A little more documentation

jgraham · jgraham · commit 91ae0891e29f · 2007-01-02T22:41:30.000Z
--HG--
extra : convert_revision : svn%3Aacbfec75-9323-0410-a652-858a13e371e0/trunk%40406
diff --git a/README b/README
@@ -24,10 +24,17 @@ f = open("mydocument.html")
 parser = html5lib.HTMLParser()
 document = parser.parse(f)
 
-The returned document is a simple DOM-like structure which can be
-navigated using the .parent and .childNode attributes on each
-element. Future releases will support a more substantial tree
-implementation. 
+By default, the returned document is a simple DOM-like structure which
+can be navigated using the .parent and .childNode attributes on each
+element. 
+
+It is also possible to generate an ElementTree tree, this requires the use of the "tree" argument to the parser:
+
+from html5lib.treebuilders import etree
+parser = html5lib.HTMLParser(tree=etree.TreeBuilder)
+
+Intrepid users may write their own treebuilder implementations - see
+help(html5lib.treebuilders) for more information
 
 More documentation is avaliable in the docstrings.
 
diff --git a/src/parser.py b/src/parser.py
@@ -35,11 +35,17 @@ class HTMLParser(object):
     """Main parser class"""
 
     def __init__(self, strict = False, tree=simpletree.TreeBuilder):
+        """HTML parser. Generates a tree structure from a stream of (possibly
+        malformed) HTML.  
+        strict - raise an exception when a parse error is encountered 
+        tree - a treebuilder class controlling the type of tree that will be 
+        returned (default - html5lib.simpletree.TreeBuilder)"""
+
         # Raise an exception on the first error encountered
         self.strict = strict
         self.errors = []
 
-        self.tree = tree()
+        self._treeCls = tree
 
         self.phases = {
             "initial": InitialPhase(self, self.tree),
@@ -70,7 +76,7 @@ def parse(self, stream, innerHTML=False):
         issues have not yet been dealt with."""
 
         # XXX - need to ensure the tree is reset here
-        # Why? -- anne
+        self.tree = self._treeCls()
 
         # We don't actually support innerHTML yet but this should allow
         # assertations
