Customizing labels¶

Customizing leaf axis labels¶

Using leaf_label_func¶

Leaf axis labels are controlled via leaf_label_func parameter of idd.create_dendrogram function. To customize labels, pass a callable that returns a desired label for a given linkage ID.

For example, to display the linkage ID itself:

In [3]:

#instantiate idendrogram object and pass clustering data
idd = idendrogram.idendrogram()
idd.set_cluster_info(cl_data)

#create a dendrogram object
dendrogram = idd.create_dendrogram(
    leaf_label_func = lambda _, linkage_id: str(linkage_id)
)
to_altair(dendrogram=dendrogram, height=200, width=629)

#instantiate idendrogram object and pass clustering data idd = idendrogram.idendrogram() idd.set_cluster_info(cl_data) #create a dendrogram object dendrogram = idd.create_dendrogram( leaf_label_func = lambda _, linkage_id: str(linkage_id) ) to_altair(dendrogram=dendrogram, height=200, width=629)

Out[3]:

Convenience functions¶

Two convenience labellers are available:

• idendrogram.callbacks.counts returns counts of observations represented by the linkage ID. This is the default.
• idendrogram.callbacks.cluster_labeller(fmt_string=[..]) returns a callable that returns label only for the first leaf in the cluster, leaving other labels empty.

Leaf labels representing observation counts¶

In [4]:

dendrogram = idd.create_dendrogram(
    leaf_label_func = idendrogram.callbacks.counts
)
to_altair(dendrogram=dendrogram, height=200, width=629)

dendrogram = idd.create_dendrogram( leaf_label_func = idendrogram.callbacks.counts ) to_altair(dendrogram=dendrogram, height=200, width=629)

Out[4]:

Leaf labels shown once per cluster¶

In [5]:

dendrogram = idd.create_dendrogram(
    leaf_label_func = idendrogram.callbacks.cluster_labeller(
        fmt_string="This is cluster {cluster}"
    )
)
to_altair(dendrogram=dendrogram, height=200, width=629)

dendrogram = idd.create_dendrogram( leaf_label_func = idendrogram.callbacks.cluster_labeller( fmt_string="This is cluster {cluster}" ) ) to_altair(dendrogram=dendrogram, height=200, width=629)

Out[5]:

Angles¶

Angles of leaf labels are defined in idendrogram.containers.ClusterNode. The easiest way to change defaults is by defining a new dataclass with appropriate values and passing it to the idendrogram() constructor.

In [6]:

from idendrogram.containers import AxisLabel
from dataclasses import dataclass

@dataclass
class MyAxisLabel(AxisLabel):
    labelAngle: int = -90
    
idd = idendrogram.idendrogram(axis_label_factory=lambda x:  MyAxisLabel(**x))
idd.set_cluster_info(cl_data)
dendrogram = idd.create_dendrogram(
    leaf_label_func = idendrogram.callbacks.cluster_labeller(
        fmt_string="This is cluster {cluster}"
    ))
to_altair(dendrogram=dendrogram, height=200, width=629)

from idendrogram.containers import AxisLabel from dataclasses import dataclass @dataclass class MyAxisLabel(AxisLabel): labelAngle: int = -90 idd = idendrogram.idendrogram(axis_label_factory=lambda x: MyAxisLabel(**x)) idd.set_cluster_info(cl_data) dendrogram = idd.create_dendrogram( leaf_label_func = idendrogram.callbacks.cluster_labeller( fmt_string="This is cluster {cluster}" )) to_altair(dendrogram=dendrogram, height=200, width=629)

Out[6]:

Text sizes¶

Text sizes are not natively integrated into idendrogram. You may be able to change them by manipulating the returned objects directly. E.g., in Altair:

In [7]:

chart = to_altair(dendrogram=dendrogram, height=200, width=629)
chart.configure_axis(
    labelFontSize=15,
    titleFontSize=15
)

chart = to_altair(dendrogram=dendrogram, height=200, width=629) chart.configure_axis( labelFontSize=15, titleFontSize=15 )

Out[7]:

Customizing node labels¶

Using node_label_func¶

Leaf axis labels are controlled via node_label_func parameter of idd.create_dendrogram function. To customize labels, pass a callable that returns an appropriate label for a given linkage ID.

For example, to display the linkage ID itself:

In [8]:

from idendrogram.containers import ClusterNode

#this is to change default node color/radius for better readability
@dataclass
class MyNode(ClusterNode):
    fillcolor: str = 'black'
    _default_leaf_radius: float = 10    
    radius: float = 20
    
idd = idendrogram.idendrogram(node_factory=lambda x: MyNode(**x))
idd.set_cluster_info(cl_data)

# pass node_label_func a callable that returns linkage_id
dendrogram = idd.create_dendrogram(
    node_label_func= lambda _, linkage_id: str(linkage_id),
)
to_altair(dendrogram=dendrogram, height=200, width=629)

from idendrogram.containers import ClusterNode #this is to change default node color/radius for better readability @dataclass class MyNode(ClusterNode): fillcolor: str = 'black' _default_leaf_radius: float = 10 radius: float = 20 idd = idendrogram.idendrogram(node_factory=lambda x: MyNode(**x)) idd.set_cluster_info(cl_data) # pass node_label_func a callable that returns linkage_id dendrogram = idd.create_dendrogram( node_label_func= lambda _, linkage_id: str(linkage_id), ) to_altair(dendrogram=dendrogram, height=200, width=629)

Out[8]:

Displaying cluster IDs¶

By default, idendrogram uses idendrogram.callbacks.cluster_id_if_cluster which returns flat cluster ID if the linkage ID corresponds to the flat cluster and empty label otherwise. All callback functions share the same signature, so you can get creative. Here's how you could combine it with idendrogram.callbacks.counts to show cluster ID and the number of observations in the cluster.

In [9]:

def custom_label(cl, id):
    cluster_id = idendrogram.callbacks.cluster_id_if_cluster(cl, id)
    obs_count = idendrogram.callbacks.counts(cl, id)

    return f"""C{cluster_id} ({obs_count})""" if cluster_id != "" else ""

# pass node_label_func a callable that returns linkage_id
dendrogram = idd.create_dendrogram(
    node_label_func= custom_label
)
to_altair(dendrogram=dendrogram, height=200, width=629)

def custom_label(cl, id): cluster_id = idendrogram.callbacks.cluster_id_if_cluster(cl, id) obs_count = idendrogram.callbacks.counts(cl, id) return f"""C{cluster_id} ({obs_count})""" if cluster_id != "" else "" # pass node_label_func a callable that returns linkage_id dendrogram = idd.create_dendrogram( node_label_func= custom_label ) to_altair(dendrogram=dendrogram, height=200, width=629)

Out[9]:

Customizing tooltips¶

Using node_hover_func¶

Leaf axis labels are controlled via node_hover_func parameter of idd.create_dendrogram function. To customize labels, pass a callable that returns a dictionary of key:value pairs that will be displayed in a tooltip for a given linkage ID.

For example, to display the linkage ID itself:

In [10]:

idd = idendrogram.idendrogram()
idd.set_cluster_info(cl_data)

dendrogram = idd.create_dendrogram(
    node_hover_func= lambda _, id: {'linkage ID': id}
)
to_altair(dendrogram=dendrogram, height=200, width=629)

idd = idendrogram.idendrogram() idd.set_cluster_info(cl_data) dendrogram = idd.create_dendrogram( node_hover_func= lambda _, id: {'linkage ID': id} ) to_altair(dendrogram=dendrogram, height=200, width=629)

Out[10]:

idendrogram uses idendrogram.callbacks.default_hover that displays the linkage ID and the # of observations associated with this linkage ID by default.

Customizing via postprocessing¶

You can also rely on post-processing of dendrograms to customize tooltips. Here's an example using two of available callback functions.

In [11]:

idd = idendrogram.idendrogram()
idd.set_cluster_info(cl_data)

dendrogram = idd.create_dendrogram(
    node_hover_func= lambda _, id: {'linkage ID': id}
)

for n in dendrogram.nodes:
    n.hovertext = {
        'Cluster': idendrogram.callbacks.cluster_assignments(cl_data, n.id),
        "Node type": n.type,
        "Node color": n.fillcolor,
        "Original tooltip": n.hovertext,
        "Number of observations": idendrogram.callbacks.counts(cl_data, n.id)
    }

to_altair(dendrogram=dendrogram, height=200, width=629)

idd = idendrogram.idendrogram() idd.set_cluster_info(cl_data) dendrogram = idd.create_dendrogram( node_hover_func= lambda _, id: {'linkage ID': id} ) for n in dendrogram.nodes: n.hovertext = { 'Cluster': idendrogram.callbacks.cluster_assignments(cl_data, n.id), "Node type": n.type, "Node color": n.fillcolor, "Original tooltip": n.hovertext, "Number of observations": idendrogram.callbacks.counts(cl_data, n.id) } to_altair(dendrogram=dendrogram, height=200, width=629)

Out[11]: