Imputing "null" values for missing keys?

[ericmjl]

Imputing "null" values for missing keys?

[yerraeee]

Hi Eric,

Thank you for this awesome nxviz package.

I have a bipartite graph of "technical-tickets" and "teams" (that work on tickets) as nodes. So "Ticket"s and "Team"s have different attributes. When I tried to Color,Order and Group the ticket-nodes based on attributes that absent in Team-nodes, faced issues due to missing keys in Team-nodes.

I have overwritten the class methods in my code with below 2-functions. you may verify and update the code as appropriate.

Below code helps :

1. Coloring, grouping and ordering the nodes irrespective of missing keys in few nodes.
2. Overrides the CMAP dictionary to take unlimited values of categories in any key, original one has a limit of 8-categories for any key.

def new_compute_node_colors(self):
        import logging
        import nxviz as nv
        import matplotlib.patches as patches
        import matplotlib.pyplot as plt
        import networkx as nx
        from matplotlib.cm import get_cmap
        from matplotlib.path import Path
        from nxviz.geometry import circos_radius, get_cartesian, node_theta
        from nxviz.utils import (cmaps, infer_data_type, is_data_diverging, num_discrete_groups)
        """Compute the node colors. Also computes the colorbar."""
        #data = [self.graph.node[n][self.node_color] if self.node_color in self.graph.node[n].keys() else 'no_key' for n in self.nodes]
        data = [self.graph.node[n][self.node_color] if self.node_color in self.graph.node[n].keys() else 'zzz_no_key' for n in self.nodes]
        data_reduced = sorted(list(set(data)))
        dtype = infer_data_type(data)
        n_grps = num_discrete_groups(data)

        if dtype == 'categorical' or dtype == 'ordinal':
            #cmap = get_cmap(cmaps['Accent_{0}'.format(n_grps)].mpl_colormap)
            cmap = get_cmap(cmaps['Paired_12'].mpl_colormap)
        elif dtype == 'continuous' and not is_data_diverging(data):
            cmap = get_cmap(cmaps['continuous'].mpl_colormap)
        elif dtype == 'continuous' and is_data_diverging(data):
            cmap = get_cmap(cmaps['diverging'].mpl_colormap)

        for d in data:
            idx = data_reduced.index(d) / n_grps
            self.node_colors.append(cmap(idx))

        # Add colorbar if required.
        logging.debug('length of data_reduced: {0}'.format(len(data_reduced)))
        logging.debug('dtype: {0}'.format(dtype))
        if len(data_reduced) > 1 and dtype == 'continuous':
            self.sm = plt.cm.ScalarMappable(cmap=cmap,
                                            norm=plt.Normalize(vmin=min(data_reduced),  # noqa
                                                               vmax=max(data_reduced)   # noqa
                                                               )
                                            )
            self.sm._A = []

def new_group_and_sort_nodes(self):
        """
        Groups and then sorts the nodes according to the criteria passed into
        the Plot constructor.
        """
        if self.node_grouping and not self.node_order:
            #self.nodes = [n for n, d in sorted(self.graph.nodes(data=True), key=lambda x: x[1][self.node_grouping])]
            self.nodes = [n for n,d in sorted([(n,self.graph.node[n][self.node_grouping]) if self.node_grouping in self.graph.node[n].keys() else (n,'zzz_no_key') for n in self.graph.nodes],key= lambda x: x[1])]

        elif self.node_order and not self.node_grouping:
            #self.nodes = [n for n, _ in sorted(self.graph.nodes(data=True), key=lambda x: x[1][self.node_order])]
            self.nodes = [n for n, _ in sorted([(n,self.graph.node[n][self.node_order]) if self.node_order in self.graph.node[n].keys() else (n,'zzz_no_key') for n in self.graph.nodes],key= lambda x: x[1])]

        elif self.node_grouping and self.node_order:
            #self.nodes = [n for n, d in sorted(self.graph.nodes(data=True), key=lambda x: (x[1][self.node_grouping], x[1][self.node_order]))]
            nodes = []
            for n in self.nodes:
                if self.node_grouping in self.graph.node[n].keys():
                    grouping = self.graph.node[n][self.node_grouping]
                else:
                    grouping = 'zzz_no_key'
                
                if self.node_order in self.graph.node[n].keys():
                    order = self.graph.node[n][self.node_order]
                else:
                    order = 'zzz_no_key'
                    
                nodes.append((n,grouping,order))
            
            self.nodes = [n for n,group,order in sorted(nodes,key= lambda x: (x[1], x[2]))]
                
                
            

new_cmaps = {
    'Accent_2': qualitative.Accent_3,
    'Accent_3': qualitative.Accent_3,
    'Accent_4': qualitative.Accent_4,
    'Accent_5': qualitative.Accent_5,
    'Accent_6': qualitative.Accent_6,
    'Accent_7': qualitative.Accent_7,
    'Accent_8': qualitative.Accent_8,
    'Paired_12': qualitative.Paired_12, #added colorbrewer set ref:https://jiffyclub.github.io/palettable/colorbrewer/qualitative/#paired_12
    'continuous': sequential.YlGnBu_9,
    'diverging': diverging.RdBu_11,
}
            
nv.plots.BasePlot.compute_node_colors = new_compute_node_colors
nv.plots.BasePlot.group_and_sort_nodes = new_group_and_sort_nodes
nv.utils.cmaps = new_cmaps

[yerraeee]

Pasting my code again, did not paste properly above.

def new_compute_node_colors(self):
        import logging
        import nxviz as nv
        import matplotlib.patches as patches
        import matplotlib.pyplot as plt
        import networkx as nx
        from matplotlib.cm import get_cmap
        from matplotlib.path import Path
        from nxviz.geometry import circos_radius, get_cartesian, node_theta
        from nxviz.utils import (cmaps, infer_data_type, is_data_diverging, num_discrete_groups)
        """Compute the node colors. Also computes the colorbar."""
        #data = [self.graph.node[n][self.node_color] if self.node_color in self.graph.node[n].keys() else 'no_key' for n in self.nodes]
        data = [self.graph.node[n][self.node_color] if self.node_color in self.graph.node[n].keys() else 'zzz_no_key' for n in self.nodes]
        data_reduced = sorted(list(set(data)))
        dtype = infer_data_type(data)
        n_grps = num_discrete_groups(data)

        if dtype == 'categorical' or dtype == 'ordinal':
            #cmap = get_cmap(cmaps['Accent_{0}'.format(n_grps)].mpl_colormap)
            cmap = get_cmap(cmaps['Paired_12'].mpl_colormap)
        elif dtype == 'continuous' and not is_data_diverging(data):
            cmap = get_cmap(cmaps['continuous'].mpl_colormap)
        elif dtype == 'continuous' and is_data_diverging(data):
            cmap = get_cmap(cmaps['diverging'].mpl_colormap)

        for d in data:
            idx = data_reduced.index(d) / n_grps
            self.node_colors.append(cmap(idx))

        # Add colorbar if required.
        logging.debug('length of data_reduced: {0}'.format(len(data_reduced)))
        logging.debug('dtype: {0}'.format(dtype))
        if len(data_reduced) > 1 and dtype == 'continuous':
            self.sm = plt.cm.ScalarMappable(cmap=cmap,
                                            norm=plt.Normalize(vmin=min(data_reduced),  # noqa
                                                               vmax=max(data_reduced)   # noqa
                                                               )
                                            )
            self.sm._A = []

def new_group_and_sort_nodes(self):
        """
        Groups and then sorts the nodes according to the criteria passed into
        the Plot constructor.
        """
        if self.node_grouping and not self.node_order:
            #self.nodes = [n for n, d in sorted(self.graph.nodes(data=True), key=lambda x: x[1][self.node_grouping])]
            self.nodes = [n for n,d in sorted([(n,self.graph.node[n][self.node_grouping]) if self.node_grouping in self.graph.node[n].keys() else (n,'zzz_no_key') for n in self.graph.nodes],key= lambda x: x[1])]

        elif self.node_order and not self.node_grouping:
            #self.nodes = [n for n, _ in sorted(self.graph.nodes(data=True), key=lambda x: x[1][self.node_order])]
            self.nodes = [n for n, _ in sorted([(n,self.graph.node[n][self.node_order]) if self.node_order in self.graph.node[n].keys() else (n,'zzz_no_key') for n in self.graph.nodes],key= lambda x: x[1])]

        elif self.node_grouping and self.node_order:
            #self.nodes = [n for n, d in sorted(self.graph.nodes(data=True), key=lambda x: (x[1][self.node_grouping], x[1][self.node_order]))]
            nodes = []
            for n in self.nodes:
                if self.node_grouping in self.graph.node[n].keys():
                    grouping = self.graph.node[n][self.node_grouping]
                else:
                    grouping = 'zzz_no_key'
                
                if self.node_order in self.graph.node[n].keys():
                    order = self.graph.node[n][self.node_order]
                else:
                    order = 'zzz_no_key'
                    
                nodes.append((n,grouping,order))
            
            self.nodes = [n for n,group,order in sorted(nodes,key= lambda x: (x[1], x[2]))]
                
                
            

new_cmaps = {
    'Accent_2': qualitative.Accent_3,
    'Accent_3': qualitative.Accent_3,
    'Accent_4': qualitative.Accent_4,
    'Accent_5': qualitative.Accent_5,
    'Accent_6': qualitative.Accent_6,
    'Accent_7': qualitative.Accent_7,
    'Accent_8': qualitative.Accent_8,
    'Paired_12': qualitative.Paired_12, #added colorbrewer set ref:https://jiffyclub.github.io/palettable/colorbrewer/qualitative/#paired_12
    'continuous': sequential.YlGnBu_9,
    'diverging': diverging.RdBu_11,
}
            
nv.plots.BasePlot.compute_node_colors = new_compute_node_colors
nv.plots.BasePlot.group_and_sort_nodes = new_group_and_sort_nodes
nv.utils.cmaps = new_cmaps