Gremlin inspired Rexster Graph Server client for NodeJS and the browser.

If you wish to try the latest changes, please checkout the develop branch.

Comment, issues and pull requests welcome!

gRex was originally started by Frank Panetta (@entrendipity) and is now maintained by Jean-Baptiste Musso (@jbmusso).

gRex tries to implement Gremlin/Groovy syntax as closely as possible within a JavaScript environment. However, there are some differences:

• All method calls require brackets (), even if there are no arguments.

• Closures do not translate to javascript. Closures need to passed in as a string argument to gRex methods.

  g.v(1).out().gather("{it.size()}");
  
  g.v(1).out().ifThenElse("{it.name=='josh'}{it.age}{it.name}");

• Comparators and Float's are not native javascript Types so they need to be passed in as a string to gRex methods. Floats need to be suffixed with a 'f'.

  g.v(1).outE().has("weight", "T.gte", "0.5f").property("weight")

• Certain methods cannot be implemented, such as aggregate, store, table, tree and fill. These methods require a local object to populate with data, which cannot be done in this environment. This may however change in the future providing some tweaks in the library.

Note: all asynchronous methods in gRex accept a dual callback/promise API (using kriskowal/q). Examples will however be given with the default Node.js callback style, with an error as first parameter.

  var options = {
    'database': 'myGraphDB',
    'host': 'my.host.com',
    'port': 8000
  };

  gRex.connect(options, function(err, g) {
    if (err) console.error(err);

    // Once connected the return value is a reference to the graph
    tx = g.begin();
    t1 = tx.addVertex({ name: 'Test1a' });
    t2 = tx.addVertex({ name: 'Test2a' });
    tx.addEdge(t1, t2, 'linked', { name: "ALabel" })

    tx.commit(function(err, result) {
        if (err) console.error(err);
        console.log("Added new vertices successfully. -> ", result);
    });
  });

A good resource to understand the Gremlin API are GremlinDocs and SQL2Gremlin.

gRex uses the Q module to return a Promise when making calls to Rexster. GET HTTP requests are invoked with get() and the callback is captured by then(success, error);. However, if you prefer Node style callbacks, simply pass the callback to get().

Warning: this API should not be considered stable and will most likely evolve in future versions as we head toward v1.0.0.

  g.V('name', 'marko').out().get()
  .then(function(result) {
    console.log(result);
  })
  .fail(function(err) {
    console.error(err);
  });

  g.V('name', 'marko').out().get(function(err, result) {
    if(err) console.error(err);
    console.log(result);
  });

Create, Update and Delete actions are batched to reduce the number of calls to the server. In order to send these types of requests, a Transaction must be created by calling var tx = g.begin();. Make changes to your data against this object. Once all changes are done, invoke tx.commit() to commit your changes.

Again, both the Promise style and Node style callbacks are available with commit().

Creating, updating or deleting Vetices or Edges. Use commit() to commit changes.

  var tx = g.begin();
  tx.addVertex(100, {k1:'v1', 'k2':'v2', k3:'v3'});
  tx.addVertex(200, {k1:'v1', 'k2':'v2', k3:'v3'});
  tx.addEdge(300,100,200,'pal',{weight:'0.75f'})
  tx.updateVertex(100, {k2: 'v4'});
  tx.removeVertex(100, ['k2', 'k3']);
  tx.removeVertex(200);

  tx.commit(function(err, result){
    if (err) console.error(err);
    console.log(result);
  });

Below are examples of Gremlin queries and their respective equivalent gRex syntax.

For simplicity the callbacks are not included in the examples below.

gremlin>  g.V('name', 'marko').out

gRex>     g.V('name', 'marko').out();

gRex>     g.V({name: 'marko'}).out();

gremlin>  g.v(1, 4).out('knows', 'created').in

gRex>     g.v(1, 4).out('knows', 'created').in();

gRex>     g.v([1, 4]).out(['knows', 'created']).in();

gremlin>  g.V[0].name

gRex>     g.V().index(0).property('name');

gremlin>  g.V[0..<2].name

gRex>     g.V().range('0..<2').property('name');

gremlin>  g.E.has('weight', T.gt, 0.5f).outV.transform{[it.id,it.age]}

gRex>     g.E().has('weight', 'T.gt', '0.5f').outV().transform('{[it.id,it.age]}');

gremlin>  g.V.and(_().both("knows"), _().both("created"))

gRex>     g.V().and(g._().both("knows"), g._().both("created"))

gremlin>  g.v(1).outE.or(_().has('id', T.eq, "9"), _().has('weight', T.lt, 0.6f))

gRex>     g.v(1).outE().or(g._().has('id', 'T.eq', 9), g._().has('weight', 'T.lt', '0.6f'));

gremlin>    g.V.out.groupBy{it.name}{it.in}{it.unique().findAll{i -> i.age > 30}.name}.cap

gRex>       g.V().out().groupBy('{it.name}{it.in}{it.unique().findAll{i -> i.age > 30}.name}').cap()

gremlin>  g.V.retain([g.v(1), g.v(2), g.v(3)])

gRex>     g.V().retain([g.v(1), g.v(2), g.v(3)])

gremlin>  g.createIndex("my-index", Vertex.class)

gRex>     g.createIndex("my-index", "Vertex.class")

gremlin>  g.idx("my-index").put("name", "marko", g.v(1))

gRex>     g.idx("my-index").put("name", "marko", g.v(1))

gremlin>  g.idx("my-index")[[name:"marko"]]

gRex>     g.idx("my-index", {name:"marko"});

gremlin>  g.dropIndex("my-index", Vertex.class)

gRex>     g.dropIndex("my-index", "Vertex.class")

gRex:

var tx = g.begin();
tx.addVertex(100, { k1:'v1', 'k2':'v2', k3:'v3' });
tx.addVertex(200, { k1:'v1', 'k2':'v2', k3:'v3' });
tx.addEdge(300, 100, 200,'pal', { weight: '0.75f' })
tx.updateVertex(100, { k2: 'v4' });
tx.removeVertex(100, ['k2', 'k3']);
tx.removeVertex(200);
tx.commit();

var tx = g.begin();
var v1, v2;

v1 = tx.addVertex({ name: 'frank' });
v2 = tx.addVertex({ name: 'Luca' });

v1.addProperty('status', 'new');
v1.setProperty('name', 'Frank');

tx.addEdge(v1, v2, 'knows', { since:"2003/06/01" })

v1 = tx.addVertex({name:'Stephen'});
v2 = tx.addVertex({name:'James'});
tx.addEdge(v2, v1, 'knows', { since:"2000/01/01" })

tx.commit()
.then(function(result) {
  console.log("New vertices -> ", result);
})
.fail(function(err) {
  console.error(err)
});

//This will return a JSON object with an array called newVertices. For example:

{ success: true,
  newVertices:
   [ { name: 'Frank', _id: '#8:334', _type: 'vertex' },
     { name: 'Luca', _id: '#8:336', _type: 'vertex' },
     { name: 'Stephen', _id: '#8:335', _type: 'vertex' },
     { name: 'James', _id: '#8:337', _type: 'vertex' } ]
}

var y = "bob";
var tx = g.begin();
var vertex = tx.addVertex({ name:y });

tx.commit()
.then(function(result) {
  console.log("Added a vertex successfully for", y);
  return g.createIndex('actor', 'Vertex.class');
})
.then(function(result) {
  return g.idx('actor').put('name', y, g.v(vertex._id);
})
.then(function(result) {
  console.log("Index added successfully for", y);
})
.fail(function(err) {
  console.error(err);
});

options object specify the location and name of the database:

• host: location of Rexster server (default: localhost)
• port: Rexster server port (default: 8182)
• graph: graph database name (default: tinkergraph)
• idRegex: this can remain as false (default value), if IDs are numbers. If IDs are not numbers (i.e. alpha-numeric or string), but still pass parseFloat() test, then idRegex must be set. This property will enable gRex to distinguish between an ID and a float expression.

  var settings = {
    host: 'myDomain',
    graph: 'myOrientdb',
    idRegex: /^[0-9]+:[0-9]+$/
  });

Rexster Graph Server supports the following Property Data Types:

• Strings
• Boolean
• Integer
• Long
• Float
• Double
• List (Array)
• Map

gRex automatically preserves data types. It uses type values obtained from the server, when data is retrieved, to ascertain data types. If a property's data type is unknown, gRex will not try to infer the data type and will simply allow the value to be passed as a string, which is the default behaviour. However, it is possible to provide a type definition to a Transaction, which will then be used to pass type information to the server during a POST.

When Rexster returns data, it will include Type information, gRex will create a Type definition based on this information to be used in subsequent POST requests. Type definitions can only be generated for Objects that have been retrieved from the server. So, if you are updating or creating a 'Person' object the type definition will only be available if a 'Person' object was previously requested and retrieved from teh server.

Also, if totally new properties need to have a Type definition, so that gRex can understand how to send the information to the server.

A Type definition is an Object and is used globally. For example, if 'age' has been defined as type integer in a 'Person' object, then whenever gRex encounters an 'age' property, regardless of the Object, it will be treated as an integer. Although, if the 'age' property is embedded in another object, then it will need to be explicitly defined.

You are only required to provide a Type definition for properties that are being added.

In order to use a Type Definition, you pass in an Object to the Transaction begin function.

A Type definition is an Object Literal. The key is the property name for the Object you are providing a Type definition for and the value is the Type that is being assigned to that property. For example, to define a property as boolean for a key called 'active' you would do the following:

{ active: 'boolean' } OR { active: 'b' }

This is the similar for all the simple Types.

• Strings = 'string' or 's'
• Boolean = 'boolean' or 'b'
• Integer = 'integer' or 'i'
• Long = 'long' or 'l'
• Float = 'float' or 'f'
• Double = 'double' or 'd'

Complex types, such as list and map are a little different.

To define a Type for are List (Array), you simply provide an Array as the value and provide the type name for each item in the array. You will need to know which index a particular Type will be located. Any items added to the array after item[3] will be added as the last Type defined in the array, in this instance the items will be added as integers.

{ items: ['string', 'string', 'boolean', 'integer'] }

Map Type's are simply object literals. To define a map type you pass in objects much the same as defining a simple type above.

{ address: { number: 'integer', street: 'string', city: 'string'} }

Both List and Map Types can have embedded list and map types.

• list with embedded map [NB. There is currently a bug for maps embedded in lists] { items: ['string', {age: 'integer'}, 'boolean', 'integer'] }

• map with embedded list { address: { number: 'integer', street: 'string', city: 'string', occupantNames:['string']} }

To use a Type definition, just pass it to the begin function of a transaction.

var typeDef = { active: 'boolean',
                items: ['string', 'string', 'boolean', 'integer'],
                address: { number: 'integer', street: 'string', city: 'string', occupantNames:['string']}
            };
var tx = new g.begin(typeDef);

If there is already a Type definition for a property, the passed in type definition is merged with the existing type definition and takes precedence.

gRex requires batch kibble (Rexster extension) to be installed first.

Move the batch-kibble-XXX.jar, located in the modules lib folder, to the /ext folder under the rexster server directory. Add an allow tag to the database extensions configuration in the rexster.xml file.

  <extensions>
    <allows>
      <allow>tp:gremlin</allow>
      <allow>tp:batch</allow>
    </allows>
  </extensions>

Batch kibble is currently required for handling transactions over http.

gRex can be loaded as:

• a <script> tag in the browser. Files are located in the client directory.

   <script type="text/javascript" src="grex.min.js"></script>

  this exposes gRex as a global variable.

• a Node.js and CommonJS module available from NPM as the grex package

  $ npm install grex
  

  then in Node

  var gRex = require('grex');
  

• a RequireJS module

  require(['gRex'], function (gRex) {
      // Do something with gRex
  });
  

You will notice that in the examples tokens are passed as string (i.e. 'T.gt'). However, gRex also exposes some objects for convenience to make it feel more natural. To access the objects reference them like so:

var T = gRex.T;
var Contains = gRex.Contains;
var Vertex = gRex.Vertex;
var Edge = gRex.Edge;

You can now use these objects in place of the string representation in your queries.

Frank Panetta - Follow @entrendipity

Jean-Baptiste Musso - Follow @jbmusso / Gulthor on GitHub

Andreas Richter - richtera

##License ###The MIT License (MIT)

Copyright (c) 2013 entrendipity pty ltd

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.