gremlin-python (gremthon) allows you to use Python syntax when traversing property graphs...
Take a test drive locally (expects curl, java, and unzip commands to be available):
$ ./test_drive/setup.sh && ./test_drive/run.shThis will download the necessary components, start titan, and run the rexster console with python scripting support enabled.
If you prefer to use Docker, you can pull the pokitdok/gremlin-python-test-drive container and run it to quickly evaluate gremlin-python:
$ docker run -i -t pokitdok/gremlin-python-test-driveCompile, test and package the gremlin-python jar:
$ mvn clean packageTo utilize Python syntax from within a rexster console, you'll need to drop in a couple of jar files and edit some xml configuration.
Edit conf/rexster-cassandra-es.xml (or the configuration file you're using) in your titan distribution directory to include:
<script-engine>
<name>gremlin-python</name>
</script-engine>See the rexster-test-drive.xml file for an example configuration that's used when you take a test drive.
You'll also need to put the files gremlin-python-{version}.jar and jython-standalone-{version}.jar into your titan lib directory. gremlin-python has been tested with jython-standalone-2.7.0.jar. You can find a gremlin-python jar file for each release at https://github.com/pokitdok/gremlin-python/releases The jython standalone jar can be found at http://www.jython.org/downloads.html
After (re)starting titan + rexster, you should see python available in your rexster console. We also included a sample graph to explore.
$ ./bin/rexster-console.sh -l python
(l_(l
(_______( 0 0
( (-Y-) <woof>
l l-----l l
l l,, l l,,
opening session [127.0.0.1:8184]
?h for help
rexster[python]> g = rexster.getGraph("graph")
==>null
rexster[python]> g.V.has('vertex_type','doctor').count()
==>25
rexster[python]> g.V.has('vertex_type','consumer').count()
==>1000
rexster[python]> g.E.has('edge_type','viewed').count()
==>3470
rexster[python]> g.E.has('edge_type','scheduled_with').count()
==>1000Exploring the properties of this graph a bit more... (note: the people in this example are fake)
rexster[python]> list(g.V.has('vertex_type','doctor').map())[0]
==>last_name=Moen
==>first_name=Shana
==>full_name=Dr. Shana Moen
==>specialty=family doctor
==>age=60
==>vertex_type=doctor
rexster[python]> list(g.V.has('vertex_type','consumer').map())[0]
==>last_name=Rogahn
==>first_name=Ozella
==>full_name=Ozella Rogahn
==>age=32
==>vertex_type=consumer
rexster[python]> list(g.V.has('age',55).full_name)[:5]
==>Enrico Homenick
==>Vincenzo Ebert
==>Dr. Annetta McGlynn
==>Bertram Jaskolski
==>Mitchel QuitzonThis example was set up to mimick a network of consumers viewing and scheduling appointments with doctors. For example:
rexster[python]> g.V.has('full_name','Myriam Daugherty').out_e('viewed').in_v().full_name
==>Dr. Alfred Dibbert
==>Dr. Abraham Casper
==>Dr. Cristobal Leffler
==>Dr. Avis Crona
rexster[python]> g.V.has('full_name','Myriam Daugherty').out_e('scheduled_with').in_v().full_name
==>Dr. Nellie HeidenreichNext, let us take a look at how to rank the providers according to how many consumers have viewed them.
rexster[python]> ranked_by_viewed = {}
rexster[python]> g.E.has('edge_type','viewed').in_v().group_count(ranked_by_viewed, lambda it: it.full_name, lambda it: it.b+1.0)
rexster[python]> sorted_viewed_results = sorted(ranked_by_viewed.items(), key=lambda x:x[1], reverse=True)
==>null
rexster[python]>sorted_viewed_results[:10]
==>(u'Dr. Cristobal Leffler', 369.0)
==>(u'Dr. Avis Crona', 366.0)
==>(u'Dr. Shana Moen', 350.0)
==>(u'Dr. Nellie Heidenreich', 337.0)
==>(u'Dr. Donny Schaefer', 266.0)
==>(u'Dr. Annetta McGlynn', 221.0)
==>(u'Dr. Israel Kiehn', 201.0)
==>(u'Dr. Roxanne Quigley', 196.0)
==>(u'Dr. Elton Zboncak', 184.0)
==>(u'Dr. Alfred Dibbert', 153.0)Next, let us take a look at how to rank the providers according to how many consumers have scheduled with them.
rexster[python]> ranked_by_scheduled = {}
rexster[python]> g.E.has('edge_type','scheduled_with').in_v().group_count(ranked_by_scheduled, lambda it: it.full_name, lambda it: it.b+1.0)
rexster[python]> sorted_scheduled_results = sorted(ranked_by_scheduled.items(), key=lambda x:x[1], reverse=True)
==>null
rexster[python]> sorted_scheduled_results[:10]
==>(u'Dr. Nellie Heidenreich', 111.0)
==>(u'Dr. Donny Schaefer', 104.0)
==>(u'Dr. Anna Collier', 100.0)
==>(u'Dr. Avis Crona', 83.0)
==>(u'Dr. Elton Zboncak', 78.0)
==>(u'Dr. Roxanne Quigley', 68.0)
==>(u'Dr. Constance Kihn', 58.0)
==>(u'Dr. Emmalee Ondricka', 57.0)
==>(u'Dr. Annetta McGlynn', 54.0)
==>(u'Dr. Kip Stoltenberg', 45.0)If you have problems connecting to a remote titan graph (that's using elasticsearch) when you're working within an interactive jython session, try placing the names.txt file from elasticsearch somewhere on the path or in your current working directory. It seems that some class loader differences exist between an interactive jython session and working within rexster. names.txt can be found properly within rexster but not when working with jython. You can grab a copy of names.txt here: https://github.com/elasticsearch/elasticsearch/blob/master/src/main/resources/config/names.txt or from within the elasticsearch jar file.
This library aims to support and is tested against these JVM versions:
- openjdk7
- oraclejdk7
- oraclejdk8
Copyright (c) 2015 PokitDok, Inc. The MIT License (MIT) (See LICENSE for details.)