使用networkx + graphviz定位/显示标签 [英] Position/showing of labels with networkx + graphviz

问题描述

我已经结合使用networkx和graphviz来完成以下绘图:

I've achieved the following plot with a combination networkx and graphviz:

我对结果感到非常满意.在该图中，您可以识别出我所说的聚合节点:那些是最新的绿色节点(所有绿色节点都汇聚在其中)比橙色节点早一跳.

I'm very happy with the result. In the plot you can identify what I call aggregation nodes: those are the latest green ones (where all the green nodes converge) one hop before the orange ones.

我要实现的目标如下:

1)在节点的侧面放置标签.如您所见，标签在它们的上方，很难阅读；

1) Put labels on the sides of the nodes. As you can see, the labels are over them and it's difficult to read;

2)仅在聚合节点和橙色节点上显示标签.

2) Only show labels on the aggregation nodes and the orange ones.

这就是我绘制图表的方式.

This is how I get to plot the diagram.

# We create the graph
G = nx.DiGraph()

# We add nodes and edges
G.add_nodes_from(nodes)
G.add_edges_from(edges)

# We establish attributes to nodes
nx.set_node_attributes(G,nodesAttrDic)

# Tune plot
nodeFontSize    = 10
nodeSize        = 20
nodeColorList   = list(getNodeColor(nodesAttrDic,G.nodes()))
edgeColorList   = getEdgeColor(G.edges())

# Graphiz tunning
prog = 'dot'
args = '-Gnodesep=1 -Granksep=2 -Gpad=0.5 -Grankdir=TD'
root = None
pos  = graphviz_layout(G, prog = prog, root = root, args = args)

nx.draw(G,
    pos         = pos,
    with_labels = True, 
    node_color  = nodeColorList, 
    edge_color  = edgeColorList, 
    font_size   = nodeFontSize,
    node_size   = nodeSize,)

plt.show()

关于如何执行此操作的任何想法?

Any ideas on how to do this?

谢谢！

卢卡斯

推荐答案

好，所以我已经部分解决了第二个问题:如何从聚合节点开始绘制.为此，我估计了最近一跳的跳数.之后，我决定标记低于阈值的节点.

Ok, so I've partially solved the second question: how to plot from the aggregation nodes onwards. To do so I estimate the number of hops towards the latest one. After that I decide to label the nodes below the threshold.

def getHopToNH(nodes):
    path        = []
    labelList   = {}

    for startNode in G.nodes():
        endNode = 'myLabel'
        try:
            p = len(nx.shortest_path(G,source=startNode,target=endNode))
        except:
            p = -1
        path.append((startNode,p))

        if p < 8:
            labelList = {**labelList,**{str(startNode):str(startNode)}}
        else:
            labelList = {**labelList,**{str(startNode):''}}

    return labelList

更新:

现在，为了重新放置标签，我必须自己修改位置.

Now, in order to re-position the labels, I had to modify the position myself.

for p in pos:

    yOffSet = -300
    xOffSet = -400

    pos[p] = (pos[p][0]+xOffSet,pos[p][1]+yOffSet)

labelDescr = nx.draw_networkx_labels(G,
    pos         = pos,
    labels      = nodeLabelDict,
    font_size   = nodeFontSize,)

for n,t in labelDescr.items():
    finDegree = 70
    t.set_rotation(finDegree)

在此之后，我将绘制以下内容:

After this, I get to plot the following:

我现在真的很喜欢这个输出...:-)

And I really like this output now ... :-)

这篇关于使用networkx + graphviz定位/显示标签的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！