This repo contains code to mirror other repos. It also contains the code that is getting mirrored.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

205 lines
6.9 KiB

  1. """This module implements a LALR(1) Parser
  2. """
  3. # Author: Erez Shinan (2017)
  4. # Email : erezshin@gmail.com
  5. from copy import deepcopy, copy
  6. from ..exceptions import UnexpectedInput, UnexpectedToken
  7. from ..lexer import Token
  8. from ..utils import Serialize
  9. from .lalr_analysis import LALR_Analyzer, Shift, Reduce, IntParseTable
  10. from .lalr_interactive_parser import InteractiveParser
  11. from lark.exceptions import UnexpectedCharacters, UnexpectedInput, UnexpectedToken
  12. from ..grammar import END
  13. ###{standalone
  14. class LALR_Parser(Serialize):
  15. def __init__(self, parser_conf, debug=False):
  16. analysis = LALR_Analyzer(parser_conf, debug=debug)
  17. analysis.compute_lalr()
  18. callbacks = parser_conf.callbacks
  19. self._parse_table = analysis.parse_table
  20. self.parser_conf = parser_conf
  21. self.parser = _Parser(analysis.parse_table, callbacks, debug)
  22. @classmethod
  23. def deserialize(cls, data, memo, callbacks, debug=False):
  24. inst = cls.__new__(cls)
  25. inst._parse_table = IntParseTable.deserialize(data, memo)
  26. inst.parser = _Parser(inst._parse_table, callbacks, debug)
  27. return inst
  28. def serialize(self, memo):
  29. return self._parse_table.serialize(memo)
  30. def parse_interactive(self, lexer, start):
  31. return self.parser.parse(lexer, start, start_interactive=True)
  32. def parse(self, lexer, start, on_error=None):
  33. try:
  34. return self.parser.parse(lexer, start)
  35. except UnexpectedInput as e:
  36. if on_error is None:
  37. raise
  38. while True:
  39. if isinstance(e, UnexpectedCharacters):
  40. s = e.interactive_parser.lexer_state.state
  41. p = s.line_ctr.char_pos
  42. if not on_error(e):
  43. raise e
  44. if isinstance(e, UnexpectedCharacters):
  45. # If user didn't change the character position, then we should
  46. if p == s.line_ctr.char_pos:
  47. s.line_ctr.feed(s.text[p:p+1])
  48. try:
  49. return e.interactive_parser.resume_parse()
  50. except UnexpectedToken as e2:
  51. if (isinstance(e, UnexpectedToken)
  52. and e.token.type == e2.token.type == END
  53. and e.interactive_parser == e2.interactive_parser):
  54. # Prevent infinite loop
  55. raise e2
  56. e = e2
  57. except UnexpectedCharacters as e2:
  58. e = e2
  59. class ParseConf(object):
  60. __slots__ = 'parse_table', 'callbacks', 'start', 'start_state', 'end_state', 'states'
  61. def __init__(self, parse_table, callbacks, start):
  62. self.parse_table = parse_table
  63. self.start_state = self.parse_table.start_states[start]
  64. self.end_state = self.parse_table.end_states[start]
  65. self.states = self.parse_table.states
  66. self.callbacks = callbacks
  67. self.start = start
  68. class ParserState(object):
  69. __slots__ = 'parse_conf', 'lexer', 'state_stack', 'value_stack'
  70. def __init__(self, parse_conf, lexer, state_stack=None, value_stack=None):
  71. self.parse_conf = parse_conf
  72. self.lexer = lexer
  73. self.state_stack = state_stack or [self.parse_conf.start_state]
  74. self.value_stack = value_stack or []
  75. @property
  76. def position(self):
  77. return self.state_stack[-1]
  78. # Necessary for match_examples() to work
  79. def __eq__(self, other):
  80. if not isinstance(other, ParserState):
  81. return NotImplemented
  82. return len(self.state_stack) == len(other.state_stack) and self.position == other.position
  83. def __copy__(self):
  84. return type(self)(
  85. self.parse_conf,
  86. self.lexer, # XXX copy
  87. copy(self.state_stack),
  88. deepcopy(self.value_stack),
  89. )
  90. def copy(self):
  91. return copy(self)
  92. def feed_token(self, token, is_end=False):
  93. state_stack = self.state_stack
  94. value_stack = self.value_stack
  95. states = self.parse_conf.states
  96. end_state = self.parse_conf.end_state
  97. callbacks = self.parse_conf.callbacks
  98. while True:
  99. state = state_stack[-1]
  100. try:
  101. action, arg = states[state][token.type]
  102. except KeyError:
  103. expected = {s for s in states[state].keys() if s.isupper()}
  104. raise UnexpectedToken(token, expected, state=self, interactive_parser=None)
  105. assert arg != end_state
  106. if action is Shift:
  107. # shift once and return
  108. # assert not is_end
  109. state_stack.append(arg)
  110. value_stack.append(token if token.type not in callbacks else callbacks[token.type](token))
  111. return
  112. else:
  113. # reduce+shift as many times as necessary
  114. rule = arg
  115. size = len(rule.expansion)
  116. if size:
  117. s = value_stack[-size:]
  118. del state_stack[-size:]
  119. del value_stack[-size:]
  120. else:
  121. s = []
  122. value = callbacks[rule](s)
  123. _action, new_state = states[state_stack[-1]][rule.origin.name]
  124. assert _action is Shift
  125. state_stack.append(new_state)
  126. value_stack.append(value)
  127. if is_end and state_stack[-1] == end_state:
  128. return value_stack[-1]
  129. class _Parser(object):
  130. def __init__(self, parse_table, callbacks, debug=False):
  131. self.parse_table = parse_table
  132. self.callbacks = callbacks
  133. self.debug = debug
  134. def parse(self, lexer, start, value_stack=None, state_stack=None, start_interactive=False):
  135. parse_conf = ParseConf(self.parse_table, self.callbacks, start)
  136. parser_state = ParserState(parse_conf, lexer, state_stack, value_stack)
  137. if start_interactive:
  138. return InteractiveParser(self, parser_state, parser_state.lexer)
  139. return self.parse_from_state(parser_state)
  140. def parse_from_state(self, state):
  141. # Main LALR-parser loop
  142. try:
  143. token = None
  144. for token in state.lexer.lex(state):
  145. state.feed_token(token)
  146. token = Token.new_borrow_pos(END, '', token) if token else Token(END, '', 0, 1, 1)
  147. while True:
  148. x = state.feed_token(token, True)
  149. if x is not None:
  150. return x
  151. except UnexpectedInput as e:
  152. try:
  153. e.interactive_parser = InteractiveParser(self, state, state.lexer)
  154. except NameError:
  155. pass
  156. raise e
  157. except Exception as e:
  158. if self.debug:
  159. print("")
  160. print("STATE STACK DUMP")
  161. print("----------------")
  162. for i, s in enumerate(state.state_stack):
  163. print('%d)' % i , s)
  164. print("")
  165. raise
  166. ###}