def convert_id(self, ):
if self.current_token == '':
raise PathExprParsingError('empty ID at position {}'.format(
self.pos))
token, self.current_token = self.current_token, ''
return token
def convert_slice_element(self):
try:
ret = None if self.current_token == '' else int(self.current_token)
self.current_token = ''
return ret
except ValueError:
raise PathExprParsingError('invalid slice syntax: {!r} at position {}'.format(self.current_token, self.pos))
def create_slice_object(self):
if len(self.current_slice_elements) == 0:
if self.bare_id_matches_all:
slc_obj = slice(None, None, None)
else:
slc_obj = 0
elif len(self.current_slice_elements) == 1:
# This must be true or the parsing logic is wrong
assert isinstance(self.current_slice_elements[0], int)
if self.current_slice_elements[0] >= 0:
slc_obj = self.current_slice_elements[0]
else:
slc_obj = slice(self.current_slice_elements[0],
self.current_slice_elements[0] + 1 if self.current_slice_elements[0] != -1 else None,
None)
elif len(self.current_slice_elements) <= 3: # 2 or 3
slc_obj = slice(*self.current_slice_elements)
else:
raise PathExprParsingError('slice can have at most three indices')
self.current_slice_elements = []
return slc_obj
def unexpected_char_error(c, idx):
return PathExprParsingError('unexpected char: {!r} at position {}'.format(
c, idx))
def parse(self, path_expr):
path_expr_stripped = path_expr.strip()
if path_expr_stripped == '':
raise PathExprParsingError('Empty path expression')
# A path expression should always start with one of the following chars
# Otherwise fail fast.
if path_expr_stripped[
0] not in '@/>0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ':
raise unexpected_char_error(path_expr_stripped[0],
path_expr.find(path_expr_stripped[0]))
self.reset()
self.node_path = NodePath(path_expr)
self.current_state = STATE_START_PARSING
self.current_token = ''
while self.pos < len(path_expr):
c = path_expr[self.pos]
if c in string.whitespace:
pass # all whitespaces are ignored
elif c == '@': # start of subset specifier
if self.current_state == STATE_START_PARSING:
self.current_state = STATE_START_SUBSET
else:
raise unexpected_char_error(c, self.pos)
elif c == '[':
self.handle_left_bracket()
elif c in (':', ']'):
self.handle_colon_and_right_bracket(c)
elif c in (PATH_SEPARATOR_CHILD, PATH_SEPARATOR_ATTRIB,
PATH_SEPARATOR_DESCEND):
self.handle_separator(c)
else:
if self.current_state in (STATE_START_ID,
STATE_START_SUBSET_SLICE_0,
STATE_START_SUBSET_SLICE_X,
STATE_START_SLICE_0,
STATE_START_SLICE_X):
self.current_token += c
elif self.current_state == STATE_START_PARSING:
self.handle_separator(PATH_SEPARATOR_DESCEND)
self.current_token += c
else:
raise unexpected_char_error(c, self.pos)
self.pos += 1
if self.current_state == STATE_START_ID:
self.current_id = self.convert_id()
self.add_new_path_component()
elif self.current_state == STATE_STOP_SLICE:
self.add_new_path_component()
elif self.current_token != '':
raise unexpected_char_error(self.current_token[0],
self.pos - len(self.current_token))
return self.node_path