Barplots and Replicates with Simulation data

.pylintrc

@@ -319,7 +319,7 @@ ignored-classes=optparse.Values,thread._local,_thread._local
 # (useful for modules/projects where namespaces are manipulated during runtime
 # and thus existing member attributes cannot be deduced by static analysis). It
 # supports qualified module names, as well as Unix pattern matching.
-ignored-modules=
+ignored-modules=matplotlib
 
 # Show a hint with possible names when a member name was not found. The aspect
 # of finding the hint is based on edit distance.

petab/visualize/helper_functions.py

@@ -221,14 +221,16 @@ def create_dataset_id_list(simcond_id_list,
     return exp_data, dataset_id_list, legend_dict
 
 
-def create_figure(uni_plot_ids):
+def create_figure(uni_plot_ids, plots_to_file):
     """
     Helper function for plotting data and simulations, open figure and axes
 
     Parameters
     ----------
     uni_plot_ids: ndarray
         Array with unique plot indices
+    plots_to_file: bool
+        Indicator if plots are saved to file
 
     Returns
     -------
@@ -248,6 +250,11 @@ def create_figure(uni_plot_ids):
     # Set Colormap
     sns.set(style="ticks", palette="colorblind")
 
+    # Check if plots are saved to file and return single subplot axis
+    if plots_to_file:
+        fig, ax = plt.subplots(1, 1, squeeze=False)
+        return fig, ax, 1, 1
+
     #  Initiate subplots
     num_subplot = len(uni_plot_ids)
 
@@ -392,8 +399,7 @@ def get_data_to_plot(vis_spec: pd.DataFrame,
                      i_visu_spec: int,
                      col_id: str):
     """
-    group the data, which should be plotted and save it in pd.dataframe called
-    'ms'.
+    group the data, which should be plotted and return it as dataframe.
 
     Parameters:
         vis_spec:

petab/visualize/plot_data_and_simulation.py

@@ -8,6 +8,8 @@
 import petab
 
 from typing import Union, Optional, List
+import matplotlib.pyplot as plt
+
 
 IdsList = List[str]
 NumList = List[int]
@@ -16,14 +18,15 @@
 def plot_data_and_simulation(
         exp_data: Union[str, pd.DataFrame],
         exp_conditions: Union[str, pd.DataFrame],
-        visualization_file_path: str = '',
+        vis_spec: Optional[Union[str, pd.DataFrame]] = '',
         sim_data: Optional[Union[str, pd.DataFrame]] = None,
         dataset_id_list: Optional[List[IdsList]] = None,
         sim_cond_id_list: Optional[List[IdsList]] = None,
         sim_cond_num_list: Optional[List[NumList]] = None,
         observable_id_list: Optional[List[IdsList]] = None,
         observable_num_list: Optional[List[NumList]] = None,
-        plotted_noise: Optional[str] = 'MeanAndSD'):
+        plotted_noise: Optional[str] = 'MeanAndSD',
+        subplot_file_path: str = ''):
     """
     Main function for plotting data and simulations.
 
@@ -38,42 +41,47 @@ def plot_data_and_simulation(
     ----------
     exp_data:
         measurement DataFrame in the PEtab format or path to the data file.
-    exp_conditions: str
+    exp_conditions:
         condition DataFrame in the PEtab format or path to the condition file.
-    visualization_file_path: str (optional)
-        Path to the visualization specification file.
-    sim_data: str (optional)
+    vis_spec: str or pandas.Dataframe (optional)
+        Visualization specification DataFrame in the PEtab format or path to
+         visualization file.
+    sim_data:
         simulation DataFrame in the PEtab format
         or path to the simulation output data file.
-    dataset_id_list: list (optional)
+    dataset_id_list:
         A list of lists. Each sublist corresponds to a plot, each subplot
         contains the datasetIds for this plot.
         Only to be used if no visualization file was available.
-    sim_cond_id_list: list (optional)
+    sim_cond_id_list:
         A list of lists. Each sublist corresponds to a plot, each subplot
         contains the simulationConditionIds for this plot.
         Only to be used if no visualization file was available.
-    sim_cond_num_list: list (optional)
+    sim_cond_num_list:
         A list of lists. Each sublist corresponds to a plot, each subplot
         contains the numbers corresponding to the simulationConditionIds for
         this plot.
         Only to be used if no visualization file was available.
-    observable_id_list: list (optional)
+    observable_id_list:
         A list of lists. Each sublist corresponds to a plot, each subplot
         contains the observableIds for this plot.
         Only to be used if no visualization file was available.
-    observable_num_list: list (optional)
+    observable_num_list:
         A list of lists. Each sublist corresponds to a plot, each subplot
         contains the numbers corresponding to the observableIds for
         this plot.
         Only to be used if no visualization file was available.
-    plotted_noise: str (optional)
+    plotted_noise:
         String indicating how noise should be visualized:
         ['MeanAndSD' (default), 'MeanAndSEM', 'replicate', 'provided']
+    subplot_file_path:
+        String which is taken as file path to which single subplots are saved.
+        PlotIDs will be taken as file names.
 
     Returns
     -------
     ax: Axis object of the created plot.
+    None: In case subplots are save to file
     """
 
     if isinstance(exp_data, str):
@@ -84,18 +92,20 @@ def plot_data_and_simulation(
         exp_conditions = petab.get_condition_df(exp_conditions)
 
     # import visualization specification, if file was specified
-    if visualization_file_path != '':
-        vis_spec = petab.get_visualization_df(visualization_file_path)
-    else:
-        # create them based on simulation conditions
-        vis_spec, exp_data = get_default_vis_specs(exp_data,
-                                                   exp_conditions,
-                                                   dataset_id_list,
-                                                   sim_cond_id_list,
-                                                   sim_cond_num_list,
-                                                   observable_id_list,
-                                                   observable_num_list,
-                                                   plotted_noise)
+    if isinstance(vis_spec, str):
+        if vis_spec != '':
+            vis_spec = pd.read_csv(vis_spec, sep="\t",
+                                   index_col=None)
+        else:
+            # create them based on simulation conditions
+            vis_spec, exp_data = get_default_vis_specs(exp_data,
+                                                       exp_conditions,
+                                                       dataset_id_list,
+                                                       sim_cond_id_list,
+                                                       sim_cond_num_list,
+                                                       observable_id_list,
+                                                       observable_num_list,
+                                                       plotted_noise)
 
     # import simulation file, if file was specified
     if isinstance(sim_data, str):
@@ -104,14 +114,25 @@ def plot_data_and_simulation(
     # get unique plotIDs
     uni_plot_ids, _ = np.unique(vis_spec.plotId, return_index=True)
 
-    fig, ax, num_row, num_col = create_figure(uni_plot_ids)
+    # Switch saving plots to file on or get axes
+    plots_to_file = False
+    if subplot_file_path != '':
+        plots_to_file = True
+    else:
+        fig, ax, _, num_col = create_figure(uni_plot_ids, plots_to_file)
 
     # loop over unique plotIds
     for i_plot_id, var_plot_id in enumerate(uni_plot_ids):
 
-        # setting axis indices
-        i_row = int(np.ceil((i_plot_id + 1) / num_col)) - 1
-        i_col = int(((i_plot_id + 1) - i_row * num_col)) - 1
+        if plots_to_file:
+            fig, ax, _, num_col = create_figure(uni_plot_ids,
+                                                plots_to_file)
+            i_row = 0
+            i_col = 0
+        else:
+            # setting axis indices
+            i_row = int(np.ceil((i_plot_id + 1) / num_col)) - 1
+            i_col = int(((i_plot_id + 1) - i_row * num_col)) - 1
 
         # get indices for specific plotId
         ind_plot = (vis_spec['plotId'] == var_plot_id)
@@ -122,12 +143,18 @@ def plot_data_and_simulation(
             ax = handle_dataset_plot(i_visu_spec, ind_plot, ax, i_row, i_col,
                                      exp_data, exp_conditions, vis_spec,
                                      sim_data)
+        if plots_to_file:
+            sns.despine()
+            plt.xticks(rotation=45, ha="right")
+            plt.tight_layout()
+            plt.savefig(f'{subplot_file_path}/{var_plot_id}.png')
+            plt.close()
 
     # finalize figure
-    fig.tight_layout()
-    sns.despine()
-
-    return ax
+    if not plots_to_file:
+        fig.tight_layout()
+        sns.despine()
+        return ax
 
 
 def plot_petab_problem(petab_problem: petab.problem.Problem,

petab/visualize/plotting_config.py

@@ -72,42 +72,42 @@ def plot_lowlevel(vis_spec: pd.DataFrame,
         # add xOffset
         conditions = conditions + vis_spec.xOffset[i_visu_spec]
 
+        # TODO sort mean and sd/sem by x values (as for simulatedData below)
+        #  to avoid crazy lineplots in case x values are not sorted by default.
+        #  cf issue #207
+        #
         # construct errorbar-plots: Mean and standard deviation
         label_base = vis_spec[ind_plot].legendEntry[i_visu_spec]
         if vis_spec.plotTypeData[i_visu_spec] == 'MeanAndSD':
             p = ax[axx, axy].errorbar(
                 conditions, ms['mean'], ms['sd'], linestyle='-.', marker='.',
                 label=label_base)
-            colors = p[0].get_color()
-            if plot_sim:
-                ax[axx, axy].plot(
-                    conditions, ms['sim'], linestyle='-', marker='o',
-                    label=label_base + " simulation", color=colors)
 
         # construct errorbar-plots: Mean and standard error of mean
         elif vis_spec.plotTypeData[i_visu_spec] == 'MeanAndSEM':
-            ax[axx, axy].errorbar(
+            p = ax[axx, axy].errorbar(
                 conditions, ms['mean'], ms['sem'], linestyle='-.', marker='.',
                 label=label_base)
 
         # plotting all measurement data
         elif vis_spec.plotTypeData[i_visu_spec] == 'replicate':
-            for ii in range(0, len(ms['repl'])):
-                for k in range(0, len(ms.repl[ii])):
-                    ax[axx, axy].plot(
-                        conditions[conditions.index.values[ii]],
-                        ms.repl[ii][ms.repl[ii].index.values[k]], 'x')
+            p = ax[axx, axy].plot(
+                conditions[conditions.index.values],
+                ms.repl[ms.repl.index.values], 'x',
+                label=label_base)
 
         # construct errorbar-plots: Mean and noise provided in measurement file
         elif vis_spec.plotTypeData[i_visu_spec] == 'provided':
             p = ax[axx, axy].errorbar(
                 conditions, ms['mean'], ms['noise_model'],
                 linestyle='-.', marker='.', label=label_base)
-            colors = p[0].get_color()
-            if plot_sim:
-                ax[axx, axy].plot(
-                    conditions, ms['sim'], linestyle='-', marker='o',
-                    label=label_base + " simulation", color=colors)
+        # construct simulation plot
+        colors = p[0].get_color()
+        if plot_sim:
+            xs, ys = zip(*sorted(zip(conditions, ms['sim'])))
+            ax[axx, axy].plot(
+                xs, ys, linestyle='-', marker='o',
+                label=label_base + " simulation", color=colors)
 
         ax[axx, axy].legend()
         ax[axx, axy].set_title(vis_spec.plotName[i_visu_spec])
@@ -116,7 +116,12 @@ def plot_lowlevel(vis_spec: pd.DataFrame,
 
         x_name = vis_spec[ind_plot].legendEntry[i_visu_spec]
 
-        ax[axx, axy].bar(x_name, ms['mean'], yerr=ms['sd'])
+        p = ax[axx, axy].bar(x_name, ms['mean'], yerr=ms['sd'])
         ax[axx, axy].set_title(vis_spec.plotName[i_visu_spec])
 
+        if plot_sim:
+            colors = p[0].get_facecolor()
+            ax[axx, axy].bar(x_name + " simulation", ms['sim'], color='white',
+                             edgecolor=colors)
+
     return ax