Skip to content

Network tables-based dashboard with customizable and switchable JSON layouts, H264 camera streaming, pose displays, and path generation/preview

Notifications You must be signed in to change notification settings

Spartronics4915/Dashboard

Repository files navigation

Dashboard

Repository with the Spartronics dashboard code. This dashboard is built atop the python implementation of networktables. Coupled with standard python webserver tech (here, tornado), the result is a webserver running on the driver's station that can read and control networktable values. Now, we can connect a standard web browser to the server and develop a custom dashboard built atop standard web technologies: html, css, javascript.

If you:

  • wonder why SendableChooser is so weird/flaky
  • were disappointed by the antiquated collection of interface elements available in SmartDashboard
  • were unwilling to invest precious time learning VFX/java GUI plugins

this approach may be for you.

To run the dashboard:

  1. make sure you have python3 installed (3.6, 3.7 tested)
  2. There is a requirements.txt file that will install versions of the required packages that are known to work. Install this with: py -3 -m pip install -r requirements.txt
  3. start dashboard server: python DashboardServer.py
  4. point your favorite web browser to http://localhost:5080, Chrome is recommended over Firefox due to improved performance of h264 streaming and faster canvas implementation.

Windows driver station

Drop a shortcut to start.bat in the startup folder: Win+R, enter shell:startup. Put a shortcut to Firefox there as well, and a shortcut to the driver station. The driver station shortcut has to be marked "Run as Administrator" in its Properties window. Also, disable UAC (the window-dimming 'are you sure' thing) otherwise the dashboard won't start up on boot.

Administration

We've found that Chrome is flaky when it comes to mjpg streams. On the other hand, we've found that Firefox is less stable for H264 and slower for canvas drawing. For 2019 season, we now recommend that you use Chrome as your primary web browser for this application.

Because so much code is reused across years, we chose to maintain a single repository for multiple years. We will tag the commit at the end of a year, to allow us to easily return to that point in history, then remove/rework the dashboard for the new game.

Upgrading pynetworktables requires simultaneous update of pynetworktables2js. Pip installs may or may not be in sync and may very according to python version. As of 1/20/2018, the python3 version of pynetworktables is still at 2017 level. For this reason python27 is preferred.

Dashboard access to robot logs may be impeded by firewall rules on your computer. Inbound UDP traffic is required.

Programming Note

From the pynetworktables2js docs:

NetworkTables is type sensitive, whereas Javascript is loosely typed. This function will not check the type of the value that you are trying to put, so you must be careful to only put the correct values (types) that are expected. If your robot tries to retrieve the value and it is an unexpected type, an exception will be thrown and your robot may crash. Make sure you test your code – you have been warned.

Helpful Links

python - is the language atop which the dashboard support resides.

pynetworktables - is the core technology. You can use standard 'pip' mechanism to install this module into your python interpreter:

python -m pip install pynetworktables

pynetworktables2js - is a small python module that forwards NetworkTables key/values over a Websocket, so that you can easily write a Driver Station Dashboard for your robot in HTML5 + JavaScript. Included in the package are JavaScript utilities to connect with the websocket. Install via:

python -m pip install pynetworktables2js

tornado - is a python package for webservice development.

Acknowledgements

Thanks to the robotpy maintainers for pynetworktables!

About

Network tables-based dashboard with customizable and switchable JSON layouts, H264 camera streaming, pose displays, and path generation/preview

Resources

Stars

Watchers

Forks

Packages

No packages published