Dynamic change scale from linear to log

Community Support
1

Hi,

After searching in various places, I still don’t really know if it’s possible to dynamically change a figure scale (in my case from linear to log). As a turnaround, I’m using a panel with two tabs but I would love to just have the possibility to switch between the two scales using a radio buttons group for example. Am I missing something obvious ? I tried to change the y_scale attribute but nothing is modified when it’s done.

Thank you

2

It is definitely possible - I’ve done it before with a separate toolbar button. Can’t get to the code right now, so will post it tomorrow.

3

You can do something like below:


def function_to_change_log():
    if log_hist_cb.active:
           if log_hist_cb.active[0] == 0:
                        log_hist = numpy.log(your_dataframe)
                            
log_hist_cb = CheckboxGroup(
            labels=["Log transform axis"], active=[])

The example is here: https://github.com/samirak93/ML_Tool/blob/ff24f178afc03ff8aeb1ba6d76f930353a8c17f8/main.py#L367

4

It doesn’t change the axis/scale, it changes the data. Often, that’s undesirable.

1 Like
5

Ah, sorry - my existing code just re-renders the whole plot with new parameters, one of which is y_axis_type.

6

Just tried to do that without re-rendering the whole plot, and now I remember why it didn’t work.
The issue is that PlotView stores all the scales internally when it’s initialized. You cannot overwrite them without creating a custom Plot class.

7

Thanks for the answers. I will try that ASAP.

8

Hi, I’m pretty interested to try this, could you give me some guidelines of which files should I look?

Thanks in advance

9

Here it is: bokeh/plot_canvas.ts at 6bf049cff1b8e16d74dc38b799fc836d76ac7fd0 · bokeh/bokeh · GitHub