如何在Python中创建带多条连接线的散点图并自动识别节点间的同值关联

Great start with your existing NetworkX code! You’ve already nailed the visualization part with curved edges—now let’s solve the core problem: automatically identifying and adding edges between nodes that share the same value (instead of hardcoding each edge manually).

Step 1: Structure Your Data #

First, let’s assume your table data is structured like a pandas DataFrame (this is standard for tabular datasets, but we can adapt if you’re using a different format). Here’s a sample that matches your original node positions and adds a "shared value" column (e.g., a group ID that defines which nodes should connect):

import pandas as pd

# Simulate your tabular data
data = pd.DataFrame({
    "node_id": [0, 1, 2, 3, 4, 5, 6],
    "pos_x": [0, 1, 3, 5, 6, 7, 9],
    "shared_value": ["A", "A", "B", "C", "B", "B", "C"]  # Nodes with same value get connected
})

Step 2: Automatically Generate Edges from Shared Values #

We’ll group nodes by their shared_value, then create edges between nodes in the same group. You can adjust the edge logic based on your exact needs:

Option 1: Connect a Central Node to All Others in the Group

This matches your original example (where node 2 connects to 4 and 5). We’ll pick the first node in each group as the central node and link it to every other node in the group:

import networkx as nx
import matplotlib.pyplot as plt

# Initialize the directed graph
G = nx.DiGraph()
G.add_nodes_from(data["node_id"])

# Auto-add edges based on shared values
for group, group_nodes in data.groupby("shared_value"):
    node_list = group_nodes["node_id"].tolist()
    if len(node_list) > 1:
        central_node = node_list[0]
        # Link central node to all other nodes in the group
        for node in node_list[1:]:
            G.add_edge(central_node, node)

# Set node positions (pull directly from your data)
pos = {row["node_id"]: [row["pos_x"], 0] for _, row in data.iterrows()}

# Draw the graph with your original styling
nx.draw(
    G,
    pos=pos,
    connectionstyle="arc3,rad=-0.7",
    edge_color="blue",
    node_size=600,
    node_color="#a8e6cf",
    with_labels=True
)
plt.ylim([-0.5, 0.5])
plt.show()

Option 2: Connect All Pairs of Nodes in the Group

If you want every node in the same group to connect to every other node (e.g., node 0 ↔ 1, node 2 ↔ 4, node 2 ↔5, node4 ↔5), use itertools.combinations to generate all pairwise edges:

from itertools import combinations

# Reset the graph for this example
G = nx.DiGraph()
G.add_nodes_from(data["node_id"])

for group, group_nodes in data.groupby("shared_value"):
    node_list = group_nodes["node_id"].tolist()
    if len(node_list) > 1:
        # Generate all unique pairs of nodes in the group
        for u, v in combinations(node_list, 2):
            G.add_edge(u, v)
            # Uncomment below if you want bidirectional edges
            # G.add_edge(v, u)

# Reuse the same position and drawing code as above
nx.draw(
    G,
    pos=pos,
    connectionstyle="arc3,rad=-0.7",
    edge_color="blue",
    node_size=600,
    node_color="#a8e6cf",
    with_labels=True
)
plt.ylim([-0.5, 0.5])
plt.show()

Key Notes #

• If your nodes have unique y-axis positions (not all 0), just add a pos_y column to your DataFrame and update the pos dictionary to [row["pos_x"], row["pos_y"]].
• Adjust connectionstyle (e.g., rad=0.7 for upward arcs) or edge colors to match your target visualization.
• If you’re not using pandas, you can group your nodes using a dictionary (map each shared value to a list of node IDs) and follow the same edge-generation logic.

内容的提问来源于stack exchange，提问作者researchcollege111