xgt_connector Package

This Python package is for connecting the Rocketgraph xGT graph analytics engine to other applications. Rocketgraph xGT can significantly speedup Neo4j queries.

The default connector provided is for Neo4j or AuraDB. The package also provides an optional ODBC connector for connecting to databases or applications that support ODBC. Information about the ODBC connector can be found in the documentation.

Requirements

• Neo4j Python
• xGT Python
• Pyarrow
• Rocketgraph xGT

Installation

You can install this python package by executing this command:

python -m pip install xgt_connector

If you want to use the ODBC connector, you can install the optional dependencies like so:

python -m pip install 'xgt_connector[odbc]'

If you don't have Rocketgraph xGT, you can install and run the Developer version from Docker:

docker pull rocketgraph/xgt
docker run --publish=4367:4367 rocketgraph/xgt

Using the xgt_connector

From any Python environment, simply importing both xgt and xgt_connector is all that is needed to operate this connector.

A simple example below shows connecting to Neo4j and xGT, transferring the whole graph database to xGT, running a query in xGT, and printing the results:

import xgt
from xgt_connector import Neo4jConnector, Neo4jDriver

# Connect to xGT and Neo4j.
xgt_server = xgt.Connection()
xgt_server.set_default_namespace('neo4j')
neo4j_server = Neo4jDriver(auth=('neo4j', 'foo'))
conn = Neo4jConnector(xgt_server, neo4j_server)

# Transfer the whole graph.
conn.transfer_to_xgt()

# Run the query.
query = "match(a:foo) return a"
job = xgt_server.run_job(query)

# Print results.
print("Results: ")
for row in job.get_data():
    print(row)

API

The available properties are:

• neo4j_relationship_types
• neo4j_node_labels
• neo4j_property_keys
• neo4j_node_type_properties
• neo4j_rel_type_properties
• neo4j_nodes
• neo4j_edges

The available methods are:

• get_xgt_schemas
• create_xgt_schemas
• copy_data_to_xgt
• transfer_to_xgt
• transfer_to_neo4j

Examples

Some examples can be found here:

• Python examples
• Jupyter notebooks