148 lines
4.1 KiB
Python
148 lines
4.1 KiB
Python
"""Recursive-descent parser for arithmetic expressions.
|
|
|
|
AST node shapes (stable contract for the eval phase):
|
|
Num(value) — numeric literal; value is int or float
|
|
BinOp(op, left, right) — binary operation; op in ('+', '-', '*', '/')
|
|
Unary(op, operand) — unary operation; op == '-'
|
|
|
|
Grammar (encodes precedence and left-associativity):
|
|
expr → term (('+' | '-') term)*
|
|
term → unary (('*' | '/') unary)*
|
|
unary → '-' unary | primary
|
|
primary → NUMBER | '(' expr ')'
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
from typing import Union
|
|
|
|
|
|
class ParseError(Exception):
|
|
pass
|
|
|
|
|
|
class Num:
|
|
__slots__ = ("value",)
|
|
|
|
def __init__(self, value: Union[int, float]) -> None:
|
|
self.value = value
|
|
|
|
def __repr__(self) -> str:
|
|
return f"Num({self.value!r})"
|
|
|
|
def __eq__(self, other: object) -> bool:
|
|
return isinstance(other, Num) and self.value == other.value
|
|
|
|
|
|
class BinOp:
|
|
__slots__ = ("op", "left", "right")
|
|
|
|
def __init__(self, op: str, left: "Node", right: "Node") -> None:
|
|
self.op = op
|
|
self.left = left
|
|
self.right = right
|
|
|
|
def __repr__(self) -> str:
|
|
return f"BinOp({self.op!r}, {self.left!r}, {self.right!r})"
|
|
|
|
def __eq__(self, other: object) -> bool:
|
|
return (
|
|
isinstance(other, BinOp)
|
|
and self.op == other.op
|
|
and self.left == other.left
|
|
and self.right == other.right
|
|
)
|
|
|
|
|
|
class Unary:
|
|
__slots__ = ("op", "operand")
|
|
|
|
def __init__(self, op: str, operand: "Node") -> None:
|
|
self.op = op
|
|
self.operand = operand
|
|
|
|
def __repr__(self) -> str:
|
|
return f"Unary({self.op!r}, {self.operand!r})"
|
|
|
|
def __eq__(self, other: object) -> bool:
|
|
return (
|
|
isinstance(other, Unary)
|
|
and self.op == other.op
|
|
and self.operand == other.operand
|
|
)
|
|
|
|
|
|
Node = Union[Num, BinOp, Unary]
|
|
|
|
|
|
class _Parser:
|
|
def __init__(self, tokens: list) -> None:
|
|
self._tokens = tokens
|
|
self._pos = 0
|
|
|
|
def _peek(self):
|
|
return self._tokens[self._pos]
|
|
|
|
def _consume(self, kind: str = None):
|
|
tok = self._tokens[self._pos]
|
|
if kind is not None and tok.kind != kind:
|
|
raise ParseError(
|
|
f"expected {kind}, got {tok.kind!r} ({tok.value!r})"
|
|
)
|
|
self._pos += 1
|
|
return tok
|
|
|
|
def parse(self) -> Node:
|
|
if self._peek().kind == "EOF":
|
|
raise ParseError("empty expression")
|
|
node = self._expr()
|
|
if self._peek().kind != "EOF":
|
|
tok = self._peek()
|
|
raise ParseError(
|
|
f"unexpected token {tok.kind!r} ({tok.value!r}) after expression"
|
|
)
|
|
return node
|
|
|
|
def _expr(self) -> Node:
|
|
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:
|
|
node = self._unary()
|
|
while self._peek().kind in ("STAR", "SLASH"):
|
|
op = self._consume().value
|
|
right = self._unary()
|
|
node = BinOp(op, node, right)
|
|
return node
|
|
|
|
def _unary(self) -> Node:
|
|
if self._peek().kind == "MINUS":
|
|
op = self._consume().value
|
|
operand = self._unary()
|
|
return Unary(op, operand)
|
|
return self._primary()
|
|
|
|
def _primary(self) -> Node:
|
|
tok = self._peek()
|
|
if tok.kind == "NUMBER":
|
|
self._consume()
|
|
return Num(tok.value)
|
|
if tok.kind == "LPAREN":
|
|
self._consume("LPAREN")
|
|
node = self._expr()
|
|
if self._peek().kind != "RPAREN":
|
|
raise ParseError("unclosed parenthesis — expected ')'")
|
|
self._consume("RPAREN")
|
|
return node
|
|
if tok.kind == "EOF":
|
|
raise ParseError("unexpected end of expression")
|
|
raise ParseError(f"unexpected token {tok.kind!r} ({tok.value!r})")
|
|
|
|
|
|
def parse(tokens: list) -> Node:
|
|
"""Parse a token list produced by calc.lexer.tokenize into an AST."""
|
|
return _Parser(tokens).parse()
|