119 lines
2.8 KiB
Python
119 lines
2.8 KiB
Python
"""Recursive-descent parser for arithmetic expressions.
|
|
|
|
Grammar (precedence low→high):
|
|
expr ::= term ( ('+' | '-') term )*
|
|
term ::= factor ( ('*' | '/') factor )*
|
|
factor ::= NUMBER | '(' expr ')' | '-' factor
|
|
|
|
AST node shapes (all dataclasses):
|
|
Num(value) — numeric leaf; value is int or float
|
|
BinOp(op, left, right) — op in ('+', '-', '*', '/'), children are nodes
|
|
Unary(op, operand) — op is '-', operand is a node
|
|
"""
|
|
|
|
from dataclasses import dataclass
|
|
from typing import Any
|
|
|
|
from calc.lexer import Token, tokenize
|
|
|
|
|
|
class ParseError(Exception):
|
|
pass
|
|
|
|
|
|
@dataclass
|
|
class Num:
|
|
value: Any
|
|
|
|
def __repr__(self):
|
|
return f"Num(value={self.value!r})"
|
|
|
|
|
|
@dataclass
|
|
class BinOp:
|
|
op: str
|
|
left: Any
|
|
right: Any
|
|
|
|
def __repr__(self):
|
|
return f"BinOp(op={self.op!r}, left={self.left!r}, right={self.right!r})"
|
|
|
|
|
|
@dataclass
|
|
class Unary:
|
|
op: str
|
|
operand: Any
|
|
|
|
def __repr__(self):
|
|
return f"Unary(op={self.op!r}, operand={self.operand!r})"
|
|
|
|
|
|
class _Parser:
|
|
def __init__(self, tokens: list):
|
|
self._tokens = tokens
|
|
self._pos = 0
|
|
|
|
def _peek(self) -> Token:
|
|
return self._tokens[self._pos]
|
|
|
|
def _consume(self) -> Token:
|
|
tok = self._tokens[self._pos]
|
|
self._pos += 1
|
|
return tok
|
|
|
|
def _expect(self, kind: str) -> Token:
|
|
tok = self._peek()
|
|
if tok.kind != kind:
|
|
raise ParseError(f"Expected {kind} but got {tok.kind!r}")
|
|
return self._consume()
|
|
|
|
def parse(self):
|
|
node = self._expr()
|
|
if self._peek().kind != 'EOF':
|
|
raise ParseError(f"Unexpected token {self._peek()!r}")
|
|
return node
|
|
|
|
def _expr(self):
|
|
node = self._term()
|
|
while self._peek().kind in ('PLUS', 'MINUS'):
|
|
op = self._consume().value
|
|
right = self._term()
|
|
node = BinOp(op, node, right)
|
|
return node
|
|
|
|
def _term(self):
|
|
node = self._factor()
|
|
while self._peek().kind in ('STAR', 'SLASH'):
|
|
op = self._consume().value
|
|
right = self._factor()
|
|
node = BinOp(op, node, right)
|
|
return node
|
|
|
|
def _factor(self):
|
|
tok = self._peek()
|
|
|
|
if tok.kind == 'NUMBER':
|
|
self._consume()
|
|
return Num(tok.value)
|
|
|
|
if tok.kind == 'LPAREN':
|
|
self._consume()
|
|
node = self._expr()
|
|
self._expect('RPAREN')
|
|
return node
|
|
|
|
if tok.kind == 'MINUS':
|
|
self._consume()
|
|
operand = self._factor()
|
|
return Unary('-', operand)
|
|
|
|
if tok.kind == 'EOF':
|
|
raise ParseError("Unexpected end of input")
|
|
|
|
raise ParseError(f"Unexpected token {tok!r}")
|
|
|
|
|
|
def parse(tokens: list):
|
|
"""Parse a token list (from tokenize()) into an AST. Raises ParseError on bad input."""
|
|
return _Parser(tokens).parse()
|