Hello all,
Gentle reminder regarding previous email.
Also, I have implemented a Plotly based frontend and Bokeh server as backend. Please see [ 1].
For demo: Go to [ 2] for Bokeh based output and [ 3] for Plotly based output. In both cases, backend is handled by Bokeh.
Procedure to setup on your system: - Turn on Bokeh server
- Run the top_block.py.
- You can review Bokeh based output at localhost:5006
- Also, opening plotly.html will display the plot using plotly.js frontend. You may have to change the socket URL in plotly.html. It can be automatically configured but I wanted to have a minimalistic prototype.
Now, I think this example solves the query of independent frontend and backend. As shown in the example, using the socket provided by Bokeh server, one can use any frontend library for plotting.
Also, any change in connected client will be pushed by Bokeh server. So, the person changing the frontend library (in future) will have to handle such changes using _javascript_.
So, I think, now we can easily use Bokeh backend for the task. Instead of writing server application on our own, Bokeh will be more structured and efficient. What is your suggestion on this?
|