Js_move = 'if(cb_obj.x >= fig.x_range.start & cb_obj.x = fig.y_range.start & cb_obj.y <= fig.y_range. This answer is for the people who liked Graeme's solution just as I did, but need to apply it to more than two figures just as I did: from bokeh.models import CustomJS, CrosshairTool To reduce to just one dimension (vertical or horizontal) remove the corresponding "if / else" part of the callback
PYTHON BOKEH TUTORIAL PLAYLIST CODE
Source = ColumnDataSource(įigure.js_on_event('mousemove', CustomJS(args = args, code = js_move))įigure.js_on_event('mouseleave', CustomJS(args = args, code = js_leave)) # change just some things about the y-gridĭec = Span(location=datetime(year-1, 12, 1, 0, 0, 0).timestamp() * 1000,ĭimension='height', line_color='grey', line_dash='dashed', line_width=1) # change just some things about the x-grid X_range=x_range, y_range=y_range, tools=TOOLS)
HoverTool(tooltips=[("Dato", Setting up the bokeh figureįig = figure(width=800, height=250, title=title, x_axis_type="datetime", # TOOLS = "resize,crosshair,xpan,xwheel_zoom,box_zoom,reset,save" # TOOLS = "resize,crosshair,pan,wheel_zoom,box_zoom,reset,box_select,lasso_select,save" Here is the plot setup and an example: #!/usr/bin/env pythonįrom otting import figure, ColumnDataSource, output_file, saveįrom bokeh.models import Span, CrosshairTool, HoverTool, ResetTool, PanTool, WheelZoomToolĭef timeline_figure(title=None, x_range=None, y_range=None): I'd just be a matter of getting the plot to refresh everytime time you use an intractable object, which might be resource intensive depending on your data set.When moving the crosshair (dimensions=width) in one plot I want to see the same position in the other plot(s). Hence, it proves to be extremely useful for developing web based dashboards. Unlike Matplotlib and Seaborn, they are also Python packages for data visualization, Bokeh renders its plots using HTML and JavaScript. I'd probably just have the plotting resource output an image into my GUI and make it intractable with PyQT objects. Bokeh Introduction - Bokeh is a data visualization library for Python. I imagine you could output your plot directly into your PyQT GUI and build in interaction WAY easier that way. You'd probably have better luck using a more basic plotting library (like matplotlib) with it. PyQT is my favorite GUI resource in python even though the documentation is kinda ass (since it's built on C++ and is typically used in C++). I imagine it would be difficult, but probably do-able? Idk.
I have 0 experience with Javascript and I didn't want to get any so I never tried it. That said though, Bokeh outputs to html, so getting it to interact in any direction will require inbedded Javascript in your bokeh code.
I haven't messed with that tool in a while. It's like the power of Javascript at your fingertips - but Pythonic.ĭisclaimer: I am not associated with Bokeh in any way I'm just excited. Plotly is pretty good too I admit, but Bokeh from the ground up clearly was developed by developers for developers with elegance, power, and all-around bad-assery in mind - it is ridiculous. You have to constantly decide if you want %matplotlib inline (dead and lifeless) or %matplotlib qt (zero persistence between runs). It has a really weird api (for instance: oh you created a plot one way use plt.title() oh wait you created it another way sorry use ax.set_title() because there are two names for everything because reasons). I realize matplotlib is ok, but when in notebook environments it sort of sucks. Seriously why aren't more people using this library? I'm working through the tutorials now:Īnd, given that so many tools are notebook-driven now and that the business-heads always want dashboards and web-facing blah blah blah, why isn't Bokeh the shiznit?
0 kommentar(er)
