The py2neo project provides bindings between Python and Neo4j via its RESTful web service interface. It attempts to be both Pythonic and consistent with the core Neo4j API.
In addition, the project provides support for the Graph Export Object File Format (Geoff) as well as a couple of command line tools for extra sugar. For further information on Geoff, visit http://py2neo.org/geoff.
- Website: http://py2neo.org/
- PyPI: http://pypi.python.org/pypi/py2neo
- GitHub: https://github.com/nigelsmall/py2neo
- Email: Nigel Small firstname.lastname@example.org
Py2neo has been built against the following software:
- Neo4j 1.4, 1.5
- Python 2.7
Other versions may work but are not guaranteed.
The easiest way to install py2neo is from the Python Package Index (PyPI). This generally requires superuser privileges, so on Debian or Ubuntu, simply execute the following (the -Z option ensures the package is unzipped):
$ sudo easy_install -Z py2neo
If you wish to make use of the command line tools, you may prefer to create symbolic links to your system path. To achieve this, execute the following (ensuring you replace the paths specified here with the appropriate equivalents for your system and versions):
$ sudo ln -s /usr/local/lib/python2.7/dist-packages/py2neo-0.99-py2.7.egg/py2neo/cypher.py /usr/bin/cypher $ sudo ln -s /usr/local/lib/python2.7/dist-packages/py2neo-0.99-py2.7.egg/py2neo/geoff.py /usr/bin/geoff
The following short programme illustrates a simple usage of the py2neo library:
#!/usr/bin/env python """ Simple first example showing connection and traversal """ # Import Neo4j modules from py2neo import neo4j # Attach to the graph db instance gdb = neo4j.GraphDatabaseService("http://localhost:7474/db/data") # Obtain a link to the reference node of the db ref_node = gdb.get_reference_node() # Obtain a traverser instance relative to reference node traverser = ref_node.traverse(order="depth_first", max_depth=2) # Output all the paths from this traversal for path in traverser.paths: print path
If symbolic links have been added, as detailed above, the
commands will be available for direct usage from a terminal or within a shell
cypher command allows Cypher queries to be executed against a
local or remote database, via its RESTful web interface, with the results
displayed on stdout:
usage: cypher [-h] [-u DATABASE_URI] [-d] [-g] [-j] [-t] query Execute Cypher queries against a Neo4j database server and output the results. positional arguments: query the Cypher query to execute optional arguments: -h, --help show this help message and exit -u DATABASE_URI the URI of the source Neo4j database server -d output all values in delimited format -g output nodes and relationships in GEOFF format -j output all values as a single JSON array -t output all results in a plain text table (default)
geoff command will allow a GEOFF data to be loaded into a
database from either a file or stdin:
usage: geoff [-h] [-u DATABASE_URI] [-f SOURCE_FILE] [name=uri [name=uri ...]] Import graph data from a GEOFF file into a Neo4j database. A source file may be specified with the -f option and a destination database with the -u option. The remainder of the arguments will be passed as hooks into the load routine. Each hook may be a node of relationship and can optionally be named. Unnamed hooks will be automatically named by their relative zero-based position. For example, "foo=/node/123" designates the node with ID 123, named as "foo", whereas "/relationship/456" designates the relationship with ID 456 and will be named "0" if it is in the first position, "1" for the second, and so on. EXAMPLE: geoff.py -f foo.geoff bar=/node/123 baz=/relationship/456 positional arguments: name=uri named relative entity URI (e.g. foo=/node/123) optional arguments: -h, --help show this help message and exit -u DATABASE_URI the URI of the destination Neo4j database server -f SOURCE_FILE the GEOFF file to load
These commands may of course be piped together, allowing data to be fed from one database into another. For example:
cypher -g "START n=node(23) match (n)-[r]-(x)--(y) return n, x.name, n, r, x, x.name, x.\`birth.date\`, y, y.name, r.\`marriage.date\`?" | geoff -u http://188.8.131.52:7474/db/data/
Note that if no database URI is specified, a default of
http://localhost:7474/db/data/ will be used.
Copyright 2011-2012 Nigel Small