LENS2-Dashboard

This repository hosts the code for a dashboard for the LENS2 dataset using FastAPI and D3.js.

To get started, clone this repository using git clone --recurse-submodules https://github.com/NCAR/LENS2-Dashboard-D3.git

Recreating the environment

The environment.yml will allow you to recreate the environment required for this project. You will need conda installed, which you can find here. This tutorial on conda by ProjectPythia will help you install and understand the basics of environment management with conda.

Once you have conda installed, you can use the following commands to recreate the environment.

1. cd LENS2-Dashboard-D3 && conda create --prefix ./.env -c conda-forge mamba This will create a directory called .env in your local copy of the LENS2-Dashboard-D3 repository and install mamba in it. mamba is faster than conda in solving environments.
2. conda activate ./.env && mamba env update --file environment.yml --prefix ./.env - This will start creating the environemnt in the .env directory based on the environment.yml file.
3. conda activate ./.env and voila! you just activated the environment required to run the back-end server. You will still need t

Get the d3-CESM-tiling submodule (if did not use --recurse-submodules while cloning the repository)

This back-end is tied to the front-end code hosted at the NCAR/d3-CESM-tiling repository. The back-end imports this code as a git submodule. The submodule contents can either be fetched during cloning the back-end this repository by specifying the --recurse-submodules flag, or manually by executing the following command: git submodule update --init

How to start the back-end server

Once you have the environment activated, the following commands are required to (1) start a dask cluster, (2) start some dask workers, and (3) start the fastapi server.

Scheduler

dask scheduler --host localhost --port 8786

Workers - 16 workers, 2GiB memory

dask worker --host localhost --nworkers 16 --memory-limit '2GB' localhost:8786

Run the server in development mode - if you're modifying the back-end code

uvicorn --reload app:app

Run the server in production mode - if you're hosting the back-end server in production

gunicorn -k uvicorn.workers.UvicornWorker app:app