Skip to main content

Run your graphs in code

Project description

Graph Runner (ALPHA)

Overview

graph-runner provides a way to add actual Python code to Graph DB nodes, which can be used to create real-time dynamically property values. Rather than only storing fixed values, you have the option to create rich property values that can be calculated on-the-fly with other vertex property values.

IMPORTANT: This codebase currently uses the exec() method to execute Python code. This was used for sake of time since this was built for a specific trusted project. Ideally, a more secure method should be used, but just be aware that arbitrary code can be executed.

Quick Start

To install graph-runner:

pip install graph-runner

While there's no full documentation yet, here's a code sample that walks you through the majority of the current features. Make sure that you export a CONN_STR environment variable that matches your connections string.

The only code this does not show is the https://github.com/troylar/graph-runner/blob/master/entities/init.py. Each custom entity maps to a node type. You simply have to define the exec_properties in the class definition, which define which properties that contain executable code.

from gremlin_python import statics
from gremlin_python.structure.graph import Graph
from gremlin_python.process.graph_traversal import __
from gremlin_python.process.strategies import *
from gremlin_python.driver.driver_remote_connection import DriverRemoteConnection
from graph_entity import GraphEntity
from entities import EventEntity, PersonEntity
import os

graph = Graph()

g = graph.traversal().withRemote(DriverRemoteConnection(os.environ.get('CONN_STR'),'g'))

# We're going to add an entity to the graph
first_event = EventEntity(Traversal=g)
print('Adding node')

# add_node() returns a traversal, so we have to next()
first_event.add_node().next()

# Now we're going to add a dynamic property called 'ignore_if' with a simple boolean piece of code
# NOTE: exec_val is always required as the return value
first_event.ignore_if = 'exec_val=1==1'

# print the id
print('Node id = {}'.format(first_event.id))

# print the node type, which maps to the EventEntity class and is saved as a property in the graph
print('Node type: {}'.format(first_event.full_self()))

# this is where it gets cool . . . the ignore_if property method was dynamically added to the python object
# and when you access the property, it executes the code and returns the value
print('ignore_if property: {}'.format(first_event.ignore_if()))

# now let's create a GraphEntity object and from scratch, load the node we just created
ge = GraphEntity(Traversal=g)
print('Getting node')
node = ge.from_node(first_event.id)
print('Node id = {}'.format(node.id))
print('Node type: {}'.format(node.full_self()))
print('ignore_if property: {}'.format(node.ignore_if()))

# let's create a person entity with an 'is_present' property
pe = PersonEntity(Traversal=g)
print('Adding person')
pe.add_node().next()
pe.is_present = 'True'

# let's create a second event
second_event = EventEntity(Traversal=g)
print('Adding another event')
second_event.add_node().next()

# let's add some more complex logic to a property
# this property is saying that if the person is present, move to the next event
next_node_code = """
from graph_entity import GraphEntity
ge = GraphEntity(Traversal=g)

if ge.from_node('{}').is_present()=='True':
    exec_val = g.V('{}')
else:
    exec_val = "N/A"
""".format(pe.id, second_event.id)

# because is_present is true, we get our next node
first_event.next_node = next_node_code
print(first_event.next_node())

# change is_present to false, and the next_node is N/A
pe.is_present = 'False'
print(first_event.next_node())

# we can also use jinja2 templating, which allows us to pass in dynamic values at run-time
# in this case, it's the same code, but we we're alloing the ids to be passed in at run-time
next_node_template = """
from graph_entity import GraphEntity
ge = GraphEntity(Traversal=g)

if ge.from_node('{{ person_node_id }}').is_present()=='True':
    exec_val = g.V('{{ next_node_id }}')
else:
    exec_val = "N/A"
"""

 # run the same code, but pass ids at run-time
first_event.next_node = next_node_template
pe.is_present = 'True'
print(first_event.next_node(Data={'person_node_id': pe.id, 'next_node_id': second_event.id}))

pe.is_present = 'False'
print(first_event.next_node(Data={'person_node_id': pe.id, 'next_node_id': second_event.id}))

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

graph-runner-0.0.15.tar.gz (9.0 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page