def regex(self, ast, *args):
pattern = ast
try:
re.compile(pattern)
except (TypeError, re.error) as e:
raise FailedSemantics('regexp error: ' + str(e))
return pattern
def symbol_or_pair(self, ast):
string = ast.token.strip()
failmsg = []
pat = re.compile(
r"""^
(?P<up>[a-zšžåäöüõA-ZÅÄÖ0-9]*
|
\{[a-zåäöüõA-ZÅÄÖØ]+\})
:
(?P<lo>[a-zšžåäöüõA-ZÅÄÖØ]*)
$""", re.X)
m = re.match(pat, string)
if m: # it is a pair with a colon
up = m.group("up")
up_quoted = re.sub(r"([{}])", r"%\1", up)
lo = m.group("lo")
if up and (up not in cfg.input_symbol_set):
failmsg.append("input symbol '{}'".format(up))
if lo and (lo not in cfg.output_symbol_set):
failmsg.append("output symbol '{}'".format(lo))
if up and lo and ((up, lo) not in cfg.symbol_pair_set):
failmsg.append("symbol pair '{}'".format(string))
if failmsg:
cfg.error_message = " and ".join(failmsg) + " not in alphabet"
raise FailedSemantics(cfg.error_message)
elif up and lo: # it is a valid pair with a colon
return "{}:{}".format(up_quoted, lo)
elif up and (not lo):
return "[{} .o. PI]".format(up_quoted)
elif (not up) and lo:
return "[PI .o. {}]".format(lo)
else:
return "PI"
m = re.fullmatch(r"[a-zåäöüõA-ZÅÄÖØ]+", string)
if m: # its either a defined sym or a surf ch
if string in cfg.definitions:
return "{}".format(string)
elif (string
in cfg.output_symbol_set) and (string
in cfg.input_symbol_set):
return "{}:{}".format(string, string)
elif string in {'BEGIN', 'END'}:
return string
cfg.error_message = "'" + string + "' is an invalid pair/definend symbol"
raise FailedSemantics(cfg.error_message)
def symbol_or_pair(self, ast):
string = ast.token.strip()
failmsg = []
pat = re.compile(
r"""^
(?P<up>[a-zšžåäöüõA-ZÅÄÖ0-9'´`]*
|
\{[a-zåäöüõA-ZÅÄÖØ'´`]+\})
:
(?P<lo>[a-zšžåäöüõA-ZÅÄÖØ'´`]*)
$""", re.X)
m = re.match(pat, string)
if m: # it is a pair with a colon
up = m.group("up")
up_quoted = re.sub(r"([{}])", r"%\1", up)
lo = m.group("lo")
if up and (up not in cfg.input_symbol_set):
failmsg.append("input symbol '{}'".format(up))
if lo and (lo not in cfg.output_symbol_set):
failmsg.append("output symbol '{}'".format(lo))
if up and lo and ((up, lo) not in cfg.symbol_pair_set):
failmsg.append("symbol pair '{}'".format(string))
if failmsg:
cfg.error_message = " and ".join(failmsg) + " not in alphabet"
raise FailedSemantics(cfg.error_message)
elif up and lo: # it is a valid pair with a colon
result_fst = hfst.regex(up_quoted + ':' + lo)
result_fst.set_name(string)
return result_fst
elif up and (not lo):
result_fst = hfst.regex(up_quoted)
result_fst.compose(cfg.all_pairs_fst)
result_fst.set_name(string)
return result_fst
elif (not up) and lo:
result_fst = cfg.all_pairs_fst.copy()
lo_fst = hfst.regex(lo)
result_fst.compose(lo_fst)
result_fst.set_name(string)
return result_fst
else:
result_fst = cfg.all_pairs_fst.copy()
result_fst.set_name("PI")
return result_fst
m = re.fullmatch(r"[a-zåäöšžüõA-ZÅÄÖØ'´`]+", string)
if m: # its either a defined sym or a surf ch
if string in cfg.definitions:
result_fst = cfg.definitions[string].copy()
result_fst.set_name(string)
return result_fst
elif (string
in cfg.output_symbol_set) and (string
in cfg.input_symbol_set):
result_fst = hfst.regex(string)
result_fst.set_name(string)
return result_fst
elif string in {'BEGIN', 'END'}:
result_fst = hfst.regex(string)
result_fst.set_name(string)
return result_fst
cfg.error_message = "'" + string + "' is an invalid pair/definend symbol"
raise FailedSemantics(cfg.error_message)
def known_name(self, name):
if name not in self.rules:
raise FailedSemantics('rule "%s" not yet defined' % str(name))
return name
def new_name(self, name):
if name in self.rules:
raise FailedSemantics('rule "%s" already defined' % str(name))
return name
def token(self, ast, *args):
token = ast
if not token:
raise FailedSemantics('empty token')
return grammars.Token(token)
def ul_marker(self, ast):
ctx = self._context
if ctx.substate is not None and not ctx.tokenizer.match(
"*" * ctx.substate):
raise FailedSemantics("not at correct level")
return ast
def error(self, msg):
raise FailedSemantics(msg)