last HMM plots. All plots refactored and checked

Author: Lblackhurst29

src/ethoscopy/behavpy_plotly.py

@@ -1840,7 +1840,6 @@ def plot_hmm_quantify(self, hmm, variable:str = 'moving', labels:list = None, co
 
         Returns:
             fig (plotly.figure.Figure): Figure object of the plot.
-                Generated by the .plot_overtime() method       
         Raises:
             Multiple assertion of ValueErrors in regards to faceting and HMM lists
         """
@@ -1931,7 +1930,6 @@ def plot_hmm_quantify_length(self, hmm, variable:str = 'moving', labels:list = N
 
         Returns:
             fig (plotly.figure.Figure): Figure object of the plot.
-                Generated by the .plot_overtime() method       
         Raises:
             Multiple assertion of ValueErrors in regards to faceting and HMM lists
         """
@@ -2014,13 +2012,14 @@ def plot_hmm_quantify_length_min_max(self, hmm, variable:str = 'moving', labels:
                     When false the colour palette is used instead as in the Seaborn version. Default is False.
                 func (str, optional): When binning to the above what function should be applied to the grouped data. 
                     Default is "max" as is necessary for the "moving" variable.
+                z_score (bool, optional): If True (Default) the z-score for each entry is found the those above/below zero are removed. 
+                    Default is True.
                 title (str, optional): The title of the plot. Default is an empty string.
                 t_column (str, optional): The name of column containing the timing data (in seconds). Default is 't'
                 grids (bool, optional): true/false whether the resulting figure should have grids. Default is False.
 
         Returns:
             fig (plotly.figure.Figure): Figure object of the plot.
-                Generated by the .plot_overtime() method       
         Raises:
             Multiple assertion of ValueErrors in regards to faceting and HMM lists
         Notes:
@@ -2106,13 +2105,14 @@ def plot_hmm_quantify_transition(self, hmm, variable:str = 'moving', labels:list
                     When false the colour palette is used instead as in the Seaborn version. Default is False.
                 func (str, optional): When binning to the above what function should be applied to the grouped data. 
                     Default is "max" as is necessary for the "moving" variable.
+                z_score (bool, optional): If True (Default) the z-score for each entry is found the those above/below zero are removed. 
+                    Default is True.
                 title (str, optional): The title of the plot. Default is an empty string.
                 t_column (str, optional): The name of column containing the timing data (in seconds). Default is 't'
                 grids (bool, optional): true/false whether the resulting figure should have grids. Default is False.
 
         Returns:
             fig (plotly.figure.Figure): Figure object of the plot.
-                Generated by the .plot_overtime() method       
         Raises:
             Multiple assertion of ValueErrors in regards to faceting and HMM lists
         Notes:
@@ -2359,34 +2359,51 @@ def plot_hmm_raw(self, hmm, variable:str = 'moving', colours:list = None, num_pl
 
         return fig
 
-    def plot_hmm_response(self, mov_df, hmm, variable = 'moving', response_col = 'has_responded', labels = None, colours = None, facet_col = None, facet_arg = None, facet_labels = None,
-        t_bin = 60, func = 'max', z_score = True, title = '', t_column = 't', col_uniform = True, grids = False):
+    def plot_hmm_response(self, mov_df, hmm, variable:str = 'moving', response_col:str = 'has_responded', labels:list = None, colours:list = None, 
+        facet_col:None|str = None, facet_arg:None|list = None, facet_labels:None|list = None, t_bin:int = 60, col_uniform:bool = False, func:str = 'max', 
+        z_score:bool = True, title:str = '', t_column:str = 't', grids:bool = False):
         """
-        Generates a plot to explore the response rate to a stimulus per hidden state from a Hidden markov Model. Y-axis is the average response rate per group / state / True or mock interactions
+        Generates a plot to explore the response rate to a stimulus per hidden state from a Hidden markov Model. 
+        Y-axis is the average response rate per group / state / True or mock interactions
+
 
             Args:
                 mov_df (behavpy dataframe): The matching behavpy dataframe containing the movement data from the response experiment
-                hmm (hmmlearn.hmm.CategoricalHMM): The accompanying trained hmmlearn model to decode the data.
-                variable (str, optional): The name of column that is to be decoded by the HMM. Default is 'moving'.
-                response_col (str, optional): The name of the coloumn that has the responses per interaction. Must be a column of bools. Default is 'has_responded'.
-                labels (list[string], optional): A list of the names of the decoded states, must match the number of states in the given model and colours. 
-                    If left as None and the model is 4 states the names will be ['Deep Sleep', 'Light Sleep', 'Quiet Awake', 'Active Awake'], else it will be ['state_0', 'state_1', ...]. Default is None
-                colours (list[string], optional): A list of colours for the decoded states, must match length of labels. If left as None and the 
-                    model is 4 states the colours will be ['Dark Blue', 'Light Blue', 'Red', 'Dark Red'], else it will be the colour palette choice. Default is None.
+                hmm (hmmlearn.hmm.CategoricalHMM): This should be a trained HMM Learn object with the 
+                    correct hidden states and emission states for your dataset
+                variable (str, optional): The column heading of the variable of interest. Default is "moving"
+                response_col (str, optional): The name of the coloumn that has the responses per interaction. 
+                    Must be a column of bools. Default is 'has_responded'.
+                labels (list[str], optional): The names of the different states present in the hidden markov model. 
+                    If None the labels are assumed to be ['Deep sleep', 'Light sleep', 'Quiet awake', 'Full awake'] if a 4 state model. 
+                    If None and not 4 states then generic labels are generated, i.e. 'state-1, state-2, state-n'.
+                    Default is None.
+                colours (list[str/RGB], optional): The name of the colours you wish to represent the different states, must be the same length as labels. 
+                    If None the colours are by default for 4 states (blue and red), if not 4 then colours from the palette are chosen. 
+                    It accepts a specific colour or an array of numbers that are acceptable to Seaborn. Default is None.
                 facet_col (str, optional): The name of the column to use for faceting, must be from the metadata. Default is None.
-                facet_arg (list, optional): The arguments to use for faceting. If None then all distinct groups will be used. Default is None.
-                facet_labels (list, optional): The labels to use for faceting, these will be what appear on the plot. If None the labels will be those from the metadata. Default is None.
-                t_bin (int, optional): The time in seconds to bin the time series data to. Default is 60,
-                func (str, optional): When binning the time what function to apply the variable column. Default is 'max'.
+                facet_arg (list, optional): The arguments to use for faceting. If None then all distinct groups will be used. 
+                    Default is None.
+                facet_labels (list, optional): The labels to use for faceting, these will be what appear on the plot. 
+                    If None the labels will be those from the metadata. Default is None.
+                t_bin (int, optional): The time in seconds you want to bin the movement data to. Default is 60 or 1 minute
+                col_uniform (bool, optional): Unique to the plotly version of this method. 
+                    When True the colour scheme is based on the states and not the faceted column. 
+                    When false the colour palette is used instead as in the Seaborn version. Default is False.
+                func (str, optional): When binning to the above what function should be applied to the grouped data. 
+                    Default is "max" as is necessary for the "moving" variable.
+                z_score (bool, optional): If True (Default) the z-score for each entry is found the those above/below zero are removed. 
+                    Default is True.
                 title (str, optional): The title of the plot. Default is an empty string.
                 t_column (str, optional): The name of column containing the timing data (in seconds). Default is 't'
-                col_uniform (bool, optional): Unique to the plotly version of this method. When True the true interaction response is coloured by the 
-                    state colour choice even with multiple groups. When false the colour palette is used instead as in the Seaborn version. Default is True.
-                grids (bool, optional): true/false whether the resulting figure should have grids. Default is False
+                grids (bool, optional): true/false whether the resulting figure should have grids. Default is False.
 
         Returns:
-            returns a plotly figure object
-
+            fig (plotly.figure.Figure): Figure object of the plot.
+        Raises:
+            KeyError:
+                If a column for the respones is not a column.   
+            Multiple assertion of ValueErrors and KeyErrors in regards to faceting and HMM lists
         Note:
             This function must be called on a behavpy dataframe that is populated by data loaded in with the stimulus_response
             analysing function.
@@ -2401,71 +2418,85 @@ def plot_hmm_response(self, mov_df, hmm, variable = 'moving', response_col = 'ha
 
         grouped_data, palette_dict, h_order = self._hmm_response(mov_df, hmm = hmm, variable = variable, response_col=response_col, labels = labels, colours = colours, 
                                             facet_col = facet_col, facet_arg = facet_arg, facet_labels = facet_labels, t_bin = t_bin, func = func, t_column = t_column)
+        
+        domains = np.arange(0, 1+(1/len(labels)), 1/len(labels))
         if facet_col is None:
             facet_col = ''
+            col_uniform = True # make it use the state colours if no facet
 
         fig = go.Figure()
         self._plot_ylayout(fig, yrange = [0, 1.01], t0 = 0, dtick = 0.2, ylabel = 'Response Rate', title = title, grid = grids)
 
 
-        for c, (col, st_lab) in enumerate(zip(colours, labels)):
-            sub_df = grouped_data[grouped_data['state'] == st_lab]
+        for c, (col, state) in enumerate(zip(colours, labels)):
+
+            sub_df = grouped_data[grouped_data['state'] == state]
 
             for lab in h_order:
 
                 sub_np = sub_df[plot_column][sub_df[facet_col] == lab].to_numpy()
-                try:
-                    mean, median, q3, q1, zlist = self._zscore_bootstrap(sub_np, z_score = z_score)
-                except KeyError:
-                    continue
-                
+                if len(sub_np) == 0: continue # normally used when no Spon. Mov. is present
+
+                mean, median, q3, q1, zlist = self._zscore_bootstrap(sub_np, z_score = z_score)
+
                 if col_uniform == True:
                     if 'Spon. Mov.' in lab:
                         marker_col = palette_dict[lab]
                     else:
                         marker_col = col
                 else:
                     marker_col = palette_dict[lab]
-                 
+
                 fig.add_trace(self._plot_meanbox(mean = [mean], median = [median], q3 = [q3], q1 = [q1], 
-                x = [lab], colour =  marker_col, showlegend = False, name = lab, xaxis = f'x{c+1}'))
+                x = [lab], colour =  marker_col, showlegend = True, name = lab, xaxis = f'x{c+1}'))
 
                 label_list = [lab] * len(zlist)
                 fig.add_trace(self._plot_boxpoints(y = zlist, x = label_list, colour = marker_col, 
                 showlegend = False, name = lab, xaxis = f'x{c+1}'))
 
-            domains = np.arange(0, 1+(1/len(labels)), 1/len(labels))
             axis = f'xaxis{c+1}'
-            self._plot_xlayout(fig, xrange = False, t0 = False, dtick = False, xlabel = st_lab, domains = domains[c:c+2], axis = axis)
+            self._plot_xlayout(fig, xrange = False, t0 = False, dtick = False, xlabel = state, domains = domains[c:c+2], axis = axis)
+            # I don't know why the above plots the x-labels at the top. Will just have to leave it
+
+        fig.update_xaxes(showticklabels=False) # remove labels as they're too bulky
+        # reorder dataframe for stats output
+        if facet_col: 
+            grouped_data = grouped_data.sort_values(facet_col)
+        grouped_data = grouped_data.drop(columns = ['previous_state'])        
 
-        stats_df = pd.DataFrame(dict([(k, pd.Series(v)) for k, v in stats_dict.items()]))
-        
         return fig, grouped_data
 
-    def plot_response_over_hmm_bouts(self, mov_df, hmm, variable = 'moving', response_col = 'has_responded', labels = None, colours = None, x_limit = 30, t_bin = 60, func = 'max', title = '', grids = False, t_column = 't'):
+    def plot_response_over_hmm_bouts(self, mov_df, hmm, variable:str = 'moving', response_col:str = 'has_responded', labels:list = None, colours:list = None, 
+        x_limit:int = 30, t_bin:int = 60, func:str = 'max', title:str = '', grids:bool = False, t_column:str = 't'):
         """ 
         Generates a plot showing the response rate per time stamp in each HMM bout. Y-axis is between 0-1 and the response rate, the x-axis is the time point
         in each state as per the time the dataset is binned to when decoded.
 
             Args:
                 mov_df (behavpy dataframe): The matching behavpy dataframe containing the movement data from the response experiment
-                hmm (hmmlearn.hmm.CategoricalHMM): The accompanying trained hmmlearn model to decode the data.
-                variable (str, optional): The name of column that is to be decoded by the HMM. Default is 'moving'.
-                response_col (str, optional): The name of the coloumn that has the responses per interaction. Must be a column of bools. Default is 'has_responded'.
-                labels (list[string], optional): A list of the names of the decoded states, must match the number of states in the given model and colours. 
-                    If left as None and the model is 4 states the names will be ['Deep Sleep', 'Light Sleep', 'Quiet Awake', 'Active Awake']. Default is None
-                colours (list[string], optional): A list of colours for the decoded states, must match length of labels. If left as None and the 
-                    model is 4 states the colours will be ['Dark Blue', 'Light Blue', 'Red', 'Dark Red']. Default is None.
-                x_limit (int, optional): A number to limit the x-axis by to remove outliers, i.e. 30 would be 30 minutes or less if t_bin is 60. Default 30.
+                hmm (hmmlearn.hmm.CategoricalHMM): This should be a trained HMM Learn object with the 
+                    correct hidden states and emission states for your dataset
+                variable (str, optional): The column heading of the variable of interest. Default is "moving"
+                response_col (str, optional): The name of the coloumn that has the responses per interaction. 
+                    Must be a column of bools. Default is 'has_responded'.
+                labels (list[str], optional): The names of the different states present in the hidden markov model. 
+                    If None the labels are assumed to be ['Deep sleep', 'Light sleep', 'Quiet awake', 'Full awake'] if a 4 state model. 
+                    If None and not 4 states then generic labels are generated, i.e. 'state-1, state-2, state-n'.
+                    Default is None.
+                colours (list[str/RGB], optional): The name of the colours you wish to represent the different states, must be the same length as labels. 
+                    If None the colours are by default for 4 states (blue and red), if not 4 then colours from the palette are chosen. 
+                    It accepts a specific colour or an array of numbers that are acceptable to Seaborn. Default is None.
+                x_limit (int, optional): A number to limit the x-axis by to remove outliers, i.e. 30 would be 30 minutes or less if t_bin is 60. 
+                    Default 30.
                 t_bin (int, optional): The time in seconds to bin the time series data to. Default is 60,
                 func (str, optional): When binning the time what function to apply the variable column. Default is 'max'.
                 title (str, optional): The title of the plot. Default is an empty string.
                 grids (bool, optional): true/false whether the resulting figure should have grids. Default is False.
                 t_column (str, optional): The name of column containing the timing data (in seconds). Default is 't'
+                figsize (tuple, optional): The size of the figure in inches. Default is (0, 0) which auto-adjusts the size.
 
         Returns:
-            fig (plotly.figure.Figure): Figure object of the plot.
-
+            fig (matplotlib.figure.Figure): Figure object of the plot.
         Note:
             This function must be called on a behavpy dataframe that is populated with data loaded with the stimulus_response
                 analysing function.