Update cmd compare #118

src/refinegems/analysis/comparison.py

@@ -5,17 +5,21 @@
 All other stats are shown in the memote report.
 """
 
-__author__ = "Famke Baeuerle"
+__author__ = "Famke Baeuerle, Gwendolyn O. Döbel and Carolin Brune"
 
 ################################################################################
 # requirements
 ################################################################################
 
-import pandas as pd
 import matplotlib
+import matplotlib.colors
+import numpy as np
+import pandas as pd
+import warnings
 
 from cobra import Model as cobraModel
 from libsbml import Model as libModel
+from typing import Union
 from venn import venn
 
 from ..utility.io import search_sbo_label
@@ -44,7 +48,9 @@ def get_sbo_mapping_multiple(models: list[libModel]) -> pd.DataFrame:
     df['SBO-Name'] = df['SBO-Term'].apply(search_sbo_label)
     return df
 
-def plot_rea_sbo_multiple(models: list[libModel], rename=None) -> matplotlib.axes.Axes:
+def plot_rea_sbo_multiple(models: list[libModel], rename=None, 
+                          color_palette:Union[str,list[str]]='Paired',
+                          figsize:tuple=(10,10)) -> matplotlib.figure.Figure:
     """Plots reactions per SBO Term in horizontal bar chart with stacked bars for the models
 
     Args:
@@ -55,20 +61,47 @@ def plot_rea_sbo_multiple(models: list[libModel], rename=None) -> matplotlib.axe
             Defaults to None.
 
     Returns:
-        matplotlib.axes.Axes: 
-            Pandas stacked barchart
+        matplotlib.figure.Figure: 
+            The results as a stacked barchart
     """
+    # set the colours 
+    match color_palette:
+        case str():
+            try:
+                cmap = matplotlib.colormaps[color_palette]
+            except ValueError:
+                warnings.warn('Unknown color palette, setting it to "Paired"')
+                cmap = matplotlib.colormaps['Paired']
+            #@TODO : make sure all types of colourmaps work
+            if isinstance(cmap,matplotlib.colors.ListedColormap):
+                cmap = cmap.colors[0:len(models)]
+                # @TODO
+                #    alternative, if colour list shorter than model list
+            else:
+                cmap = cmap(np.linspace(0,1,len(models)))
+        case list():
+            cmap = color_palette
+        case _:
+            mes = 'Unkown type for color_palette'
+            raise TypeError(mes)
+
+    # prepare the data
     map = get_sbo_mapping_multiple(models)
     id_list = [mod.id for mod in models]
     map = map[(map[id_list]>3).all(axis=1)]
     map = map.drop('SBO-Term', axis=1).sort_values(id_list[0]).set_index('SBO-Name')
     if rename is not None:
         map = map.rename(rename, axis=1)
-    fig = map.plot.barh(stacked=True, width=.8, figsize=(8,10))
+    # make the figure
+    fig = map.plot.barh(stacked=True, width=.8, figsize=figsize, color=cmap)
+    for patch in fig.patches:
+        colour = patch.get_facecolor()
+        patch.set_edgecolor(colour)
     fig.set_ylabel('')
     fig.set_xlabel('number of reactions', fontsize=16)
     fig.legend(loc='lower right')
-    return fig
+    return fig.get_figure()
+
 
 def plot_venn(models: list[cobraModel], entity: str, perc: bool=False, rename=None) -> matplotlib.axes.Axes:
     """Creates Venn diagram to show the overlap of model entities

src/refinegems/cmd_access.py

@@ -15,6 +15,7 @@
 from pathlib import Path
 import pandas as pd
 import logging
+import sys
 
 from refinegems.curation.db_access import biocyc
 from refinegems.utility.io import write_model_to_file
@@ -406,16 +407,24 @@ def pancore(modelpath, pcpath, based_on,dir):
 
 @analyse.command()
 @click.argument('modelpaths', nargs=-1, type=click.Path(exists=True))
-@click.option('--type','-t',required=False,type=click.Choice(['sboterm','entities']), 
-              show_default=True,  default=None, help='Type of comparison to be performed.')
+# @TODO 
+#    extend / add different comparison options
+@click.option('--type','-t',required=False,type=click.Choice(['sboterm']), multiple=True,
+              show_default=True,  default=[], help='Type of comparison to be performed.')
 @click.option('--all',required=False, is_flag=True, show_default=True, default=False, 
               help='Shortcut to run all comparisons. Overwrites input of type.')
-def compare(modelpaths,type,all):
+@click.option('-d', '--dir', required=False, type=click.Path(file_okay=False), show_default=True, default='comparison', help='Path to the output dir.')
+@click.option('-c','--colour', required=False, type=str, default='Paired', show_default=True, help='Name of Matplotlib colour palette for the plot.')
+def compare(modelpaths,type,all,dir,colour):
    """Compare models.
    """
+   Path(dir).mkdir(parents=True, exist_ok=True)
+   # compare SBOterms
    if all or 'sboterm' in type:
-      models = rg.utility.io.load_model(modelpaths)
-      fig = rg.analysis
+      models = rg.utility.io.load_model(list(modelpaths),'libsbml')
+      fig = rg.analysis.comparison.plot_rea_sbo_multiple(models,color_palette=colour)
+      fig.savefig(Path(dir, 'sboterm'), dpi=400, bbox_inches='tight')
+
 
 # analyse growth
 # --------------

src/refinegems/utility/io.py

@@ -101,11 +101,11 @@ def load_libsbml_model(modelpath:str) -> libModel:
         case list():
 
             loaded_models = []
-            for modelpath in modelpath:
+            for m in modelpath:
                 if package == 'cobra':
-                    loaded_models.append(load_cobra_model(modelpath))
+                    loaded_models.append(load_cobra_model(m))
                 elif package == 'libsbml':
-                    loaded_models.append(load_libsbml_model(modelpath))
+                    loaded_models.append(load_libsbml_model(m))
             return loaded_models
 
         # read in a single model