Skip to content
Snippets Groups Projects

nuBoard

Welcome to the nuBoard repository!

nuBoard is a Python repo based on Bokeh. It is a visualization tool for nuPlan metrics and simulation.


nuBoard structure

Our code is organized in these directories:

nuboard
├── base              - Data and base classes shared by other code. 
├── resource          - Resources such as images, javascript and css files.
├── tabs              - Python code for each page.
├── templates         - HTML code for each page.
├── utils             - Common fucntions used by other code.
├── nuboard.py        - Code to start a Bokeh application server.
└── style.py          - Default color and size for templates.

Usage

To start using nuBoard, please run the following command in cli:

  • python nuplan_devkit/nuplan/planning/script/run_nuboard.py
  • nuBoard takes nuboard files from simulation as input. To import nuboard files, we have two options:
    • GUI: Use the upload file button in the side navigation. However, this method is not supported if:
      • Run nuBoard on a remote server.
      • Move simulation data to different folders.
    • Command line: Set simulation_path in the command:
      • python nuplan_devkit/nuplan/planning/script/run_nuboard.py simulation_path="[path_1, path_2]"

Pages

There are three pages in nuBoard, which are Overview, Histograms, and Scenarios.

  • Overview:
    • Show a summary of aggregated metric scores for each scenario type and planner.
  • Histograms:
    • Display distributions of available metric scores for different planners across all scenarios.
  • Scenarios:
    • Visualize simulation and metric scores for a selected scenario.