Skip to main content
Join the official 2019 Python Developers SurveyStart the survey!

Python Library to manipulate directed graphs in redis

Project description


Simple python library to manipulate directed graphs in redis

Travis CI Number of PyPI downloads


pygraph_redis is released under the MIT Public License


pygraph_redis is a simple library to manipulate directed graphs inside a redis database.

In this library, a graph is a bunch of nodes, each node knows its predecessors and its successors. A node can store some attributs (strings or sets of strings).


pygraph_redis relies on redis and redis-py.

For atomicity of transaction, it requires lua scripting support (redis-py >= 2.7.0 and redis >= 2.6.0), but it provides a legacy mode, without atomicity for older redis and redis-py.

Write atomicity

With proper versions, pygraph_redis provides the atomicity of transaction when adding or removing a node.


to install:

$ python install


$ pip install pygraph_redis

How to use

First you need a redis database, it’s up to you to install it.

The library itself is quite simple:

Cheat Sheet

#              initialization
#       arg1      |    arg2    |     arg3
# redis connexion | graph_name |    logger
#    redis obj    |  unicode   |  logger obj

mygraph1 = Directed_graph(r_server, u'mygraph1', logger)

#optional args:
#   arg4    |    arg5
# separator | has_root
# unicode   |   bool

mygraph1 = Directed_graph(r_server,
    u'mygraph1', logger, u'mysep', True)
#                    create or add elements to a node
#    arg1   |     arg2     |     arg3     |             arg4
# node name |  successors  | predecessors |           attributs
#  unicode  | unicode list | unicode list |      dictionnary of unicode
#           |              |              | or set of unicode (key: unicode)

                       {u'a3': set([u'69']), u'a2': u'42'}
#             delete elements from a node
#    arg1   |     arg2     |     arg3     |      arg4
# node name |  successors  | predecessors | attributs names
#  unicode  | unicode list | unicode list | list of unicode

mygraph1.write_off_node(u'm1', [u's2'], [u'p1'], [u'attr3', u'attr2']
# delete a node
#     arg1
#  node name
#   unicode

# get attributs list
#     arg1
#  node name
#   unicode

# get an attribut
#     arg1     |     arg2
#  node name   | attribut name
#   unicode    |    unicode

mygraph1.get_attribut(u'm1', u'a2')
# get an attribut length
#     arg1     |     arg2
#  node name   | attribut name
#   unicode    |    unicode

mygraph1.get_attribut_len(u'm1', u'a2')
# get successors
#     arg1
#  node name
#   unicode

# get predecessors
#     arg1
#  node name
#   unicode



Create an instance of “Directed_graph”:

#importing directed_graph
from pygraph_redis.directed_graph import Directed_graph
import redis

#creating a basic logger
import logging
logging.basicConfig(format = u'%(message)s')
logger = logging.getLogger(u'redis')

#creating the redis connexion
r_server = redis.Redis("localhost")

#creating the graph object
mygraph1 = Directed_graph(r_server, u'mygraph1', logger)

#creating the graph object with a different separator
mygraph2 = Directed_graph(r_server, u'mygraph2', logger, separator = u'mysep')

#creating the graph object with a "root" (improper name, I know)
mygraph2 = Directed_graph(r_server, u'mygraph2', logger, has_root = True)
#"has_root = True" ensures that every node has a predecessor
#if enabled, a node has at least root as a predecessor,
#but if it has any other predecessor it doesn't have root as predecessor

Node manipulation

Node creation:

#add node 'm1' to 'mygraph1' with:
#successors: 's1' and 's2'
#predecessors: 'p1' and 'p2'
#   * 'attr1': set([u'51',u'69'])
#   * 'attr2': '42'

    [u's1', u's2'],
    [u'p1', u'p2'],
    {u'attr1': set([u'51', u'69']), u'attr2': u'42'}

About successors and predecessors, if node was already declared as a predecessor of one of its successors, it’s not necessary to add this successor in node successors set. Same with predecessors.



Gives the same result that:


Node edition:

#add new elements or edit existing elements of a node
#it's exactly the same function as before
    {u'attr3': set([u'16', u'32', u'64']), u'attr2': u'5150'}

#remove some elements of a node (successors, predecessors, attributs)
mygraph1.write_off_node(u"m1", [u"s1"], [u"p2"],[u'attr2'])

#completely delete a node

Node attributs manipulation

To manipulate the attributs of a node:

#create the node 'm2'
    [u's1', u's2'],
    [u'p1', u'p2'],
    {u'attr1': set([u'51', u'69']), u'attr2': u'42'}

#get the set of attribut names
set_of_attributs = mygraph1.get_attributs_list(u'm2')
print set_of_attributs

#get a specific attribut
attr2 = mygraph1.get_attribut(u'm2', u'attr2')
print attr2

#get a specific attribut length
# 1 if it's a string
# cardinal of set if it's a set
# 0 if attribut doesn't exists
attr2 = mygraph1.get_attribut_len(u'm2', u'attr2')
print attr2

Graph navigation

To navigate inside the graph, you have two functions:

#get the predecessors of 'm2'
predecessors = mygraph1.get_predecessors(u'm2')
print predecessors

#get the successors of 'm2'
successors = mygraph1.get_successors(u'm2')

if you have the has_root flag enable:

#get the "root" name
root = mygraph1.get_root_name()

print root

#get the successors of 'root'
successors = mygraph1.get_successors(root)
print successors

About the redis keys

Redis key format:

<graph name><sep><node_name><sep><variable_name>[<sep><other>]*

<graph name>: name of the graph
<sep>: the key fields separator
     (this string should not be in node_name or variable_name,
      otherwise, there is a redis key collision possibility)
<node_name>: name of the node
<variable_name>: name of the variable
[<sep><other>]: optional extension

To avoid key collision, you must carefully choose the key separator, it must not be included in any node name or node attribut name (possible redis key collision).

About the logs

This library provides a lot of logs, mainly debug, some info (ex: legacy modes), some warning (ex: possible key collision)

Bitdeli badge

Project details

Download files

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

Files for pygraph_redis, version 0.2.1
Filename, size File type Python version Upload date Hashes
Filename, size pygraph_redis-0.2.1.tar.gz (10.2 kB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page