Pete has just been hired as Product Manager for Search at electronics retailer Chorus Electronics. His boss, a Vice President, has asked him to build a ‘best-in-class’ e-commerce search engine, to ‘increase customer satisfaction and drive improved revenue’.
An open source e-commerce search engine
Pete is considering using an open source search engine at Chorus Electronics: but all he sees out there is a kit of parts, not a complete software solution. Luckily Pete has a great technical team, but he doesn’t have the budget or time for them to build something from scratch.
In this post we’re going to show Pete’s team how to use a collection of open source, freely available software components we’ve named Chorus to build a complete, tuneable, testable online shop. Note that for a real world application you’d have to add a a few other components – for example a shopping cart – but Chorus is a great start and acts as a reference implementation for the search tuning methods we’ll be demonstrating in the next few posts.
All the components we’ll select are in being used in production, today, by organisations across the world, often at huge scale:
- Project Blacklight for the user interface
- Apache Solr as the search engine
- Querqy & SMUI for search management (boosts, synonyms etc.)
- Quepid for gathering relevance ratings & testing new configurations
- RRE for large-scale regression testing with detailed metrics
All are open source and freely available for download. Chorus contains customised versions of these components & sample data.
See how it’s done
At this point Pete is going to hand things over to his technical team – if you’d also like to install Chorus, you’ll need to be comfortable with the Linux command line, Docker and Git. You’ll need to start by cloning or downloading the Chorus Github repository which contains a README file – you can follow the instructions under ‘Getting Set Up to Play with Chorus’.
Eric Pugh has also recorded a video that shows all these steps – even if you’re not going to be installing Chorus yourself, you can see how it can be done in under 20 minutes and you can take a first look at the components in action: