dcggraph 0.0.3

Directed Cyclic Graph Utilities for Python

Directed Cyclic Graph Utilities for Python

This package currently used for doing predicated searches from a given node in a directed graph that is allowed to be cyclic. Some points to clear up confusion:

• directed: means an edge between 2 nodes can only be travelled in one direction
• predicated: means that even if a node is reachable, it still has to pass a prediate test (i.e. predicate(node) == True) in order to be returned.
  • this means there are really 2 search filters in the graph:
    1. the node has to be reachable without any filtering
    2. the node has to pass the predicate test
• cyclic: means that if you start from a node, and follow a bunch of paths, you can end up at the same node. This makes things a lot harder than if you were working with DAGs.

installing

pip install dcggraph

API

The main feature is predicated graph compression and search. In addition to the snippet below, look at the tests for example usage

import re
from dcggraph import DCGnode, DCGedge, DCGgraph
from dcggraph.search import PredicateResult, DCGsearch

g = DCGgraph()
n0 = g.create_node("n0")
n1 = g.create_node("n1")
n2 = g.create_node("n2")
n3 = g.create_node("n3")
n4 = g.create_node("n4")
g.create_edge(n0, n1)
g.create_edge(n0, n2)
g.create_edge(n1, n3)
g.create_edge(n1, n4)

def even(node, visited, passed, failed):
   if f re.match("[a-z][02468]", node.name):
       return PredicateResult.PASS
   else:
       return PredicateResult.FAIL

s1 = DCGsearch(g, even)

# reuse the search object to cache FAIL/PASS results between searches
assertEqual(len(s1.search(n0)), 3)
assertEqual(len(s1.search(n1)), 1)
assertEqual(len(s1.search(n2)), 1)
assertEqual(len(s1.search(n3)), 0)
assertEqual(len(s1.search(n4)), 1)