<HTML>
<!--
     Copyright (c) Jeremy Siek, Lie-Quan Lee, and Andrew Lumsdaine 2000
    
     Distributed under the Boost Software License, Version 1.0.
     (See accompanying file LICENSE_1_0.txt or copy at
     http://www.boost.org/LICENSE_1_0.txt)
  -->
<Head>
<Title>DFS Visitor</Title>
<BODY BGCOLOR="#ffffff" LINK="#0000ee" TEXT="#000000" VLINK="#551a8b" 
        ALINK="#ff0000"> 
<IMG SRC="../../../boost.png" 
     ALT="C++ Boost" width="277" height="86"> 

<BR Clear>

<H1><img src="figs/python.gif" alt="(Python)"/>DFS Visitor Concept</H1>

This concept defines the visitor interface for <a
href="./depth_first_search.html"><tt>depth_first_search()</tt></a>.
Users can define a class with the DFS Visitor interface and pass an
object of the class to <tt>depth_first_search()</tt>, thereby
augmenting the actions taken during the graph search.

<h3>Refinement of</h3>

<a href="../../utility/CopyConstructible.html">Copy Constructible</a>
(copying a visitor should be a lightweight operation).

<h3>Notation</h3>

<Table>
<TR>
<TD><tt>V</tt></TD>
<TD>A type that is a model of DFS Visitor.</TD>
</TR>

<TR>
<TD><tt>vis</tt></TD>
<TD>An object of type <tt>V</tt>.</TD>
</TR>

<TR>
<TD><tt>G</tt></TD>
<TD>A type that is a model of Graph.</TD>
</TR>

<TR>
<TD><tt>g</tt></TD>
<TD>An object of type <tt>G</tt>.</TD>
</TR>

<TR>
<TD><tt>e</tt></TD>
<TD>An object of type <tt>boost::graph_traits&lt;G&gt;::edge_descriptor</tt>.</TD>
</TR>

<TR>
<TD><tt>s,u</tt></TD>
<TD>An object of type <tt>boost::graph_traits&lt;G&gt;::vertex_descriptor</tt>.</TD>
</TR>

</table>

<h3>Associated Types</h3>

none
<p>

<h3>Valid Expressions</h3>

<table border>
<tr>
<th>Name</th><th>Expression</th><th>Return Type</th><th>Description</th>
</tr>

<tr>
<td>Initialize Vertex</td>
<td><tt>vis.initialize_vertex(s, g)</tt></td>
<td><tt>void</tt></td>
<td>
This is invoked on every vertex of the graph before the start of the
graph search.
</td>
</tr>

<tr>
<td>Start Vertex</td>
<td><tt>vis.start_vertex(s, g)</tt></td>
<td><tt>void</tt></td>
<td>
This is invoked on the source vertex once before the start of the
search.
</td>
</tr>

<tr>
<td>Discover Vertex</td>
<td><tt>vis.discover_vertex(u, g)</tt></td>
<td><tt>void</tt></td>
<td>
This is invoked when a vertex is encountered for the first time.
</td>
</tr>

<tr>
<td>Examine Edge</td>
<td><tt>vis.examine_edge(e, g)</tt></td>
<td><tt>void</tt></td>
<td>
This is invoked on every out-edge of each vertex after it is discovered.
</td>
</tr>


<tr>
<td>Tree Edge</td>
<td><tt>vis.tree_edge(e, g)</tt></td>
<td><tt>void</tt></td>
<td>
This is invoked on each edge as it becomes a member of the edges that
form the search tree.</td>
</tr>

<tr>
<td>Back Edge</td>
<td><tt>vis.back_edge(e, g)</tt></td>
<td><tt>void</tt></td>
<td>
This is invoked on the back edges in the graph.  For an undirected
graph there is some ambiguity between tree edges and back edges since
the edge <i>(u,v)</i> and <i>(v,u)</i> are the same edge, but both the
<tt>tree_edge()</tt> and <tt>back_edge()</tt> functions will be
invoked. One way to resolve this ambiguity is to record the tree
edges, and then disregard the back-edges that are already marked as
tree edges.  An easy way to record tree edges is to record
predecessors at the <tt>tree_edge</tt> event point.
</td>
</tr>

<tr>
<td>Forward or Cross Edge</td>
<td><tt>vis.forward_or_cross_edge(e, g)</tt></td>
<td><tt>void</tt></td>
<td>
This is invoked on forward or cross edges in the graph. In an
undirected graph this method is never called.
</td>
</tr>

<tr>
<td>Finish Vertex</td>
<td><tt>vis.finish_vertex(u, g)</tt></td>
<td><tt>void</tt></td>
<td>
This is invoked on vertex <tt>u</tt> after <tt>finish_vertex</tt> has
been called for all the vertices in the DFS-tree rooted at vertex
<tt>u</tt>. If vertex <tt>u</tt> is a leaf in the DFS-tree, then
the <tt>finish_vertex</tt> function is called on <tt>u</tt> after
all the out-edges of <tt>u</tt> have been examined.
</td>
</tr>

</table>

<h3>Models</h3>

<ul>
 <li><a href="./dfs_visitor.html"><tt>dfs_visitor</tt></a>
</ul>

<a name="python"></a>
<h3>Python</h3>

To implement a model of the <tt>DFSVisitor</tt> concept in Python,
create a new class that derives from the <tt>DFSVisitor</tt> type of
the graph, which will be
named <tt><i>GraphType</i>.DFSVisitor</tt>. The events and syntax are
the same as with visitors in C++. Here is an example for the
Python <tt>bgl.Graph</tt> graph type:

<pre>
class count_tree_edges_dfs_visitor(bgl.Graph.DFSVisitor):
  def __init__(self, name_map):
    bgl.Graph.DFSVisitor.__init__(self)
    self.name_map = name_map

  def tree_edge(self, e, g):
    (u, v) = (g.source(e), g.target(e))
    print "Tree edge ",
    print self.name_map[u],
    print " -> ",
    print self.name_map[v]
</pre>

<br>
<HR>
<TABLE>
<TR valign=top>
<TD nowrap>Copyright &copy; 2000-2001</TD><TD>
<A HREF="http://www.boost.org/people/jeremy_siek.htm">Jeremy Siek</A>,
Indiana University (<A
HREF="mailto:jsiek@osl.iu.edu">jsiek@osl.iu.edu</A>)<br>
<A HREF="http://www.boost.org/people/liequan_lee.htm">Lie-Quan Lee</A>, Indiana University (<A HREF="mailto:llee@cs.indiana.edu">llee@cs.indiana.edu</A>)<br>
<A HREF="http://www.osl.iu.edu/~lums">Andrew Lumsdaine</A>,
Indiana University (<A
HREF="mailto:lums@osl.iu.edu">lums@osl.iu.edu</A>)
</TD></TR></TABLE>

</BODY>
</HTML>