def get_node_type_id(self, node):
if isinstance(node, ASTNode):
type_repr = typename(node.type)
return self.node_type_to_id[type_repr]
else:
# assert isinstance(node, str)
# it is a type
type_repr = typename(node)
return self.node_type_to_id[type_repr]
def pretty_print_helper(self, sb, depth, new_line=False):
if new_line:
sb.write('\n')
for i in range(depth):
sb.write(' ')
sb.write('(')
sb.write(typename(self.type))
if self.label is not None:
sb.write('{%s}' % self.label)
if self.value is not None:
sb.write('{val=%s}' % self.value)
if len(self.children) == 0:
sb.write(')')
return
sb.write(' ')
new_line = True
for child in self.children:
child.pretty_print_helper(sb, depth + 2, new_line)
sb.write('\n')
for i in range(depth):
sb.write(' ')
sb.write(')')
def __repr__(self):
parent = typename(self.type)
if self.label is not None:
parent += '{%s}' % self.label
if self.value is not None:
parent += '{val=%s}' % self.value
return '%s -> %s' % (parent, ', '.join(
[repr(c) for c in self.children]))
def __init__(self, rules):
"""
instantiate a grammar with a set of production rules of type Rule
"""
self.rules = rules
self.rule_index = defaultdict(list)
self.rule_to_id = OrderedDict()
node_types = set()
lhs_nodes = set()
rhs_nodes = set()
for rule in self.rules:
self.rule_index[rule.parent].append(rule)
# we also store all unique node types
for node in rule.nodes:
node_types.add(typename(node.type))
lhs_nodes.add(rule.parent)
for child in rule.children:
rhs_nodes.add(child.as_type_node)
root_node = lhs_nodes - rhs_nodes
try:
assert len(root_node) == 1
except AssertionError:
print(root_node)
raise AssertionError
self.root_node = next(iter(root_node))
self.terminal_nodes = rhs_nodes - lhs_nodes
self.terminal_types = set([n.type for n in self.terminal_nodes])
self.node_type_to_id = OrderedDict()
for i, type in enumerate(node_types, start=0):
self.node_type_to_id[type] = i
for gid, rule in enumerate(rules, start=0):
self.rule_to_id[rule] = gid
self.id_to_rule = OrderedDict(
(v, k) for (k, v) in list(self.rule_to_id.items()))
def __repr__(self):
repr_str = ''
# if not self.is_leaf:
repr_str += '('
repr_str += typename(self.type)
if self.label is not None:
repr_str += '{%s}' % self.label
if self.value is not None:
repr_str += '{val=%s}' % self.value
# if not self.is_leaf:
for child in self.children:
repr_str += ' ' + child.__repr__()
repr_str += ')'
return repr_str