A small web application developed as a requisite for a computer science course. Protein tokenizer fetches protein amino acid sequences from the UniProtKB database and trains a byte pair encoding tokenizer on the results. It is a minimum viable product written soley to practice implementing a microservice architecture.
requirements
- make
- python
- virtualenv
- pip
- node
setup
make install
tests
For unit tests:
make test/unit
For integration tests:
make test/integration
Note that the frontend, datacollector, and dataanalyzer need to be running to perform the integration test.
- first start rabbitmq, postgres, and prometheus
docker compose up
- run the services
make run backend
make run datacollector
make run dataanalyzer
make run frontend
- navigate to the frontend web address
- prometheus should be exposed at http://localhost:9090
This application is composed of four major components: a ReactJS frontend, a FastAPI backend, a data collector, and a data analyzer. The latter three services communicate with one another via AMQP (RabbitMQ). The frontend sends messages to the backend via a simple REST API and the backend responds aysnchronously through server-sent events. Persistent data in the form of search queries and search results are stored in a Postgres database. Endpoints are exposed for service monitoring, which is handled by Prometheus. Continuous integration and continous deployment are handled through Github Actions.
