Plasma/plasma_parser.py

from lexer import Lexer
from lexer_token import Token, TokenType
from typing import Callable
from enum import Enum, auto

from AST import Statement, Expression, Program
from AST import ExpressionStatement, AssignmentStatement
from AST import InfixExpression
from AST import IntegerLiteral, FloatLiteral, IdentifierLiteral

class PrecedenceType(Enum):
	P_LOWEST = 0
	P_EQUALS = auto()
	P_LESSGREATER = auto()
	P_SUM = auto()
	P_PRODUCT = auto()
	P_EXPONENT = auto()
	P_PREFIX = auto()
	P_CALL = auto()
	P_INDEX = auto()

PRECEDENCES: dict[TokenType, PrecedenceType] = {
	TokenType.PLUS: PrecedenceType.P_SUM,
	TokenType.MINUS: PrecedenceType.P_SUM,
	TokenType.ASTERISK: PrecedenceType.P_PRODUCT,
	TokenType.SLASH: PrecedenceType.P_PRODUCT,
	TokenType.MODULUS: PrecedenceType.P_PRODUCT,
	TokenType.POW: PrecedenceType.P_EXPONENT
}

class Parser:
	def __init__(self, lexer: Lexer) -> None:
		self.lexer: Lexer = lexer

		self.errors: list[str] = []

		self.current_token: Token = None
		self.peek_token: Token = None

		self.prefix_parse_functions: dict[Token, Callable] = { # -1

			TokenType.INT: self.__parse_int_literal,
			TokenType.FLOAT: self.__parse_float_literal,
			TokenType.LPAREN: self.__parse_grouped_expression

		} 
		self.infix_parse_functions: dict[Token, Callable] = { # 5 + 5

			TokenType.PLUS: self.__parse_infix_expression,
			TokenType.MINUS: self.__parse_infix_expression,
			TokenType.SLASH: self.__parse_infix_expression,
			TokenType.ASTERISK: self.__parse_infix_expression,
			TokenType.POW: self.__parse_infix_expression,
			TokenType.MODULUS: self.__parse_infix_expression

		} 

		self.__next_token()
		self.__next_token()

	# region Parser helpers
	def __next_token(self) -> None:
		self.current_token = self.peek_token
		self.peek_token = self.lexer.next_token()

	def __current_token_is(self, tt: TokenType) -> bool:
		return self.current_token.type == tt

	def __peek_token_is(self, tt: TokenType) -> bool:
		return self.peek_token.type == tt
	
	def __expect_peek(self, tt: TokenType) -> bool:
		if self.__peek_token_is(tt):
			self.__next_token()
			return True
		else:
			self.__peek_error(tt)
			return False
		
	def __current_precedence(self) -> PrecedenceType:
		prec = PRECEDENCES.get(self.current_token.type)
		if prec is None:
			return PrecedenceType.P_LOWEST
		return prec
	
	def __peek_precedence(self) -> PrecedenceType:
		prec = PRECEDENCES.get(self.peek_token.type)
		if prec is None:
			return PrecedenceType.P_LOWEST
		return prec
		
	def __peek_error(self, tt: TokenType):
		self.errors.append(f"Expected next token to be {tt}, got {self.peek_token.type} instead.")

	def __no_prefix_parse_function_error(self, tt: TokenType):
		self.errors.append(f"No Prefix Parse Function for {tt} found.")
	# endregion

	def parse_program(self) -> None:
		program: Program = Program()

		while self.current_token.type != TokenType.EOF:
			stmt: Statement = self.__parse_statement()

			if stmt is not None:
				program.statements.append(stmt)

			self.__next_token()

		return program
	
	# region Statement Methods
	def __parse_statement(self) -> Statement:
		match self.current_token.type:
			case TokenType.IDENT:
				return self.__parse_assignment_statement()
			case _:
				return self.__parse_expression_statement()
	
	def __parse_expression_statement(self) -> ExpressionStatement:
		expr = self.__parse_expression(PrecedenceType.P_LOWEST)

		if self.__peek_token_is(TokenType.SEMICOLON):
			self.__next_token()

		stmt: ExpressionStatement = ExpressionStatement(expr=expr)

		return stmt
	
	def __parse_assignment_statement(self) -> AssignmentStatement:
		# x: Int = 10;
		stmt: AssignmentStatement = AssignmentStatement(name=IdentifierLiteral(self.current_token.literal))

		if not self.__expect_peek(TokenType.COLON):
			return None
		
		if not self.__expect_peek(TokenType.TYPE):
			return None
		
		stmt.value_type = self.current_token.literal
		
		if not self.__expect_peek(TokenType.EQ):
			return None
		
		self.__next_token()

		stmt.value = self.__parse_expression(PrecedenceType.P_LOWEST)

		while not self.__current_token_is(TokenType.SEMICOLON) and not self.__current_token_is(TokenType.EOF):
			self.__next_token()

		return stmt
	# endregion

	# region Expression Methods
	def __parse_expression(self, precedence: PrecedenceType) -> Expression:
		prefix_func: Callable | None = self.prefix_parse_functions.get(self.current_token.type)
		if prefix_func is None:
			self.__no_prefix_parse_function_error(self.current_token.type)
			return None
		
		left_expr: Expression = prefix_func()
		while not self.__peek_token_is(TokenType.SEMICOLON) and precedence.value < self.__peek_precedence().value:
			infix_func: Callable | None = self.infix_parse_functions.get(self.peek_token.type)
			if infix_func is None:
				return left_expr
			
			self.__next_token()

			left_expr = infix_func(left_expr)
		
		return left_expr

	def __parse_infix_expression(self, left_node: Expression) -> Expression:
		infix_expr: Expression = InfixExpression(left_node=left_node, operator=self.current_token.literal)

		precedence = self.__current_precedence()

		self.__next_token()

		infix_expr.right_node = self.__parse_expression(precedence)

		return infix_expr
	
	def __parse_grouped_expression(self) -> Expression:
		self.__next_token()

		expr: Expression = self.__parse_expression(PrecedenceType.P_LOWEST)

		if not self.__expect_peek(TokenType.RPAREN):
			return None
		
		return expr
	# endregion

	# region Prefix Methods
	def __parse_int_literal(self) -> Expression:
		int_lit: IntegerLiteral = IntegerLiteral()

		try:
			int_lit.value = int(self.current_token.literal)
		except:
			self.errors.append(f"Could not parse \"{self.current_token.literal}\" as an integer.")
			return None
		
		return int_lit
	
	def __parse_float_literal(self) -> Expression:
		float_lit: FloatLiteral = FloatLiteral()

		try:
			float_lit.value = float(self.current_token.literal)
		except:
			self.errors.append(f"Could not parse \"{self.current_token.literal}\" as an float.")
			return None
		
		return float_lit
	# endregion