How to get the AST-tree instead of a list when parsing in Python with ANTLR?

Question

I get simple antlr3 grammar MicroXpath and build lexer and parser for Python.

Then I wrote a simple test code:

import antlr3
from XPathLexer import XPathLexer
from XPathParser import XPathParser

def print_level_order(tree, indent):
    print('{0}{1}'.format('   '*indent, tree.text, tree.getType()))
    for child in tree.getChildren():
        print_level_order(child, indent+1)

input = 'descendant::name[class/name[test="x"]="File"]'
char_stream = antlr3.ANTLRStringStream(input)
lexer = XPathLexer(char_stream)
tokens = antlr3.CommonTokenStream(lexer)
parser = XPathParser(tokens)
tree = parser.xPath().tree

print_level_order(tree, 0)

Result:

None
   descendant
   :
   name
   [
   class
   /
   name
   [
   test
   =
   "x"
   ]
   =
   "File"
   ]

Where the tree? This is a linear list! What am I doing wrong? Or using ANTLR can not build a tree?

Bart Kiers Bart Kiers · Accepted Answer · 2013-03-04T16:28:41

Only adding output=AST; to the options{...} is not enough: you'll have to tell ANTLR which nodes/tokens to exclude from the AST (if any), and which nodes/tokens you want to make the root of a (sub) tree. Not doing so results in a flat tree, as you already observed.

Checkout this Q&A to find out how to create a hierarchy in your tree: How to output the AST built using ANTLR?

How to get the AST-tree instead of a list when parsing in Python with ANTLR?

1 Answers