Changing Axis Order

Users may want to specify the order of each axis in a hive plot.

Additionally, users may be interested in looking at a subset of possible axes for the current partition.

This notebook demonstrates how to handle both of these cases when working with a HivePlot instance.

from hiveplotlib.datasets import example_hive_plot

We will base this discussion on the following toy hive plot:

# keep inter group edge colors consistent throughout notebook
INTER_EDGE_COLORS = {
    "A": {"B": "limegreen"},
    "B": {"C": "darkorange"},
    "C": {"A": "darkviolet"},
}

hp = example_hive_plot()
for p1 in INTER_EDGE_COLORS:
    for p2 in INTER_EDGE_COLORS[p1]:
        hp.update_edges(
            partition_id_1=p1,
            partition_id_2=p2,
            color=INTER_EDGE_COLORS[p1][p2],
        )

hp.plot();

Changing Order of All Partition Values

Users can set the axes’ order both when instantiating a new HivePlot as well as on an existing HivePlot instance.

The axes names we provide will correspond to the values used to partition the nodes i.e. the partition_variable used from the node dataset.

hp.nodes.data.head()

	unique_id	low	med	high	partition_0
0	0	6.363247	14.795079	23.193620	B
1	1	2.695169	12.321405	21.873202	A
2	2	0.409326	18.010787	26.718541	A
3	3	0.165111	19.226066	21.949123	A
4	4	8.124570	12.658641	25.771102	C

sorted(hp.nodes.data[hp.partition_variable].unique())

['A', 'B', 'C']

Thus for this example, we can only use the variables A, B, and C when setting the axes’ order.

Setting Order of Axes on Hive Plot Initialization

On initialization, we can set the axes_order parameter by providing these unique values from our partition data column in whatever order we choose.

hp = example_hive_plot(
    axes_order=["B", "A", "C"],  # change order on initialization
)
# update edge colors same as above
for p1 in INTER_EDGE_COLORS:
    for p2 in INTER_EDGE_COLORS[p1]:
        hp.update_edges(
            partition_id_1=p1,
            partition_id_2=p2,
            color=INTER_EDGE_COLORS[p1][p2],
        )

hp.plot();

Changing Order of Axes on an Existing Hive Plot

On an existing hive plot, we need only call the HivePlot.set_axes_order() method:

hp.set_axes_order(axes=["C", "A", "B"])
hp.plot();

Using a Subset of Partition Values

What if we want to show some, but not all of the partition values? For example, suppose there’s more than 3 partition values:

hp = example_hive_plot(
    cutoffs=4,
    labels=["A", "B", "C", "D"],
)
hp.plot();

With more than 3 axes, we are only looking at a subset of the edges in the above hive plot, omitting any edges between A and C as well as B and D.

Solving this problem is discussed in greater detail on the Hive Plots with More Than 3 Groups page, but here we will only demonstrate how one can choose the axes order as a subset of these partition values.

Instantiating Hive Plot Using a Subset of Partition Values

We can choose to instantiate the above example only choosing a subset of the values under the axes_order parameter:

hp = example_hive_plot(
    cutoffs=4,
    labels=["A", "B", "C", "D"],
    axes_order=["D", "B", "A"],
)
hp.plot();

Updating an Existing Hive Plot Using a Subset of Partition Values

For an existing hive plot, we cannot subset axes with a HivePlot.set_axes_order() call, but we can rerun HivePlot.set_partition() with the desired axes_order subset:

import traceback

from hiveplotlib.exceptions import InvalidAxesOrderError

hp = example_hive_plot(
    cutoffs=4,
    labels=["A", "B", "C", "D"],
)

try:
    hp.set_axes_order(axes=["D", "B", "A"])
except InvalidAxesOrderError:
    traceback.print_exc()

Traceback (most recent call last):
  File "/tmp/ipykernel_20164/3413509945.py", line 11, in <module>
    hp.set_axes_order(axes=["D", "B", "A"])
  File "/home/garyk/repos/hiveplotlib/src/hiveplotlib/hiveplot.py", line 2519, in set_axes_order
    raise InvalidAxesOrderError(msg)
hiveplotlib.exceptions.hive_plot.InvalidAxesOrderError: Provided axes order ['D', 'B', 'A'] does not include all partition names: ['A', 'B', 'C', 'D'].
Missing partition names: ['C'].
If you want to only show a subset of the partition names, then run set_partition() with the desired `axes_order`.

hp.set_partition(
    partition_variable=hp.partition_variable,  # same partition variable
    sorting_variables=hp.sorting_variables,  # same sorting variables
    axes_order=["D", "B", "A"],  # subset order allowed here
)
hp.plot();