Bokeh - 如何让 HoverTool 工具提示在点击时保持在点上?

Bokeh - how to make HoverTool tooltips stick to points on click?

以下代码来自jupyter notebook:

from bokeh.io import show, output_notebook
from bokeh.plotting import ColumnDataSource, figure
from bokeh.models import HoverTool, Range1d

output_notebook()
fig = figure(tools=[HoverTool(tooltips=[("html", '@html{safe}')])])

fig.quad(left="left", top="top", bottom="bottom", right="right",
         source=ColumnDataSource({"left": [1,3], "bottom": [1,3],
                                  "right": [2,4], "top": [2,4],
                                  "html":["<b>I'm bold</b>", "<span 

style='color:red;font-size:32px;'>BIG RED TEXT</span>"]}))
    show(fig)

我需要让 HoverTool 工具提示准确地固定在单击该点时的位置,因此如果用户想要突出显示并复制工具提示中的文本,他们可以这样做。这个 codepen has an example of the type of behavior I would like to see. I know that this must be possible by either injecting some type of CustomJS or altering BokehJS coffescript 并从头开始构建 BokehJS,但我一直无法弄清楚。有没有人知道如何做到这一点?

更新:
可能会创建一个 custom tool using the tap_tool.coffee, hover_tool.coffee or tooltip.coffee。如果我弄明白了,我会更新这个。

这是在 models.HoverTool 的 'callback' 函数中使用 models.Label 创建您自己的悬停文本的解决方法。 models.TapTool 也用于切换更新标签文本。您无法编辑字形 models.Label 中的文本,但添加了一个 TextInput widget,当鼠标悬停在图形字形中时,文本会更新。

import bokeh
import bokeh.plotting
fig = bokeh.plotting.figure()

d_source = bokeh.models.ColumnDataSource({"left": [1,3,1],
                "bottom": [1,3,3],"right": [2,4,2],"top": [2,4,4]})
h_source = bokeh.models.ColumnDataSource({"do_hover":[True,True]  })
fig.quad(left="left", top="top", bottom="bottom", right="right",
         source=d_source)

myToolTip = bokeh.models.Label(x=2.5,y=2.5, text="",
                               background_fill_color = "#ffffff")

HoverCallback = bokeh.models.CustomJS(args=dict(d_source=d_source,
                    myToolTip=myToolTip,h_source=h_source),code="""
        function findWhereIam(x,y,s){
            // To find where the cursor is. 
            selection = -1;
            for (i = 0; i < s.data.left.length; i++) {
                x0 = s.data.left[i];
                x1 = s.data.right[i];
                y0 = s.data.bottom[i];
                y1 = s.data.top[i];
                if (x>x0 && x<x1 && y>y0 && y<y1){
                    // It's inside rectangle!!!
                    selection = i;
                }
            }
            return selection
        }
        if (h_source.data.do_hover[0]){
            x_data = cb_data['geometry'].x;
            y_data = cb_data['geometry'].y;
            var selection = findWhereIam(x_data,y_data,d_source)
            if (selection>=0){
                x0 = d_source.data.left[selection];
                x1 = d_source.data.right[selection];
                y0 = d_source.data.bottom[selection];
                y1 = d_source.data.top[selection];
                myToolTip.x = 0.5 * (x0+x1);
                myToolTip.y = 0.5 * (y0+y1);
                myToolTip.text = "on:"+selection;
                myToolTip.text_font_size = "24pt";
                myToolTip.background_fill_color = "#ffffff";
                myToolTip.border_line_color = "#000000";
                myToolTip.text_align = "center";
                myToolTip.trigger('change');

                current_selection.value = myToolTip.text;
                current_selection.trigger('change');
            }else{
                myToolTip.text = ""; //erase
                myToolTip.trigger('change');
                current_selection.value = myToolTip.text;
                current_selection.trigger('change');
            }
        }
    """)

TapCallback = bokeh.models.CustomJS(args=dict(h_source=h_source), code="""
        h_source.data.do_hover[0] = !h_source.data.do_hover[0];
    """)

current_selection = bokeh.models.widgets.TextInput(value="",
                                                   title="Selection")

HoverCallback.args.update(dict(current_selection=current_selection))

fig.add_tools(bokeh.models.HoverTool(tooltips=None,callback=HoverCallback))
fig.add_tools(bokeh.models.TapTool(callback=TapCallback))
fig.add_layout(myToolTip)
page = bokeh.plotting.gridplot([[fig],[current_selection]])
bokeh.io.output_notebook()
bokeh.io.show(page)