGUI(QWidget):
+    def __init__(
+        self,
+    ):
+        super().__init__()
+        self.trials = None
+        self.init_ui()
+
+    def set_data(self, processed_signal, times):
+        assert processed_signal is not None
+        assert times is not None
+        self.processed_signal = processed_signal
+        self.times = times
+
+    def init_ui(self):
+        # Create layout
+        main_layout = QVBoxLayout()
+
+        # Layout for file loading and selection
+        file_layout = QHBoxLayout()
+        self.load_button = QPushButton('Load File', self)
+        self.load_button.clicked.connect(self.open_dialog)
+        file_layout.addWidget(self.load_button)
+
+        main_layout.addLayout(file_layout)
+
+        # Set up plots layout
+        self.plot_layout = QGridLayout()
+        self.plotobj = plots.PlotSignalResponse()
+        self.figure, self.axes = self.plotobj.set_fig_layout()
+        self.canvas = FigureCanvas(self.figure)
+        self.plot_layout.addWidget(self.canvas, 0, 0, 1, 3)
+
+        # Create a NavigationToolbar
+        self.toolbar = NavigationToolbar(self.canvas, self)
+
+        main_layout.addLayout(self.plot_layout)
+        self.setLayout(main_layout)
+
+        self.setWindowTitle('Behavior Visualizer')
+        self.setGeometry(300, 100, 800, 600)
+
+    def load_trials(self, trials):
+        assert trials is not None
+        self.trials = trials
+        self.update_plots()
+
+    def load_file(self, file_path):
+        # load a trial file
+        try:
+            if file_path.endswith('.pqt') or file_path.endswith('.parquet'):
+                self.load_trials(pd.read_parquet(file_path))
+            else:
+                raise ValueError('Unsupported file format')
+        except Exception as e:
+            print(f'Error loading file: {e}')
+
+    def open_dialog(self):
+        file_path, _ = QFileDialog.getOpenFileName(
+            self, 'Open File', '', 'CSV and PQT Files (*.csv *.pqt);;All Files (*)'
+        )
+        if file_path:
+            self.load_file(file_path)
+
+    def update_plots(self):
+        self.figure.clear()
+
+        self.plotobj.set_data(
+            self.trials,
+            self.processed_signal,
+            self.times,
+        )
+        # NOTE: we need to update the layout as it depends on the data
+        self.figure, self.axes = self.plotobj.set_fig_layout(figure=self.figure)
+        self.plotobj.plot_trialsort_psth(self.axes)
+
+        self.canvas.draw()
+
 
 if __name__ == '__main__':
     app = QApplication(sys.argv)
     window = DataFrameVisualizerApp()
+    if len(sys.argv) >= 2:
+        window.load_file(sys.argv[1])
     window.show()
     sys.exit(app.exec_())

src/iblphotometry/plots.py

@@ -41,9 +41,6 @@ def set_axis_style(ax, fontsize=10, **kwargs):
 
 
 class PlotSignal:
-    # def __init__(self, *args, **kwargs):
-    #     self.set_data(*args, **kwargs)
-
     def set_data(
         self, raw_signal, times, raw_isosbestic=None, processed_signal=None, fs=None
     ):
@@ -143,7 +140,10 @@ def raw_processed_figure2(self, axd):
 
 
 class PlotSignalResponse:
-    def __init__(
+    def __init__(self):
+        self.psth_dict = {}
+
+    def set_data(
         self, trials, processed_signal, times, fs=None, event_window=np.array([-1, 2])
     ):
         self.trials = trials
@@ -187,10 +187,20 @@ def update_psth_dict(self, event):
         except KeyError:
             warnings.warn(f'Event {event} not found in trials table.')
 
-    def plot_trialsort_psth(self):
-        fig, axs = plt.subplots(2, len(self.psth_dict.keys()) - 1)
+    def set_fig_layout(self, figure=None):
+        n = max(1, len(self.psth_dict.keys()) - 1)
+        if figure is None:
+            figure, axs = plt.subplots(2, n, squeeze=False)
+        else:
+            axs = figure.subplots(2, n, squeeze=False)
+        figure.tight_layout()
+        return figure, axs
 
+    def plot_trialsort_psth(self, axs):
         signal_keys = [k for k in self.psth_dict.keys() if k != 'times']
+        if axs.shape[1] < len(signal_keys):
+            raise ValueError('Error, skipping PSTH plotting')
+
         for iaxs, event in enumerate(signal_keys):
             axs_plt = [axs[0, iaxs], axs[1, iaxs]]
             plot_psth(self.psth_dict[event], self.psth_dict['times'], axs=axs_plt)
@@ -206,17 +216,12 @@ def plot_trialsort_psth(self):
             if iaxs > 0:
                 axs[0, iaxs].axis('off')
                 axs[1, iaxs].set_yticks([])
-        fig.tight_layout()
-        return fig, axs
 
-    def plot_processed_trialtick(self, event_key='stimOn_times'):
-        fig, ax = plt.subplots(1, 1)
-        plt.figure(figsize=(10, 6))
+    def plot_processed_trialtick(self, ax, event_key='stimOn_times'):
         events = self.trials[event_key]
         ax.set_ylim([-0.2, 0.1])
         plot_event_tick(events, ax=ax, color='#FFC0CB', ls='-')
         plot_processed_signal(self.processed_signal, self.times, ax=ax)
-        return fig, ax
 
 
 """

src/iblphotometry_tests/test_plots.py

@@ -1,3 +1,4 @@
+import unittest
 import pandas as pd
 from pathlib import Path
 
@@ -6,6 +7,8 @@
 
 from iblphotometry.behavior import psth, psth_times
 import iblphotometry.plots as plots
+
+# from gui.rawdata_visualizer import BehaviorVisualizerGUI
 from iblphotometry.synthetic import synthetic101
 import iblphotometry.preprocessing as ffpr
 from iblphotometry_tests.base_tests import PhotometryDataTestCase
@@ -94,9 +97,13 @@ def test_class_plotsignalresponse(self):
         # eid = '77a6741c-81cc-475f-9454-a9b997be02a4'
         # trials = one.load_object(eid, 'trials')
         trials = pd.read_parquet(self.paths['trials_table_kcenia_pqt'])
-        plotobj = plots.PlotSignalResponse(trials, processed_signal, times)
-        plotobj.plot_trialsort_psth()
-        plotobj.plot_processed_trialtick()
+        plotobj = plots.PlotSignalResponse()
+        plotobj.set_data(trials, processed_signal, times)
+        _, axs = plotobj.set_fig_layout()
+        plotobj.plot_trialsort_psth(axs)
+        _, ax = plt.subplots(1, 1)
+        plotobj.plot_processed_trialtick(ax)
+        # plt.show()
         plt.close('all')
 
     """
@@ -186,3 +193,26 @@ def test_plot_event_tick(self):
         df_nph, t_events, fs = self.get_synthetic_data()
         plots.plot_event_tick(t_events)
         plt.close('all')
+
+    # def test_gui(self):
+    #     df_nph, _, fs = self.get_test_data()
+    #     processed_signal = df_nph['signal_processed'].values
+    #     times = df_nph['times'].values
+    #     trials = pd.read_parquet(self.paths['trials_table_kcenia_pqt'])
+
+    #     from PyQt5.QtWidgets import QApplication
+    #     app = QApplication(sys.argv)
+    #     window = BehaviorVisualizerGUI()
+    #     window.set_data(processed_signal, times)
+    #     window.load_trials(trials)
+    #     window.show()
+    #     # Uncomment to debug
+    #     app.exec_()
+
+
+if __name__ == '__main__':
+    unittest.main()
+    # suite = unittest.TestSuite()
+    # suite.addTest(TestPlotters())
+    # runner = unittest.TextTestRunner()
+    # runner.run(suite)