Aquila: Spatial single-cell database and analysis platform

Data preproces /aquila-preprocess

The scripts are written in Python, the database use PostgreSQL.

API Server aquila-api and aquila-fastapi

The main API Server is written in Rust using Actix.

The side API Server is written in Python using FastAPI to run following analysis:

• Spatial variable gene: SpatialDE
• Ripley's Function K, F, G, L
• Spatial community detection
• Cell centrality

Two server is registered under the same domain: https://api.cheunglab.org through reverse proxy provided by Nginx.

To run a dev build with HMR

cd aquila-api
cargo watch -x 'run --bin aquila-api' --workdir .

To run a production build

cargo build --release && ./target/release/aquila-api

The computed server for spatial variable gene is in FastAPI and run as docker image.

UI aquila-ui

The UI part is implemented using Next.js, the analysis part is powered by indexedDB.

We also use WebWorker to run tasks parallel on the web.

To run a dev build with HMR

8G RAM is needed to run a dev build,

$Env:NODE_OPTIONS="--max-old-space-size=8192" # powershell
export NODE_OPTIONS="--max-old-space-size=8192" # linux

cd aquila-ui
yarn dev

Deployment

The UI part will be deployed automatically by Vercel.

To deploy actix server

cd aquila-api

SERVER_PID="$(pidof aquila-api)"

echo "Terminate current running process at $SERVER_PID"

kill $SERVER_PID

cargo build --release

nohup ./target/release/aquila-api > log.out 2>&1 &