344 lines
11 KiB
Python
Executable File
344 lines
11 KiB
Python
Executable File
import re
|
|
import util
|
|
|
|
valid_name_re = re.compile('^[a-zA-Z_][a-zA-Z0-9_]*$')
|
|
full_name_re = re.compile('^([a-zA-Z_]+)([0-9]*)$')
|
|
reserved_names = ['start', 'middle', 'end', 'null']
|
|
|
|
class Token(object):
|
|
def __init__(self, name, rule=None, y=0, x=0, s="", parent=None, matchd={}):
|
|
self.name = name
|
|
self.rule = rule
|
|
self.y = y
|
|
self.x = x
|
|
self.string = s
|
|
self.parent = parent
|
|
self.matchd = matchd
|
|
def parents(self):
|
|
if self.parent is not None:
|
|
parents = self.parent.parents()
|
|
parents.append(self.parent)
|
|
return parents
|
|
else:
|
|
return []
|
|
def domain(self):
|
|
if self.parent is not None:
|
|
names = self.parent.domain()
|
|
else:
|
|
names = []
|
|
if self.name != 'middle':
|
|
names.append(self.rule.name)
|
|
return names
|
|
def fqlist(self):
|
|
if self.parent is not None:
|
|
names = self.parent.domain()
|
|
else:
|
|
names = []
|
|
if self.name == 'start':
|
|
names.append(self.rule.name)
|
|
names.append(self.name)
|
|
return names
|
|
def fqname(self):
|
|
names = self.fqlist()
|
|
return '.'.join(names)
|
|
def copy(self):
|
|
return Token(self.name, self.rule, self.y, self.x, self.string,
|
|
self.parent, self.matchd)
|
|
def add_to_string(self, s):
|
|
self.string += s
|
|
def end_x(self):
|
|
return self.x + len(self.string)
|
|
def __eq__(self, other):
|
|
return (self.y == other.y and self.x == other.x
|
|
and self.name == other.name and self.parent is other.parent and
|
|
self.string == other.string)
|
|
def __repr__(self):
|
|
if len(self.string) < 10:
|
|
s = self.string
|
|
else:
|
|
s = self.string[:10] + '...'
|
|
fields = (self.fqname(), self.rule, self.y, self.x, s)
|
|
return "<Token(%r, %r, %d, %d, %r)>" % fields
|
|
|
|
class Rule:
|
|
reflags = 0
|
|
def __init__(self, name, group=None):
|
|
assert valid_name_re.match(name), 'invalid name %r' % name
|
|
assert name not in reserved_names, "reserved rule name: %r" % name
|
|
self.name = name
|
|
if group is None:
|
|
self.group = name
|
|
else:
|
|
self.group = group
|
|
def match(self, lexer, parent):
|
|
raise Exception, "not implemented"
|
|
def lex(self, lexer, parent, match):
|
|
raise Exception, "not implemented"
|
|
def make_token(self, lexer, s, name, parent=None, matchd={}):
|
|
t = Token(name, self, lexer.y, lexer.x, s, parent, matchd)
|
|
lexer.x += len(s)
|
|
return t
|
|
def get_line(self, lexer):
|
|
return lexer.lines[lexer.y] + '\n'
|
|
|
|
class PatternRule(Rule):
|
|
def __init__(self, name, pattern, group=None):
|
|
Rule.__init__(self, name)
|
|
self.pattern = pattern
|
|
self.re = re.compile(self.pattern, self.reflags)
|
|
def match(self, lexer, parent):
|
|
return self.re.match(self.get_line(lexer), lexer.x)
|
|
def lex(self, lexer, parent, m):
|
|
if m:
|
|
yield self.make_token(lexer, m.group(0), self.name, parent)
|
|
raise StopIteration
|
|
|
|
class NocasePatternRule(PatternRule):
|
|
reflags = re.IGNORECASE
|
|
|
|
class ContextPatternRule(PatternRule):
|
|
def __init__(self, name, pattern, fallback, group=None):
|
|
Rule.__init__(self, name)
|
|
self.pattern = pattern
|
|
self.fallback_re = re.compile(fallback, self.reflags)
|
|
def match(self, lexer, parent):
|
|
try:
|
|
r = re.compile(self.pattern % parent.matchd)
|
|
except KeyError:
|
|
r = self.fallback_re
|
|
return r.match(self.get_line(lexer), lexer.x)
|
|
class NocaseContextPatternRule(ContextPatternRule):
|
|
reflags = re.IGNORECASE
|
|
|
|
class RegionRule(Rule):
|
|
def __init__(self, name, start, grammar, end, group=None):
|
|
Rule.__init__(self, name)
|
|
self.grammar = grammar
|
|
self.end = end
|
|
self.start_re = re.compile(start, self.reflags)
|
|
def match(self, lexer, parent):
|
|
return self.start_re.match(self.get_line(lexer), lexer.x)
|
|
def lex(self, lexer, parent, m):
|
|
t1 = self.make_token(lexer, 'start', None, m, m.groupdict())
|
|
yield t1
|
|
if self.end:
|
|
endre = re.compile(self.end % t1.matchd, self.reflags)
|
|
else:
|
|
endre = None
|
|
for t2 in self._lex(lexer, [t1], 'start', 'end'):
|
|
yield t2
|
|
raise StopIteration
|
|
def resume(self, lexer, toresume):
|
|
assert toresume
|
|
t1 = toresume[0]
|
|
if self.end:
|
|
endre = re.compile(self.end % t1.matchd, self.reflags)
|
|
else:
|
|
endre = None
|
|
for t2 in self._lex(lexer, t1, 'end', endre):
|
|
yield t2
|
|
raise StopIteration
|
|
|
|
def _lex(self, lexer, toresume, stopname, stopre):
|
|
parent = toresume[0]
|
|
reenter = len(toresume) > 1
|
|
null_t = None
|
|
|
|
done = False
|
|
while not done and lexer.y < len(lexer.lines):
|
|
old_y = lexer.y
|
|
line = self.get_line(lexer)
|
|
while not done and lexer.y == old_y and lexer.x < len(line):
|
|
if reenter:
|
|
reenter = False
|
|
for t in toresume[1].rule.resume(lexer, toresume[1:]):
|
|
yield t
|
|
if lexer.y >= len(lexer.lines):
|
|
raise StopIteration
|
|
elif lexer.x >= len(line):
|
|
lexer.y += 1
|
|
lexer.x = 0
|
|
line = self.get_line(lexer)
|
|
|
|
if stopre:
|
|
m = stopre.match(line, lexer.x)
|
|
if m:
|
|
if null_t:
|
|
yield null_t
|
|
null_t = None
|
|
yield self.make_token(lexer, stopname, parent, m, {})
|
|
done = True
|
|
break
|
|
|
|
m = None
|
|
for rule in self.grammar.rules:
|
|
m = rule.match(lexer, parent)
|
|
if m:
|
|
if null_t:
|
|
yield null_t
|
|
null_t = None
|
|
for t in rule.lex(lexer, parent, m):
|
|
yield t
|
|
break
|
|
|
|
if not m:
|
|
if not null_t:
|
|
null_t = Token('null', None, lexer.y, lexer.x, '', parent)
|
|
if lexer.x < len(line):
|
|
null_t.add_to_string(line[lexer.x])
|
|
lexer.x += 1
|
|
if null_t:
|
|
yield null_t
|
|
null_t = None
|
|
if not done and old_y == lexer.y:
|
|
lexer.y += 1
|
|
lexer.x = 0
|
|
raise StopIteration
|
|
class NocaseRegionRule(RegionRule):
|
|
reflags = re.IGNORECASE
|
|
|
|
class DualRegionRule(RegionRule):
|
|
def __init__(self, name, start, grammar1, middle, grammar2, end, group=None):
|
|
Rule.__init__(self, name)
|
|
self.start_re = re.compile(start, self.reflags)
|
|
self.grammar1 = grammar1
|
|
self.middle = middle
|
|
self.grammar2 = grammar2
|
|
self.end = end
|
|
def match(self, lexer, parent):
|
|
return self.start_re.match(self.get_line(lexer), lexer.x)
|
|
def lex(self, lexer, parent, m):
|
|
assert m
|
|
t1 = self.make_token(lexer, 'start', parent, m, m.groupdict())
|
|
yield t1
|
|
|
|
t2 = None
|
|
if self.middle:
|
|
stopre = re.compile(self.middle % t1.groupdict(), self.reflags)
|
|
else:
|
|
stopre = None
|
|
for t2 in self._lex(lexer, [t1], 'middle', stopre):
|
|
yield t2
|
|
|
|
if t2 is not None and t2.name == 'middle':
|
|
if self.end:
|
|
stopre = re.compile(self.end % t2.groupdict(), self.reflags)
|
|
else:
|
|
stopre = None
|
|
for t3 in self._lex(lexer, [t2], 'end', stopre):
|
|
yield t3
|
|
|
|
raise StopIteration
|
|
def resume(self, lexer, toresume):
|
|
assert toresume, "can't resume without tokens to resume!"
|
|
t1 = t2 = None
|
|
if toresume[0].name == 'start':
|
|
t1 = toresume[0]
|
|
elif toresume[0].name == 'middle':
|
|
t2 = toresume[0]
|
|
else:
|
|
raise Exception, "invalid name %r" % toresume[0].name
|
|
|
|
if t1 is not None:
|
|
#assert t1.name == 'start'
|
|
if self.middle:
|
|
stopre = re.compile(self.middle, self.reflags)
|
|
else:
|
|
stopre = None
|
|
for t2 in self._lex_first(lexer, toresume, 'middle', stopre):
|
|
yield t2
|
|
toresume = [t2]
|
|
if t2 is not None:
|
|
assert t2.name == 'middle'
|
|
if self.end:
|
|
stopre = re.compile(self.end, self.reflags)
|
|
else:
|
|
stopre = None
|
|
for t3 in self._lex_second(lexer, toresume, 'end', stopre):
|
|
yield t3
|
|
#toresume = [t3]
|
|
raise StopIteration
|
|
class NocaseDualRegionRule(DualRegionRule):
|
|
reflags = re.IGNORECASE
|
|
|
|
class Grammar:
|
|
rules = []
|
|
def __init__(self):
|
|
# XYZ maybe this is unnecessary
|
|
for rule in self.rules:
|
|
if hasattr(rule, 'grammar') and rule.grammar is None:
|
|
rule.grammar = self
|
|
if hasattr(rule, 'grammar1') and rule.grammar is None:
|
|
rule.grammar = self
|
|
if hasattr(rule, 'grammar2') and rule.grammar is None:
|
|
rule.grammar = self
|
|
grammar = Grammar()
|
|
|
|
class Lexer:
|
|
def __init__(self, name, grammar):
|
|
self.name = name
|
|
self.grammar = grammar
|
|
self.y = 0
|
|
self.x = 0
|
|
self.lines = None
|
|
def get_line(self):
|
|
return self.lines[lexer.y] + '\n'
|
|
def lex(self, lines, y=0, x=0):
|
|
self.y = y
|
|
self.x = x
|
|
self.lines = lines
|
|
self.tokens = []
|
|
for t in self._lex():
|
|
yield t
|
|
raise StopIteration
|
|
def resume(self, lines, y, x, token):
|
|
self.y = y
|
|
self.x = x
|
|
self.lines = lines
|
|
self.tokens = []
|
|
toresume = token.parents()
|
|
|
|
# this is a special case for the "middle" rule of a dual region rule
|
|
i = 1
|
|
while i < len(toresume):
|
|
if toresume[i].name == 'middle' and toresume[i-1].name == 'start':
|
|
del toresume[i-1]
|
|
else:
|
|
i += 1
|
|
|
|
if toresume:
|
|
for t in toresume[0].rule.resume(self, toresume):
|
|
yield t
|
|
for t in self._lex():
|
|
yield t
|
|
raise StopIteration
|
|
|
|
def _lex(self):
|
|
null_t = None
|
|
while self.y < len(self.lines):
|
|
line = self.get_line()
|
|
while not done and self.x < len(line):
|
|
m = None
|
|
for rule in self.grammar.rules:
|
|
m = rule.match(self, parent)
|
|
if m:
|
|
if null_t:
|
|
yield null_t
|
|
null_t = None
|
|
for t in rule.lex(self, parent, m):
|
|
yield t
|
|
break
|
|
|
|
if not m:
|
|
if not null_t:
|
|
null_t = Token('null', None, self.y, self.x, '', parent)
|
|
if self.x < len(line):
|
|
null_t.add_to_string(line[self.x])
|
|
self.x += 1
|
|
if null_t:
|
|
yield null_t
|
|
null_t = None
|
|
self.y += 1
|
|
self.x = 0
|
|
raise StopIteration
|