Merge pull request #28 from openalea-incubator/waffle

Add waffle representation & partial plot

src/openalea/strawberry/analysis.py

@@ -7,13 +7,21 @@
 import matplotlib.pyplot as plt
 from itertools import chain
 
-from openalea.mtg.algo import orders
-from openalea.mtg import stat, algo, traversal
+from pandas.core.groupby.groupby import DataError
+
+import numpy as np
+from matplotlib.colors import to_rgb
+import matplotlib.patches as mpatches
+import plotly.express as px
+import plotly.graph_objs as go
 
 import six
 from six.moves import map
 from six.moves import range
 
+from openalea.mtg.algo import orders
+from openalea.mtg import stat, algo, traversal
+
 
 convert = dict(Stade='Stade',
                Fleurs_ouverte='FLWRNUMBER_OPEN',
@@ -364,7 +372,8 @@ def _module_variables(g):
     module_variables['type_of_crown'] = type_of_crown # Type de crowns (Primary Crown:1, Branch crown:2 extension crown:3)
     module_variables['crown_status'] = Crown_status
     module_variables['complete_module'] = complete #(True: complete, False: incomplete)
-
+    module_variables['stage']= stage
+
     return module_variables
 
 
@@ -599,6 +608,12 @@ def mean_leaf_area(vid,g):
 def complete(vid, g):
     return g.property("complete").get(vid, False)
 
+
+def stage(vid, g):
+    _stage = g.property('Stade')
+    return next((_stage[cid] for cid in g.components(vid) if cid in _stage), None)
+
+
 ########################## Extraction on node scale ############################################
 
 def extract_at_node_scale(g, vids=[], convert=convert):
@@ -612,7 +627,7 @@ def extract_at_node_scale(g, vids=[], convert=convert):
     orders = algo.orders(g,scale=2)
 
     # Define all the rows
-    props = ['node_id', 'rank', 'branching_type', 'complete','nb_modules_branching','nb_branch_crown_branching','nb_extension_crown_branching','branching_length', 'Genotype', 'order',  'date','plant']
+    props = ['node_id', 'rank', 'branching_type', 'complete','nb_modules_branching','nb_branch_crown_branching','nb_extension_crown_branching','branching_length', 'stage', 'Genotype', 'order',  'date','plant']
     for prop in props:
         node_df[prop] = []
 
@@ -635,6 +650,7 @@ def extract_at_node_scale(g, vids=[], convert=convert):
             node_df['order'].append(orders[g.complex(vid)]) #scale=2
             node_df['plant'].append(plant(vid, g)) #scale=1
             node_df['date'].append(date(vid, g)) #scale=1
+            node_df['stage'].append(stage(vid, g)) # scale=3
 
     df = pd.DataFrame(node_df)
 
@@ -795,29 +811,252 @@ def nb_visible_leaves_tree(v, g):
     else:
         return sum(nb_visible_leaves(m,g) for m in module_tree(v, g))
 
-def genotype(vid, g):
-    cpx = g.complex_at_scale(vid, scale=1)
-    _genotype = property(g, 'Genotype')[cpx]
-    return _genotype
 
+def stage_tree(vid, g):
+    return list(stage(m,g) for m in module_tree(v, g))
 
-def plant(vid, g):
-    cpx = g.complex_at_scale(vid, scale=1)
-    return property(g, 'Plante')[cpx]
 
-def date(vid, g):
-    # Capriss: 1:'2014/12/10', 2:'2015/01/07',3:'2015/02/15',4:'2015/03/02',5:'2015/04/03',6:'2015/05/27
-    # Ciflorette: 1:'2014/12/04',2:'2015/01/07',3:'2015/02/13',4:'2015/03/02',5:'2015/03/30',6:'2015/05/27'
-    # Cir107: 1:'2014/12/10',2:'2015/01/08',3:'2015/02/11',4:'2015/03/04',5:'2015/04/02',6:'2015/05/20'
-    # Clery: 1:'2014/12/10', 2:'2015/01/07',3:'2015/02/15',4:'2015/03/02',5:'2015/04/03',6:'2015/05/27'
-    # Darselect: 1:'2014/12/10', 2:'2015/01/09', 3:'2015/02/11', 4:'2015/03/06',5:'2015/04/03',6:'2015/05/20'
-    # Gariguette: 1:'2014/12/10', 2:'2015/01/08', 3:'2015/02/12',4:'2015/03/06',5:'2015/04/02',6:'2015/05/19'
-    # d = {'2014/12/10':1,'2015/01/07':2,'2015/02/15':3,'2015/03/02':4,'2015/04/03':5,'2015/05/27':6,
-    #      '2014/12/04':1,'2015/02/13':3,'2015/03/30':5,
-    #      '2015/01/08':2,'2015/02/11':3,'2015/03/04':4,'2015/04/02':5,'2015/05/20':6,
-    #      '2015/01/09':2,'2015/02/12':3,'2015/03/06':4,'2015/05/19':6}           
-    cpx = g.complex_at_scale(vid, scale=1)
-    # _date = g.property('Sample_date')[cpx]
-    return g.property('Sample_date')[cpx]
+######################### Transformation of dataframe ######################################
+def df2waffle(df, date, index, variable, order=None, aggfunc=None, crosstab=None, *args, **kwargs):
+    '''
+        Transpose dataframe by variable with plant in columns and rank or order in index
+        This function are available for extraction at node scale (index='rank') and 
+        extraction at module scale (index= 'order')
+        Parameters:
+        -----------
+            df: dataframe from extract function at differente scale (modules and nodes scale)
+            date_selected: date which must be processed
+            variable: variable which must be processed
+        
+        Returns:
+        --------
+            a dataframe in "waffle" shape: index=date, & columns=variable
+    '''
+
+    if order:
+        data=df[(df['date']==date) & (df['order']==order)]
+    else:
+        data=df[df['date']==date]
+
+    if index=='rank':
+        res = data.pivot(index='rank',columns='plant',values=variable)
+    elif index=='order':
+        if crosstab:
+            res = pd.crosstab(index=data['order'], columns=data[variable], normalize='index')
+            res=res*100
+            res = res.round(2)
+        else:
+            # Catch data error: when values are string and aggfunc compute numbers
+            try:
+                res= data.pivot_table(index='order',columns='plant',values=variable, aggfunc=aggfunc)
+            except DataError:
+                print("ERROR, the aggregate function does not handle the data type (float func on str?)")
+                return pd.DataFrame()
+
+    else:
+        res = data.pivot(index=index,columns='plant',values=variable)
+
+    # If use plotly heatmap -> comment "res = res.fillna('')"
+    if res.isnull().values.any():
+        res = res.fillna('')
+    res = res.sort_index(ascending=False)
+    return res
+
+
+
+def plot_waffle_plotly_heatmap(df, layout={}, legend_name={}):
+
+    def df_to_plotly(df):
+        return {'z': df.values.tolist(),
+                'x': df.columns.tolist(),
+                'y': df.index.tolist()}
+
+    height = layout.get('height', 500)
+    width = layout.get('width', 500)
+    xlabel = layout.get('xlabel', 'Plant')
+    xticks = layout.get('xticks', range(0,len(df.columns)))
+    xticks_label = layout.get('xticks_label', list(df.columns))
+    ylabel = layout.get('ylabel', '')
+    yticks = layout.get('yticks', [l-1 for l in list(df.index)])
+    yticks_label = layout.get('yticks_label', list(range(0,len(df.index))))
+    title = layout.get('title', '')    
+
+    hm_layout = go.Layout(plot_bgcolor='rgba(0,0,0,0)',
+    #                    xaxis=dict(zeroline=False),
+    #                    yaxis=dict(zeroline=False, ), 
+                       autosize=False,
+                       width=width, height=height
+                      )
+
+    data = go.Heatmap(df_to_plotly(df),
+                       xgap=1,
+                       ygap=1,
+                       colorscale="aggrnyl"
+                       )
+
+    fig = go.Figure(data=data, layout=hm_layout)
+
+    return fig
+
+
+def plot_waffle_plotly_imshow(df, layout={}, legend_name={}):
+    colormap_used = plt.cm.coolwarm
+
+    values = list(set(df.values.flatten()))
+    if '' in values:
+        values.remove('')
+    try:
+        values.sort()
+    except TypeError:
+        values = [str(i) for i in values]
+        values.sort()
+    values.insert(0,'')
+
+    color_map = {val: colormap_used(i/len(values)) for i, val in enumerate(values)}
+
+    # Add the "empty" variable - and set its color as white
+    color_map[''] = (1., 1., 1., 1.)
+
+    data = np.array(df)
+
+    # Create an array where each cell is a colormap value RGBA 
+    data_3d = np.ndarray(shape=(data.shape[0], data.shape[1], 4), dtype=float)
+    for i in range(0, data.shape[0]):
+        for j in range(0, data.shape[1]):
+            data_3d[i][j] = color_map[data[i][j]]
+
+    # drop the A
+    data_3d_rgb = np.array([[to_rgb([v for v in row]) for row in col] for col in data_3d], dtype=np.float64)
+
+    yticks = list(range(0,data.shape[0]))
+    yticks.reverse()
+
+    fig = px.imshow(data,
+                    labels={'x':'Plant', 'y':'Node'},
+                    x=list(range(1,data.shape[1]+1)),
+                    y=yticks,
+                    origin='lower',
+                    color_continuous_scale='aggrnyl',
+    #                 colorbar={}
+                    )
+    fig.update_layout(plot_bgcolor='rgba(0,0,0,0)',
+                      )
+    return fig
+
+
+def plot_waffle_matplotlib(df, layout={}, legend_name={}):
+    height = layout.get('height', 500)
+    width = layout.get('width', 500)
+    xlabel = layout.get('xlabel', 'Plant')
+    xticks = layout.get('xticks', range(0,len(df.columns)))
+    xticks_label = layout.get('xticks_label', list(df.columns))
+    ylabel = layout.get('ylabel', '')
+    yticks = layout.get('yticks', [l-1 for l in list(df.index)])
+    yticks_label = layout.get('yticks_label', list(range(0,len(df.index))))
+    title = layout.get('title', '')    
+
+    colormap_used = plt.cm.coolwarm
+
+    # Sort the variables. When variables are int or float, remove the str('') (that replaced the NaN) before sorting
+    values = list(set(df.values.flatten()))
+    if '' in values:
+        values.remove('')
+    try:
+        values.sort()
+    except TypeError:
+        values = [str(i) for i in values]
+        values.sort()
+    values.insert(0,'')
+
+    w_height = len(df.index)
+    w_width = len(df.columns)
+    color_map = {val: colormap_used(i/len(values)) for i, val in enumerate(values)}
+
+    # Add the "empty" variable - and set its color as white
+    color_map[''] = (1., 1., 1., 1.)
+
+    data = np.array(df)
+
+    # Create an array where each cell is a colormap value RGBA 
+    data_3d = np.ndarray(shape=(data.shape[0], data.shape[1], 4), dtype=float)
+    for i in range(0, data.shape[0]):
+        for j in range(0, data.shape[1]):
+            data_3d[i][j] = color_map[data[i][j]]
+
+#     display the plot 
+    fig, ax = plt.subplots(1,1)
+    fig.set_size_inches(height, width)
+    fig = ax.imshow(data_3d)
+
+    # Get the axis.
+    ax = plt.gca()
+
+    # Minor ticks
+    ax.set_xticks(np.arange(-.5, (w_width), 1), minor=True);
+    ax.set_yticks(np.arange(-.5, (w_height), 1), minor=True);
+
+    # Gridlines based on minor ticks
+    ax.grid(which='minor', color='w', linestyle='-', linewidth=2)
+
+    # Manually constructing a legend solves your "catagorical" problem.
+    legend_handles = []
+
+    for i, val in enumerate(values):
+        if val!= "":
+            color_val = color_map[val]
+            legend_handles.append(mpatches.Patch(color=color_val, label=legend_name.get(val, val)))
+
+    # Add the legend. 
+    plt.legend(handles=legend_handles, loc=(1,0))
+    plt.xlabel(xlabel)
+    plt.ylabel(ylabel)
+
+    plt.xticks(ticks=xticks, labels=xticks_label)
+    plt.yticks(ticks=yticks, labels=yticks_label)
+
+    plt.title(title)
+
+    plt.show()
+
+    return fig
+
+
+def plot_waffle(df, layout={}, legend_name={}, savepath=None, plot_func='matplotlib'):
+    """
+    Plot a dataframe in "waffle" shape
+
+    layout: dict of layout parameters:
+            height/width: size of the picture in inch
+            x/ylabel: label of the x/y axis
+            x/yticks: ticks of the x/y axis
+            x/yticks_labels: labels of the ticks on the x/y axis
+            title: title
+    plot_func: library used for the ploting:
+            matplotlib: matplotlib.pyplot.subplot.imshow
+            plotly.imshow: plotly.express.imshow
+            plotly.heatmap: plotly.graph_objs.heatmap
+    """
+
+    ## Axes not working - Plotly heatmap
+    if plot_func=='plotly.heatmap':
+        fig= plot_waffle_plotly_heatmap(df=df, layout=layout, legend_name=legend_name)
+
+    # Plotly imshow
+    elif plot_func=='plotly.imshow':
+        fig= plot_waffle_plotly_imshow(df=df, layout=layout, legend_name=legend_name)
+
+    # With matplotlib
+    elif plot_func=='matplotlib':
+        try:
+            fig= plot_waffle_matplotlib(df=df, layout=layout, legend_name=legend_name)
+        except ValueError:
+            fig={}
+
+    if savepath:
+        plt.savefig(savepath)
+
+    return fig
 
 
+def plot_pie(df):
+    return px.pie(df, values=df.mean(axis=0), names=df.columns)

src/openalea/strawberry/visu3d.py

@@ -241,8 +241,3 @@ def color_code(g):
                     nid.color = (0, int(127+127/(len(stades)-1)*(i)),255)
                 else:
                     nid.color = (153, 102, 51)
-
-
-# 2D visualization
-#############################################################################
-# TODO: add visualization functions

test/test_waffle.py

@@ -0,0 +1,26 @@
+from pathlib import Path
+import os
+from openalea.mtg.io import read_mtg_file, write_mtg
+from openalea.strawberry.analysis import extract_at_node_scale, extract_at_module_scale
+from openalea.deploy.shared_data import shared_data
+import openalea.strawberry
+
+from openalea.strawberry.analysis import df2waffle
+
+def name(f):
+    return f.basename().splitext()[0]
+
+def test_df2waffle():
+    files = shared_data(openalea.strawberry).glob('*.mtg')
+    mtg_path = dict((name(f), f) for f in files)
+    mtg = read_mtg_file(mtg_path['Capriss'])
+
+    df = extract_at_node_scale(mtg)
+
+    node_scale = df2waffle(df, index='rank', date='2015/03/02', variable='branching_type')
+    assert node_scale.shape == (20, 9)
+
+    df = extract_at_module_scale(mtg)
+    module_scale=df2waffle(df, index='order', date='2015/03/02', variable='crown_status', aggfunc='median')
+    assert module_scale.shape == (3, 9)
+