Edgeable is an easy to use, in memory, peristable graph database. It is perfect for prototyping, exploring, and quick implementation in Python applications. The library is written completely in Python, supporting 3.6+.

In Edgeable, graphs are made of nodes with a string or numeric identifier and can have any number of properties. Nodes are connected by directed or non-directed edges that can also have any number of properties. Additional convenience capabilities, like route detection through the graph, are built-in. Edgeable is thread safe for multi-threaded applications and easily handles databases with tens to hundreds of thousands of nodes.

Installation

The Edgeable database library can easily be installed with pip using pip install edgeable.

Examples

Create a graph instances and add nodes and edges

from edgeable import GraphDatabase

graph = new GraphDatabase()
A = graph.put_node('A', {'type': 'building'})
B = graph.put_node('B', {'type': 'classroom'})
A.attach(B, {'relationship': 'contains'})

Loading from a csv file of edges

from edgeable import GraphDatabase
import csv

graph = GraphDatabase()
with open('edges.csv') as csvfile:
    reader = csv.reader(csvfile)
    for row in reader:
        node_a = graph.put_node(row[0].strip())
        node_b = graph.put_node(row[1].strip())
        node_1.attach(node_b)

Write a csv file of edges

with open('edges.csv', 'w') as csvfile:
    writer = csv.writer(csvfile)
    for node in graph.get_nodes():
        for edge in node.get_edges():
            writer.writerow([
                edge.get_source().get_id(), 
                edge.get_destination().get_id()
            ])

Classes

Graph Class

The GraphDatabase class represents the entire graph. Instances of GraphDatabase provide the ability to retrieve and creates nodes, as well as save or load from the file system.

Graph Constructor

The GraphDatabase(filename="graph.db") constructor is used to create a new graph database instance. The file name used to persist the database can optionally be ovewritten. If a persisted database file already exists, it will be loaded.

The with syntax can be used to initialize the graph and ensure it is saved once the block is exited.

from edgeable import GraphDatabase

with new GraphDatabase(filename="mygraph.db") as graph:
    A = graph.put_node('A', {'type': 'building'})
    B = graph.put_node('B', {'type': 'classroom'})
    A.attach(B, {'relationship': 'contains'})

Graph Persistance

• reload() - Reload the graph from the file system. The persisted file is loaded automatically when the GraphDatabase instance is initialized, calling load() is only neccessary to reload it, overwriting any changes.
• save() - Save the graph to the file system. When using the with syntax, this is called automatically. Save operations are done in a way to be thread-safe and avoid corruption if the application terminates during a write.

Graph Nodes and Edges

• put_node(id, properties={}) - Creates or retrieves the instance of GraphNode with the provided identifier. The optionally provided properties are set or updated on the node.
• has_node(id) - Taking a node identifier, return an instance of type GraphNode.
• get_node(id) - Taking a node identifier, return an instance of type GraphNode if it exists in the database. Returns None otherwise.
• get_nodes(filter_fn=lambda node: True) - Retrieve a list of GraphNode instances from the database. If the optional filter function is not provided, all nodes are returned, otherwise the filter function is used to return only matching nodes.
• edges(edge_filter_fn=lambda edge: True, node_filter_fn=lambda node: True) - Retrieve a list of GraphEdge instances from the database. If the optional filter functions are not provided, all edges are returned, otherwise the function is used to return only matching edges.
• get_node_count() - Return the number of nodes in the database.
• get_edge_count() - Return the number of edges in the database.

Graph Properties

• set_property(key, value) - Set a property on the node with the provided key and value.
• set_properties(properties) - Provide a dict to set multiple properites on the database.
• get_property(key) - Retrieve the node's property value for the provided key.
• get_properties() - Retrieve a dict containing all properties set on the node.
• has_property(key) - Returns a boolean indicating whether the property key is defined.
• delete_property(key) - Removes a property.

Graph Callbacks

Event callbacks can be provided that will have the opportunity to process objects in the graph. Multiple callbacks can be defined for an event type. Each callbacks has an identifier, this can be specified when created or one will automatically be generated and returned. The id can be later used to overwrite or remove (set to None) the callback.

• on_create_node(fn, id=None) - Sets or changes a callback that is called when a node is added to the graph.
• on_create_edge(fn, id=None) - Sets or changes a callback that is called when a edge is added to the graph.
• on_delete_node(fn, id=None) - Sets or changes a callback that is called when a node is deleted from the graph.
• on_delete_edge(fn, id=None) - Sets or changes a callback that is called when a edge is deleted from the graph.

Callback functions that return a False value will cancel the event. Callback functions are not persisted with the database.

Node Class

The GraphNode class represent nodes and associated properties within the graph. Nodes can be connected through directed or non-directed edges.

• delete() - Delete the node from the database, removing any associated edges.

Node Edges

• attach(destination, properties={}, directed=False) - Attach the current node to the destination node with a new edge. The optionally provided properties will be set on the edge, whether it exists or is created new. The releationship will be nondirectional unless directed is set to True. The method returns a boolean indicating if a new edge was created between the nodes.
• detach(destination=None, directed=False) - Detach the current node from the destination node, removing any connecting edge. If no destination is provided, all attached nodes are detached. If directed is set to True, then the edge is only removed in the current direction, any edge in the other directions is untouched. The method returns a boolean indicating if any edges were removed.
• get_edges(filter_fn=lambda edge: True) - Retrieve a list of GraphEdge instances from the database. If the optional filter function is not provided, all edges for the node are returned, otherwise the filter function is used to return only matching edges.
• get_edge(destination) - Retrieve the edge that leads from the current node to the provided destination node.
• has_edge(destination) - Returns a boolean indicating whether there is an edge from the current node to the destination node.

Node Properties

• set_property(key, value) - Set a property on the node with the provided key and value.
• set_properties(properties) - Provide a dict to set multiple properites on the node.
• get_property(key) - Retrieve the node's property value for the provided key.
• get_properties() - Retrieve a dict containing all properties set on the node.
• has_property(key) - Returns a boolean indicating whether the property key is defined.
• delete_property(key) - Removes a property.

Node Routes

• find_routes_to(end, effort=5)
• find_route_to(end, skip=[])

Node Neighbors

• find_neighbors(distance=1, distance_fn=lambda edge: 1) - Find all nodes within a specified distance. By default, each edge is considered a distance of one. A custom function can be provided to provide a custom distance calculation.

Edge Class

The GraphEdge class representes the connections between GraphNode instances in the database. All methods assume edges are nondirected, so interactions with edges equally manipulates an edge in each direction between nodes. This behavior can be overridden for specific calls, resulting in the creation, modification, or deletion of directional edges.

• get_destination() - Retrieve the GraphNode instance that is the destination of this edge.
• get_source() - Retrieve the GraphNode instance that is the source of this edge.
• delete(directed=False) - Delete the edge from the database. If directed is True than only the edge in this direction is deleted.

Edge Properties

• set_property(key, value, directed=False) - Set a property on the node with the provided key and value. If directed is True than the property is set only on the edge in this direction.
• set_properties(properties, directed=False) - Provide a dict to set multiple properites on the edge.
• get_property(key) - Retrieve the node's property value for the provided key.
• get_properties() - Retrieve a dict containing all properties set on the edge.
• has_property(key) - Returns a boolean indicating whether the property key is defined.
• delete_property(key) - Removes a property.

Resources

Build process: https://app.travis-ci.com/github/LeeAdcock/edgeable

PyPi: https://pypi.org/project/edgeable/