我正在尝试使用带有 Bokeh 的 Holoview 制作和弦图。如here所述，我可以做第一步；但是，我无法创建组，也无法显示颜色和标签。首先，我的数据 *chord = * ：
[['GK', 'WD ', 2],
['GK', 'C ', 1],
['GD', 'GK ', 1],
['GD', 'C ', 2],
['GD', 'WA ', 4],
['WD', 'GD ', 4],
['WD', 'C ', 1],
['WD', 'WA ', 2],
['WD', 'GA ', 1],
['C', 'GD ', 1],
['C', 'WD ', 1],
['C', 'WA ', 3],
['C', 'GA ', 2],
['C', 'GS ', 10],
['WA', 'C ', 2],
['WA', 'GA ', 9],
['WA', 'GS ', 7],
['GA', 'C ', 3],
['GA', 'WA ', 4],
['GA', 'GS ', 5],
['GS', 'C ', 6],
['GS', 'WA ', 1]]
值是[源、目标、值]。让我陷入困境的是 Chord 图表不会自动对源进行分组，如果我改用 Sankey 图表就会发生这种情况。所以，我创建了我的组和标签：
group = [1,1,2,2,2,3,3,3,3,4,4,4,4,4,5,5,5,6,6,6,7,7]
label = ['GK','GK','GD','GD','GD','WD','WD','WD','WD','C','C','C','C','C','WA','WA','WA','GA','GA','GA','GS','GS']
index =list(np.arange(0,22)
然而，结果远非预期：
import numpy as np
import pandas as pd
import holoviews as hv
hv.extension('bokeh')
%output size=200
chords = pd.DataFrame(chord, columns=['source','target','value'])
%opts Chord [label_index='source' color_index='target' edge_color_index='source']
%opts Chord (cmap='Category20' edge_cmap='Category20')
hv.Chord(chords)
如果我通过
hv.Chord(chords, nodes)
我收到了这个错误：
AttributeError: 'DataFrame' object has no attribute 'ndims'
谢谢您的帮助！