如何使用Python的Multinetx库绘制指定样式的多层网络？

Here's a polished, complete version of your code to draw the multilayer network as intended, plus explanations for key parts and troubleshooting tips:

import Multinetx as mx
import matplotlib.pyplot as plt

# Assuming mst_pearson, mst_kendall, mst_tail are pre-built NetworkX graphs (your MSTs)
mg = mx.MultilayerGraph(list_of_layers=[mst_pearson, mst_kendall, mst_tail])

# Optional: Add inter-layer edges to connect identical nodes across layers (remove if not needed)
for node in mst_pearson.nodes():
    mg.add_edge((0, node), (1, node))  # Connect Pearson layer to Kendall layer
    mg.add_edge((1, node), (2, node))  # Connect Kendall layer to Tail layer

# Hide axis for a cleaner visualization
plt.axis('off')

# Calculate node positions: base layout from Pearson MST, shift layers vertically
pos = mx.get_position(
    mg,
    mx.fruchterman_reingold_layout(mst_pearson),
    layer_vertical_shift=0.2,
    layer_horizontal_shift=0.0,
    proj_angle=47
)

# Draw the full multilayer network
mx.draw_networkx(
    mg,
    pos=pos,
    node_size=50,
    with_labels=False,  # Set to True if you need node labels (adjust node_size if so)
    node_color=['#1f78b4', '#33a02c', '#e31a1c'],  # Distinct colors per layer for clarity
    edge_color='gray',
    alpha=0.7  # Soften edges to avoid visual clutter
)

# Add layer labels to make the plot easier to interpret
plt.text(-1.2, 0.4, 'Pearson MST', fontsize=10)
plt.text(-1.2, 0.0, 'Kendall MST', fontsize=10)
plt.text(-1.2, -0.4, 'Tail MST', fontsize=10)

# Render the final plot
plt.show()

Key Improvements & Notes: #

• Completed with_labels: Set to False here (small node size would make labels unreadable). Flip to True if you need node identifiers, and increase node_size to ~150+.
• Inter-layer edges: This optional block creates connections between the same node across layers—great for showing correspondence between your MSTs. Remove it if you want strictly separate layers.
• Layer-specific colors: Assigning unique colors to each layer makes it instantly clear which nodes belong to which MST.
• Layer labels: Added text annotations next to each layer so viewers don’t have to guess which is which.
• Final plt.show(): Critical step to actually display the plot (your original code was missing this).

Troubleshooting Common Issues: #

• Overlapping nodes: Increase layer_vertical_shift to 0.3 or higher, or switch the base layout algorithm to mx.spring_layout for more spread-out nodes.
• Messy inter-layer edges: If you don’t want all cross-node connections, filter which nodes to link (e.g., only nodes with significant correlations).
• Import errors: Make sure Multinetx is installed correctly (pip install multinetx) and you’re using a compatible NetworkX version (Multinetx works best with NetworkX 2.x).

内容的提问来源于stack exchange，提问作者taylor