Ogre is a domain specific language for traversing property graphs in Clojure. Ogre wraps Gremlin, a library which enables all sorts of groovy ways to work with graphs. The documentation and samples presented here attempt to stay current with the most current, stable release of Ogre. Please join the Titanium users group for Ogre related discussions. Please use the Ogre issue page for reporting bugs and discussing features. For any errors or corrections with this documentation, please use the issue page. Pull requests will be celebrated, scrutinized, and hopefully accepted. Ogre currently powers Archimedes, a Clojure library for Blueprints, and Titanium, a Clojure library built on top of Archimedes for working with Titan.
Before going down the rabbit hole, we offer the briefest of warnings:
the Tinkerpop folks have
been working on Gremlin,
Blueprints for a few
years now and the stack has become incredibly intertwined. Ogre and
Archimedes try to hide
all of this from you at some hide level, but every abstraction leaks.
Your stack traces will speak of
com.tinkerpop.gremlin.GremlinPipeline, and there is nothing we can
(well, should) do about it. The Tinkerpop stack is fantastic and
really well done, but we wanted to warn you that the following is just
the tip of the iceberg in terms of what you will need to know to
understand what is really going on when you use Ogre.
Thankfully, it's a joy to work with the stack. So, let's get started!
The version scheme for Ogre is as follows:
[Full Gremlin version].[Ogre major version]. Philosophically, the
authors of Gremlin are the ones who will be causing most of the
changes to Ogre to hapepn. Ogre is a mere wrapper around their work
and the version scheme acknowledges that directly. Thus, the current
184.108.40.206, meaning that the current release uses Gremlin
2.3.0 and has undergone one major versions itself so far since
To get started with Ogre, include the following dependency for
Unless otherwise noted, all samples reference
ogre.core as follows:
(require '[ogre.tinkergraph :as g]) (require '[ogre.core :as q]) (g/use-new-tinker-graph!)
We recommend that you open this image up into a new tab. It will serve as the main reference for the majority of the examples below.
At a high level, Ogre let's you easily ask complex questions about certain types of graphs and get back answers. That's it.
At a low level, Ogre is a library that takes in Blueprint Vertices and Edges and let's you build up GremlinPipeline objects that ask questions about those objects in the language of traversals, transformations, filters, and branching on the graph. Ogre allows you to annotate various steps of the pipeline to allow for incredibly useful queries in a few terse lines. Ogre also carefully deals with some of the side effects that the Gremlin library can launch.
At the lowest level, Ogre is probably equivalent to some crazy Turing machine. Some poor grad student has probably had to try and write the JVM as a Turing machine. Poor gal.
This series of guide is organized to be read mostly linearly. That means that you can probably read it from start to finish and understand what is going on. We start from the basics, with traversals, transformations, query executions, and filters. Then we transition into the more advanced topics of annotations and side effects. At the same time, this documentation is meant to serve as a reference for anyone using the library. These examples were are developed at the command line or inside emacs with a REPL, so they are meant to be run and experimented with.
Ogre let's you build up Gremlin queries from scratch. The main method
for doing this is
q/query. Here is a simple query on the
Tinkergraph. It takes in the vertex with id
1, finds the vertices
that the starting vertex points out to, and then returns the result in
(require '[ogre.tinkergraph :as g]) (require '[ogre.core :as q]) (g/use-new-tinker-graph!) (q/query (g/find-by-id 1) q/--> q/into-vec!) ;;[#<TinkerVertex v> #<TinkerVertex v> #<TinkerVertex v>]
Let's break this down:
q/queryis a simple macro combining of
(GremlinPipeline.). It takes in a single element or a Collection and creates a new pipeline around them.
g/find-by-idis a function that goes and asks the vertex for the element of id 1.
q/-->is a function which adds on an outwards traversal step to the pipe. This means that the Gremlin query will take all the vertices it is currently thinking about and then think about all the vertices that the vertices in it's mind pointed to.
q/into-vec!executes the query and returns the results inside of a vector. Before this call, the Gremlin query hasn't actually done anything yet. Only when a function that ends with a bang is passed in does anything actually happen beyond just a GremlinPipeline getting built up.
So far so good. But, I wonder, who is the dashing rogue behind
(q/query (g/find-by-id 1) q/--> q/into-vec! first ((q/prop :name))) ;;"vadas"
q/query isn't just about running Gremlin queries. Remember, it's
really just a glorified
->, and a bunch of provided helper
functions. That means we can stick a
first in there to get the first
vertex of the vector.
(q/prop :name) takes a property key and
returns a function which takes a vertex and returns the given
"vadas" is the charming face of
You should read about traversals next.
Let us know what was unclear or what has not been covered. Reader feedback is key to making the documentation better. If we know people are reading the documentation, we'll be much more inclined to make the documentation that much better. Please speak up!