next up previous contents index
Next: Boost Up: Eco Lab Documentation Previous: Basic Gaussrand is correlated   Contents   Index

Graph library

The Eco Lab graph library is a library providing a simple and lightweight structure for representing graphs (aka networks). A graph consists of a set of nodes labelled $N=\{0\ldots n-1\}$, and a set of edges $E\subset N\times N$, which have an optional weight factor attached.

An edge is represented by

struct Edge: public std::pair<unsigned,unsigned>
{
    unsigned source();
    unsigned target();
    float weight;
};

The abstract Graph interface has the following definition:

struct Graph
{
  struct const_iterator
  {
    Edge operator*() const;
    const Edge* operator->() const;
    const_iterator& operator++();
    bool operator==(const const_iterator& x) const;
    bool operator!=(const const_iterator& x) const;
  };
 
  virtual const_iterator begin() const;
  virtual const_iterator end() const;

  virtual unsigned nodes() const;
  virtual unsigned links() const;
  virtual void push_back(const Edge& e);
  virtual bool contains(const Edge& e) const;
  virtual bool directed() const;
  virtual void clear(unsigned nodes=0);
  const Graph& operator=(const Graph& x);
  void input(const std::string& format, const std::string& filename);
  void output(const std::string& format, const std::string& filename) const;
  template <class BG> Graph_back_insert_iterator<Graph,BG>
    back_inserter(const BG& bg);
}

The begin/end methods allow one to iterate over the edges. Only a const_iterator is supplied, as it is an error to change the value of an edge. One can only reset a graph to the empty graph via clear, and construct the graph incrementally using the push_back() method.

The contains method allows one to test whether a given edge is in the graph, and directed indicates whether the underlying graph structure has directed edges or not. A bidirectional graph otherwise appears as a directed graph where each edge appears twice, once for each direction.

The input/output methods allow for the graph to be read/written from/to a file, in a variety of formats, given by the format parameter. Currently, the following formats are supported:
name description
pajek Pajek's .net format
lgl LGL's .lgl format
dot Graphviz format
gengraph Gengraph
Also graphs can be streamed to/from standard I/O streams, and will appear in dot format. In particular, this means that TCL scripts can access Graph objects as strings containing the Graphviz representation. See some of the examples in the models/netcomplexity_scripts directory.

back_inserter creates an output iterator suitable for use with Boost Graph algorithms. As a simple example, to construct an Eco Lab graph from a Boost Graph, do

  std::pair<BG::edge_iterator,BG::edge_iterator> r=edges(bg);
  ConcreteGraph<DiGraph> g1;
  copy(r.first, r.second, g1.back_inserter(bg));

This interface can be used in both a dynamic polymorphism fashion (ie Graph is an abstract base class) and in a static polymorphism fashion.

The graph.h header file provides two concrete graph types - DiGraph and BiDirectionalGraph, which differ just in whether each edge is directed or not.



Subsections
next up previous contents index
Next: Boost Up: Eco Lab Documentation Previous: Basic Gaussrand is correlated   Contents   Index
Russell Standish 2016-09-02