A PHP framework for building complex recommendation engines on top of Neo4j graph databases.
GraphAware Reco4PHP is a PHP framework for building complex, production-ready recommendation engines on top of Neo4j graph databases. It solves the problem of architecting scalable recommendation systems by providing a structured pipeline for discovering, filtering, and scoring items, handling all the database plumbing so developers can implement custom business logic.
PHP developers and data engineers who need to build personalized recommendation features (like "users who bought this also bought" or "people you may know") within applications backed by a Neo4j database.
Developers choose Reco4PHP because it offers a battle-tested, opinionated framework specifically for Neo4j, eliminating the need to build the complex recommendation pipeline from scratch. Its modular design (Discovery Engines, Filters, Post-Processors) provides flexibility while ensuring maintainability and performance for graph-based recommendations.
Neo4j based Recommendation Engine Framework for PHP
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Enforces a clear separation with Discovery Engines, Filters, and Post-Processors, making recommendation logic modular and maintainable, as demonstrated in the MovieLens example pipeline.
Handles Neo4j Cypher transactions automatically, reducing boilerplate code and ensuring data consistency during recommendation computations.
Provides built-in capabilities to evaluate recommendation quality, essential for tuning and improving engines over time.
Allows developers to implement custom business logic at each pipeline stage, enabling tailored recommendations for specific use cases like rewarding same-city users.
The framework is tightly coupled to Neo4j, making migration to other databases nearly impossible without significant rewrites, limiting technology flexibility.
Requires developers to write and optimize custom Cypher queries for discovery and scoring, which demands deep Neo4j expertise and can be error-prone.
Focuses primarily on graph-based recommendations; lacks built-in support for other common techniques like collaborative filtering or deep learning models.
Involves a non-trivial configuration process, including Neo4j dataset import and custom class wiring, which can slow down prototyping compared to drop-in solutions.