Zoom-level dependent labels

If there are many data points, their labels may be difficult to read. To selectively display labels depending on the zoom level, use the plotly.graph_objects.FigureWidget class and register a callback.

Hint

You will need to install the anywidget package to use the plotly.graph_objects.FigureWidget class. See the Plotly documentation for details.

import functools

import plotly
import plotly.graph_objects as go
import scipy

from pdendro import attach_dendrogram

X = plotly.data.iris()[["sepal_length", "sepal_width", "petal_length", "petal_width"]]
Z = scipy.cluster.hierarchy.linkage(X)

fig = go.FigureWidget()
_ = attach_dendrogram(fig, Z)

The following code registers a callback to change labels on zooming:

def update_ticks(axis, axis_range, all_ticktext, max_n_ticks):
    """Update ticks for limiting the number."""
    step = int(abs(axis_range[1]-axis_range[0])//max_n_ticks) + 1
    axis.update(tickvals=list(range(0, len(all_ticktext), step)), ticktext=all_ticktext[::step])

# axis for data point labels
axis = fig.layout.xaxis

# get all labels
all_ticks = zip(axis.tickvals, axis.ticktext, strict=True)
all_ticks = sorted(all_ticks, key=lambda item: item[0])
_, all_ticktext = zip(*all_ticks, strict=True)

# register a callback
callback = functools.partial(update_ticks, all_ticktext=all_ticktext, max_n_ticks=5)
axis.on_change(callback, "range")

# initialize
callback(axis, axis.range)

The figure can be shown as follows:

fig

Note

The figure is not displayed in the documentation because the plotly.graph_objects.FigureWidget object can only be displayed in Jupyter. The following images illustrate the appearance at various zoom levels: