问在具有给定位置坐标的节点图中设置边标签的问题
EN

Stack Overflow用户

提问于 2018-12-27 14:02:57

回答 1查看 142关注 0票数 1

我有一个名为"loc“的11x11数据帧，其中包含节点的位置坐标(X、Y和Z)，节点名称作为索引。我还有另一个dataframe "dist“，它包含节点之间的距离，同时也包含列标题和索引中的节点名称。我想绘制网络图，以便每个节点都标有其名称(loc的索引)。此外，每个节点都应该连接到每个其他节点。

但是，我希望每条边都标有两个节点之间的距离。我

我已经创建了图形，用标签绘制了节点，并在它们之间绘制了边。但是，无法为图例绘制标签。

#####  Example data similar to my actual data
coord     = np.random.randint(low=1, high=100, size=(11,3))  
lab       = ['AP', 'STA1', 'STA2', 'STA3', 'STA4', 'STA5', 'STA6', 'STA7', 
'STA8', 'STA9', 'STA10']
loc       = pd.DataFrame.from_records(coord, columns=['X', 'Y', 'Z'], 
index=lab)
d         = distance_matrix(loc,loc) # Distance between each device
dist      = pd.DataFrame.from_records(d, columns=lab)
dist      = dist.set_index(dist.columns)
dist      = dist.round(decimals=1)

##########  Network Plot without edges
coord         = np.random.randint(low=1, high=100, size=(11,3))  
lab           = ['AP', 'STA1', 'STA2', 'STA3', 'STA4', 'STA5', 'STA6', 
'STA7', 'STA8', 'STA9', 'STA10']
loc           = pd.DataFrame.from_records(coord, columns=['X', 'Y', 'Z'], 
index=lab)


d         = distance_matrix(loc,loc) # Distance between each device
dist      = pd.DataFrame.from_records(d, columns=lab)
dist      = dist.set_index(dist.columns)
dist      = dist.round(decimals=1)

# Add the nodes to graph
G=nx.Graph() 
for i in range(len(loc)):
    G.add_node(loc.index[i])

pos = loc.ix[:,0:2].transpose().to_dict(orient='list')

# Add edges
for i in range(len(loc)):
    for j in range(len(loc)):
        G.add_edge(loc.index[i], loc.index[j])

# Draw the network
fig, ax = plt.subplots(figsize=(5,5))
ax.axis('equal')
nx.draw_networkx(G, pos=pos, arrows= True, with_labels=True, node_size=600, 
node_shape='o', alpha=0.5, font_size=10) 
plt.show()

1)给定的代码生成所需的图，但没有边标签。我想在边缘的中心或更易读的地方绘制标签。请记住，edge_label表示两个节点之间的距离(即"dist“数据帧中的值，其索引和列标题具有相同的值)。2)我们能否在3D中绘制相同的网络，因为节点有三个坐标(X，Y和Z)。和在我的代码中一样，我只绘制X和Y坐标。

python

matplotlib

networkx

回答 1

Stack Overflow用户

发布于 2018-12-30 15:57:49

对于图形边的标签，您需要首先将此数据添加到边：

G.add_edge(loc.index[i], loc.index[j], weight=d[i][j])

然后，对于绘制标签，可以调用nx.draw_networkx_edge_labels()

所有的一切：

import numpy as np
import pandas as pd
import networkx as nx
import scipy
import matplotlib.pyplot as plt

#####  Example data similar to my actual data
coord     = np.random.randint(low=1, high=100, size=(11,3))
lab       = ['AP', 'STA1', 'STA2', 'STA3', 'STA4', 'STA5', 'STA6', 'STA7',
'STA8', 'STA9', 'STA10']
loc       = pd.DataFrame.from_records(coord, columns=['X', 'Y', 'Z'],
index=lab)
d         = scipy.spatial.distance_matrix(loc,loc) # Distance between each device
dist      = pd.DataFrame.from_records(d, columns=lab)
dist      = dist.set_index(dist.columns)
dist      = dist.round(decimals=1)

##########  Network Plot without edges
coord         = np.random.randint(low=1, high=100, size=(11,3))
lab           = ['AP', 'STA1', 'STA2', 'STA3', 'STA4', 'STA5', 'STA6',
'STA7', 'STA8', 'STA9', 'STA10']
loc           = pd.DataFrame.from_records(coord, columns=['X', 'Y', 'Z'],
index=lab)


d         = scipy.spatial.distance_matrix(loc,loc) # Distance between each device
dist      = pd.DataFrame.from_records(d, columns=lab)
dist      = dist.set_index(dist.columns)
dist      = dist.round(decimals=1)

# Add the nodes to graph
G=nx.Graph()
for i in range(len(loc)):
    G.add_node(loc.index[i])

pos = loc.ix[:,0:2].transpose().to_dict(orient='list')

# Add edges
for i in range(len(loc)):
    for j in range(len(loc)):
        G.add_edge(loc.index[i], loc.index[j], weight='%.2f' % d[i][j])

# Draw the network
fig, ax = plt.subplots(figsize=(5,5))
ax.axis('equal')
nx.draw_networkx(G, pos=pos, arrows= True, with_labels=True, node_size=600, node_shape='o', alpha=0.5, font_size=10)
edge_labels = {(u, v): d['weight'] for u, v, d in G.edges(data=True)}
nx.draw_networkx_edge_labels(G, pos, edge_labels=edge_labels, font_size=8, rotate=False)
plt.show()