sandbox.bio

Interactive bioinformatics command-line tutorials.

---

Local development

Branches

• main: Development branch, pushing there runs tests
• stg: Staging branch, merge into this branch to deploy to stg.sandbox.bio
• prd: Production branch, merge into this branch to deploy to sandbox.bio

Environment setup

Define environment variables in .env:

# Supabase database URL
PUBLIC_SUPABASE_URL=...
# Supabase database public key
PUBLIC_SUPABASE_API_KEY=...
# Supabase database admin key
SUPABASE_API_KEY=...

Launch the web server:

npm install
npm run dev

Tests

• npm run test will launch all the tests in headless way
• npx playwright test --ui opens Playwright UI

Deploy

Generate Debian assets:

# Generate static assets (see https://github.com/sandbox-bio/v86/blob/master/NOTES.md)
git clone https://github.com/sandbox-bio/v86.git && cd v86
make all
make build/xterm.js

# Generate .bin files
cd tools/docker/debian/
./generate.sh

Deploy:

# Upload .bin files
export CLOUDFLARE_ACCOUNT_ID=ID_GOES_HERE

# Deploy main to stg
SOURCE=main
DEST=stg

# Deploy stg to prd
SOURCE=stg
DEST=prd

# Deploy assets and new debian image first (to avoid users caching a 404 for a non-existent image)
./bin/deploy-v86.sh $DEST
# Deploy code changes
git push origin --delete $DEST
git checkout -b $DEST --track origin/$SOURCE
git push origin $DEST
git checkout main
git branch -d $DEST

---

Infrastructure

Database

Table	Description	Access
logs	Log all calls to sandbox.bio/*	RLS
pings	Analytics for tutorial progress	RLS
state	Save tutorial progress	RLS

Append _stg to table names for dev/stg environments.

Grafana

Dashboard

• Import grafana.json into new Grafana instance: Depending on the setup, you might need to find/replace the postgres connection uid and replace it with the uid of the connection created.
• Export grafana.json to repo: Remove top-level id, uid, and version fields to avoid the import error The dashboard has been changed by someone else.