Skip to content

Latest commit

 

History

History
87 lines (65 loc) · 3.05 KB

README.md

File metadata and controls

87 lines (65 loc) · 3.05 KB
Raw

Times Square Demo Notebooks for USDF

These are notebooks that demonstrate Times Square using production-grade datasets at the Rubin USDF.

View online: https://usdf-rsp-dev.slac.stanford.edu/times-square/

Contributing

Setting up the repository

Clone this repository:

git clone https://github.com/lsst-sqre/times-square-usdf
cd times-square-usdf

Install the pre-commit hooks:

pip install pre-commit
pre-commit install

Adding a notebook

All ipynb Jupyter Notebook files in this repository's main branch are eligible to be published to Times Square on USDF Dev at https://usdf-rsp-dev.slac.stanford.edu/times-square/. Setting up a notebook requires two files: the ipynb file itself and a yaml file that provides metadata for the notebook. Both files must have the same root name (e.g my-notebook.ipynb and my-notebook.yaml). Notebooks can be organized in directories, and those directories are reflected on Times Square.

Preparing the notebook

The first code cell of a Times Square notebook must exclusively assign values to parameters. When Times Square runs a notebook on the Rubin Science Platform, it will replace this code cell with variable assignments requested by the user. For example, if the notebook uses a date variable as a string and a threshold variable as a float, the first code cell should look like this:

date = '2023-01-01'
threshold = 0.5

The notebook can then use these variables in subsequent cells.

Markdown cells can also use these variables as Jinja variables in the params namespace. For example, a title markdown cell might be:

# Report for {{ params.date }}

The notebook must be able to run independently from the standard Notebook Aspect kernel. Your notebook can't load files from this repository, rely on private files in a specific user's directory, or import Python packages that aren't available in the standard Notebook Aspect kernel.

Preparing the metadata

The yaml file that corresponds to the ipynb file must contain the following keys:

  • title: The title of the notebook, as it will appear on Times Square.
  • description: A short description of the notebook, as it will appear on Times Square.
  • authors: A list of authors for the notebook, as they will appear on Times Square. Each item is a dictionary with name and slack keys.
  • tags: A list of tags for the notebook, as they will appear on Times Square.
  • parameters: A mapping of parameters that the notebook accepts, as they will appear on Times Square. The key of each parameter is its Python variable name, and maps to type, description, and default keys. The type must be one of string, integer, number, or boolean. More types are coming

For example:

title: My Notebook
description: A notebook that does something.
authors:
  - name: An Engineer
    slack: @engineer
tags:
  - demo
parameters:
  date:
    type: string
    description: The date to use.
    default: 2023-01-01
  threshold:
    type: number
    description: The threshold to use.
    default: 0.5