Bokeh 动态改变 BoxAnnotation

Question

是否可以在 IPython 的交互功能中更新 Bokeh 图的渲染。我有如下代码:

x = [0, 1, 2, 3, 4]
y = [0, 1, 2, 3, 4]
source = ColumnDataSource(data=dict(x=x, y=y)
f = figure()
f.line(x, y, source=source)
show(f)

def update_func(selected_data):
   source.data['y'] = ...
   source.push_notebook()
   <here I would like to add BoxAnnotation to figure f, and rerender it>

interactive(update_func, selected_data=[0,1,2])

Answer 1

截至 Bokeh 0.11 push_notebook函数现在可以更新任意 Bokeh 模型对象。因此，只需设置/更新框注释上的属性并调用 push_notebook .您的代码将类似于:

x = [0, 1, 2, 3, 4]
y = [0, 1, 2, 3, 4]
source = ColumnDataSource(data=dict(x=x, y=y)
f = figure()
f.line(x, y, source=source)
show(f)

def update_func(selected_data):
    source.data['y'] = ...
    # update box annotation (or any other properties) here
    push_notebook()  # note, just a function, not a method on "source"

interactive(update_func, selected_data=[0,1,2]) 

你可以在这里看到一些howto notebooks:

https://github.com/bokeh/bokeh/blob/master/examples/howto/notebook_comms/Basic%20Usage.ipynb

https://github.com/bokeh/bokeh/blob/master/examples/howto/notebook_comms/Continuous%20Updating.ipynb

https://github.com/bokeh/bokeh/blob/master/examples/howto/notebook_comms/Jupyter%20Interactors.ipynb