Improve NWB explorer Python api

[filippomc]

NWB explorer makes available the loaded nwb file on the embedded Jupyter notebook through the variable nwbfile. The file can be processed through the pynwb api. We want the same feature of NetPyNE with plotting data from the python console as widgets. Before that we may need to connect the backend with the notebook kernel

• [ ] analyse the netpyne solution
• [ ] design the new backend/frontend communication
• [ ] add the plotting functionality

[filippomc]

Plotting into the javascript workspace from the kernel requires to set up a communication from the Python notebook kernel to the geppetto fronted. In order to set up a proper backend-frontend communication we need to revise the way that Geppetto communicates with the backend, now based on websocket for everything that is related to the project and model. We implemented the websocket methods but the websocket and the notebook are running on different kernels, which prevents from sharing the same variable space. We have two possibile solutions:

• Run the websocket directly inside the notebook kernel. This must happen on a different port.
• Move everything on the python-javascript message layer which is being used by NetPyNE