First, we need data for our application. We'll make use of the Neo4j Recommendations Sandbox, which contains a dataset of movie reviews.
The goal for this section of the workshop is to familiarize yourself with Cypher and Neo4j by writing Cypher queries to support the following features:
To support the features of our web application, we need to
- Search for movies by title substring and return movie details
- For a given movie, find other recommended movies
We can use the CONTAINS string comparison operator in Cypher to find movies whose title contains a search string. We'll use the toLower() function to make our comparison case insensitive.
MATCH ... WHERE ... CONTAINS $subString RETURN ...
For a given movie, what are other movies that someone who likes that movie also likely to enjoy? We use a combination of overlapping genres and actors (similar movie) and a simple collaborative filtering query (users who like this movie also like...).
MATCH ...
...
RETURN ... ORDER BY score DESC ...
If you get stuck, you can find the completed queries here. Note that the recommendations query could be completed several different ways - there is no correct answer as long as your queries returns Movies!
Once you have these queries completed, we're ready to integrate them into our GraphQL service.
Continue to the GraphQL section of the workshop.
If you are unable to use the Recommendations Neo4j Sandbox, you may download the database here.
To run neo4j locally with the above dataset, you may try these commands. They will download the dataset, and then launch neo4j inside of a docker image using that data.
curl https://s3.amazonaws.com/neo4j-sandbox-usecase-datastores/v3_2/recommendations.db.zip > /tmp/recommendations.zip
cd /tmp
unzip recommendations.zip
mkdir -p neo4j/data/databases/
mv recommendations.db neo4j/data/databases/graph.db
docker run \
--publish=7474:7474 --publish=7687:7687 \
--volume=/tmp/neo4j/data:/data \
neo4j:3.2.3
You can then change the default neo4j password by visiting http://localhost:7474/browser/