"""
3D layouts
==========

This example shows how to visualise graphs or networks in 3D using `iplotx`. Of course, a 3D layout is needed
for this. Here, we use the Fruchterman-Reingold layout algorithm from ``igraph`` to generate a 3D layout.

.. note::
    3D visualisation is most useful when used **interactively**, so you can rotate and pan the plot to inspect
    it from different angles. Matplotlib supports this both in Jupyter notebooks and in IPython via
    multiple interactive backends (e.g., TkAgg, Qt5Agg, etc.). These plots can also be saved as static
    images (the ones you see below were generated this way), however these static images can be quite
    difficult to interpret.
"""

import igraph as ig
import iplotx as ipx

# Make the graph
g = ig.Graph.Erdos_Renyi(30, m=50)

# Make a 3D layout
layout = g.layout_fruchterman_reingold_3d()

# Visualise the graph in 3D
ipx.network(
    g,
    layout,
    vertex_alpha=0.7,
    edge_alpha=0.4,
    figsize=(8, 8),
)

# %%
# Below is a variation using arrows and vertex labels:

import igraph as ig
import iplotx as ipx

# Make the graph
g = ig.Graph.Erdos_Renyi(30, m=50, directed=True)

# Make a 3D layout
layout = g.layout_fruchterman_reingold_3d()

# Visualise the graph in 3D
ipx.network(
    g,
    layout,
    vertex_alpha=0.3,
    edge_alpha=0.5,
    vertex_labels=True,
    figsize=(8, 8),
)

# %%
# .. warning::
#    3D visualisation does not support all features of 2D visualisation yet. Curved edges, waypoints, and edge
#    labels are currently unsupported. PRs are welcome!