amzn-recs

Bootstrap from a clean DB

• make sure a Neo4j server is available in the location specified eg in .env

Project layout

• api: nodejs server to provide /crawl API
• lib:
• scripts:

Tests for api/

• make the API available on localhost: node api/crawl.js
• run the tests: mocha test/api_tests

Tests for lib/

• make sure a Neo4j server is available in the location specified eg in .env
• run the tests: mocha test/lib_tests

TODO

• move /crawl API to separate 'microservice'

• website to generate token by OAuth login (eg Google, Facebook etc as auth provider) -- any way to provide status updates?? (ie through queue...)

• API/script: add all books by author

• also maintain price history separately, and price last checked date in graphdb -- don't bother repeatedly checking price for anything < £2

• graph: series -- :IN_SERIES_WITH, properties sequence and series name

• script: expand entire graph by crawling each leaf node to depth n

Docker

Build a Docker image for this app:

1. Build image (optional, if image already exists) Tag the image with the short commit hash of HEAD revision used to build it: export COMMIT=$(git rev-parse --short HEAD) docker build -t amzn-recs-app:$COMMIT .
2. Run tests in a container using that image: docker run --rm --name populate amzn-recs-app npm test Or run eg the populate script in a container using that image: docker run --rm --name populate --net=host amzn-recs-app node scripts/populate.js