Crates.io | graph_match |
lib.rs | graph_match |
version | 0.5.0 |
source | src |
created_at | 2016-05-17 21:58:26.934485 |
updated_at | 2016-05-25 09:17:05.184517 |
description | A library for matching patterns in directed graphs. |
homepage | |
repository | https://github.com/charlieegan3/graph_match |
max_upload_size | |
id | 5095 |
size | 31,253 |
A library for matching patterns in directed graphs. The library implements 'queries' for graphs as graphs that represent in part the nodes in the graph to be matched.
The API for this functionality is still changing and the project is in early development. Integration tests can be found here - these give a reasonable overview of the functionality.
There are two broad classes of query implemented, these can be described loosely as:
As part of my honors project I needed to answer the above questions for a graph data structure. In my case, the graph was a dependency graph. I opted to use Neo4j for the purpose, the Cypher query language is really nice to write queries in. This introduces temporary persistence for graphs during analysis, and that slows things down quite a bit. Running a Neo4j container also uses more RAM than I'd like. So, in my efforts to get standpoint up and running on a cheap cloud instance, this project, that implements the features of Cypher that I use, was created.
This project has been written for a tool that I'm working on where it's possible to make a number of assumptions on the type of graphs used. I'm working with dependency parse graphs and have made the following assumptions: