pipes

Neo4j API offers methods for graph traversal out-of-the-box which are likely the same ones that most people will use. Interestingly, Marko didn't use Neo4j's native API (http://api.neo4j.org) -- he used a dataflow framework he wrote called Pipes (https://github.com/tinkerpop/pipes/wiki/). You probably have heard of the graph programming language Gremlin (https://github.com/tinkerpop/gremlin/wiki). Gremlin is a thin wrapper over Pipes.

That's a great book. The difference is that the examples in Programming Collective Intelligence don't make use of a real graph database (meaning a DB where you don't have to hit an external index to find adjacent nodes), and all of the algorithm examples are in standard Python rather than in a more expressive graph programming language like Gremlin (https://github.com/tinkerpop/gremlin/wiki). Gremlin is a domain-specific language written in Groovy (it's like SQL for graphs), and you can use it from Python like you can use SQL from Python (see http://bulbflow.com/overview); however, there is a Jython version in the works, as well as a Scala version (https://groups.google.com/d/msg/gremlin-users/LXNdbT_m7Gk/QZ...). Marko is the author Gremlin, and he has been working to refactor most of its heavy lifting down into a dataflow framework calle