Spread

Spread vizualizes how viruses and other pathogens are spreading in time and space. It creates shareable, interactive and time-animated vizualisation.

Table of Contents

• Citing Spread
• Technical Overview
• Development
• Contributors
• License

Citing Spread

@article{10.1093/ve/veac088,
    author = {Nahata, Kanika D and Bielejec, Filip and Monetta, Juan and Dellicour, Simon and Rambaut, Andrew and Suchard, Marc A and Baele, Guy and Lemey, Philippe},
    title = "{SPREAD 4: online visualisation of pathogen phylogeographic reconstructions}",
    journal = {Virus Evolution},
    volume = {8},
    number = {2},
    year = {2022},
    month = {09},
    issn = {2057-1577},
    doi = {10.1093/ve/veac088},
    url = {https://doi.org/10.1093/ve/veac088},
    note = {veac088},
    eprint = {https://academic.oup.com/ve/article-pdf/8/2/veac088/46671630/veac088.pdf},
}

Technical Overview

To be able to easily pick up stack used to build spread, one should be familiar with following topics:

• Java
• Clojure
• ClojureScript
• GraphQL
• MySQL
• Docker
• S3
• SQS

The diagram below presents an overview of the architecture of spread:

• API is a gateway service, exposing graphql endpoints and publishing messages to the SQS queue.
• Worker is a messaging service wrapping the phylogeographic tree graphs parsing library libspread, multiple workers compete for the messages published by the API.
• Relational Database and S3 object storage are used for persistance.
• User-facing interface facilitates communicating with the API from the client.
• Visualization engine accepts S3 stored output, animates and displays it on the geographical map.

Development

Start all

Make sure you have tmux multiplexer installed. Execute ./start_all_components in your terminal window and it will start all the components in separate tmux windows.

Backend services

Make sure you have docker and docker-compose installed. You will also need maven and clojure cli-tool.

Source the environment variables and start the dev infrastructure:

source env/dev
docker-compose -f deployments/dev/docker-compose.yml up

Deploy the database schema changes:

cd services/db-migration
source ../../env/dev
mvn package
mvn liquibase:update

Compile and package libspread:

mvn clean package

Start an instance of the worker-service:

clj -A:run-worker

Start an instance of the api-service from the comand-line:

clj -A:run-api

Start an instance of the api-service from the REPL:

M+x cider-jack-in
C-c M-n-n api.main
(restart)

In the default dev environment a GraphQL IDE is started at: http://127.0.0.1:3001/ide

Browser client

Make sure you have yarn installed.

Install dependencies:

yarn deps

Start watcher and local server

yarn watch

Open in browser: http://localhost:8020

To get the cljs REPL:

M+x cider-connect-cljs

Select localhost and the nREPL port printed by the watcher (e.g. 46000), select shadow and :ui as the build.

Viewer browser client

Start watcher and local server

yarn watch:viewer

Tests

Start a watcher on the libspread test suite:

mvn fizzed-watcher:run

Start a watcher on the spread's integration test suite:

bin/kaocha --watch

Contributors

This project exists thanks to all the people who contribute.

License

MIT © Filip Bielejec