134 lines
3.3 KiB
Python
134 lines
3.3 KiB
Python
"""Recursive-descent parser for calc expressions.
|
|
|
|
AST node shapes (all are dataclasses with __repr__):
|
|
Num(value: int|float)
|
|
BinOp(op: str, left: Node, right: Node) -- op in {'+','-','*','/'}
|
|
Unary(op: str, operand: Node) -- op == '-'
|
|
|
|
Grammar:
|
|
expr ::= term (( '+' | '-' ) term)*
|
|
term ::= unary (( '*' | '/' ) unary)*
|
|
unary ::= '-' unary | primary
|
|
primary::= NUMBER | '(' expr ')'
|
|
"""
|
|
from __future__ import annotations
|
|
from dataclasses import dataclass
|
|
from typing import Union
|
|
|
|
from calc.lexer import Token
|
|
|
|
|
|
class ParseError(Exception):
|
|
pass
|
|
|
|
|
|
# ---------- AST nodes ----------
|
|
|
|
@dataclass
|
|
class Num:
|
|
value: Union[int, float]
|
|
|
|
def __repr__(self) -> str:
|
|
return f"Num({self.value!r})"
|
|
|
|
|
|
@dataclass
|
|
class BinOp:
|
|
op: str
|
|
left: "Node"
|
|
right: "Node"
|
|
|
|
def __repr__(self) -> str:
|
|
return f"BinOp({self.op!r}, {self.left!r}, {self.right!r})"
|
|
|
|
|
|
@dataclass
|
|
class Unary:
|
|
op: str
|
|
operand: "Node"
|
|
|
|
def __repr__(self) -> str:
|
|
return f"Unary({self.op!r}, {self.operand!r})"
|
|
|
|
|
|
Node = Union[Num, BinOp, Unary]
|
|
|
|
|
|
# ---------- parser ----------
|
|
|
|
class _Parser:
|
|
def __init__(self, tokens: list[Token]) -> None:
|
|
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!r} but got {tok.kind!r} ({tok.value!r})"
|
|
)
|
|
return self._consume()
|
|
|
|
def parse(self) -> Node:
|
|
if self._peek().kind == 'EOF':
|
|
raise ParseError("empty input")
|
|
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()
|
|
node = self._expr()
|
|
self._expect('RPAREN')
|
|
return node
|
|
if tok.kind == 'EOF':
|
|
raise ParseError("unexpected end of input")
|
|
raise ParseError(
|
|
f"unexpected token {tok.kind!r} ({tok.value!r})"
|
|
)
|
|
|
|
|
|
def parse(tokens: list[Token]) -> Node:
|
|
"""Parse a list of tokens into an AST. Raises ParseError on malformed input."""
|
|
return _Parser(tokens).parse()
|