Sheerka-Old/src/parsers/BaseParser.py

from dataclasses import dataclass
from typing import Union

from core.builtin_concepts import BuiltinConcepts, ParserResultConcept
from core.concept import Concept
from core.global_symbols import ErrorObj
from core.sheerka.ExecutionContext import ExecutionContext
from core.sheerka.services.SheerkaExecute import ParserInput
from core.tokenizer import TokenKind, Token, LexerError


class ErrorSink:
    def __init__(self):
        self.sink = []

    def __repr__(self):
        return f"Errors({self.sink})"

    def add_error(self, error):
        self.sink.append(error)

    def clear(self):
        self.sink.clear()

    @property
    def has_error(self):
        return len(self.sink) > 0


@dataclass()
class Node:
    pass


@dataclass()
class ParsingError(Node, ErrorObj):
    pass


@dataclass()
class UnexpectedTokenParsingError(ParsingError):
    message: str
    token: Union[Token, str]
    expected_tokens: list

    def __eq__(self, other):
        if id(other) == id(self):
            return True

        if not isinstance(other, UnexpectedTokenParsingError):
            return False

        if self.message != other.message:
            return False

        to_compare = self.token.repr_value if isinstance(other.token, str) else self.token
        if to_compare != other.token:
            return False

        return self.expected_tokens == other.expected_tokens

    def __hash__(self):
        return hash((self.message, self.token, self.expected_tokens))


@dataclass()
class UnexpectedEofParsingError(ParsingError):
    message: str = None

    def __repr__(self):
        return f"UnexpectedEofParsingError({self.message})"


class BaseParser:
    PREFIX = "parsers."

    def __init__(self, name, priority: int, enabled=True, yield_eof=False, hints=None):
        """

        :param name:
        :param priority:
        :param enabled:
        :param yield_eof:
        :param hints: Dictionary context_hint: priority. When not null, priority is taken from it
        """
        # self.log = get_logger("parsers." + self.__class__.__name__)
        # self.init_log = get_logger("init." + self.PREFIX + self.__class__.__name__)
        # self.verbose_log = get_logger("verbose." + self.PREFIX + self.__class__.__name__)

        self.name = BaseParser.get_name(name)
        self.short_name = name
        self.priority = priority
        self.enabled = enabled
        self.yield_eof = yield_eof
        self.hints = hints

    def __eq__(self, other):
        if not isinstance(other, self.__class__):
            return False
        return self.name == other.name

    def __hash__(self):
        return hash(self.name)

    def __repr__(self):
        return self.name

    def log_result(self, context, source, ret):
        pass
        # if not self.log.isEnabledFor(logging.DEBUG):
        #     return
        #
        # if ret.status:
        #     value = context.return_value_to_str(ret)
        #     context.log(f"Recognized '{source}' as {value}", self.name)
        # else:
        #     context.log(f"Failed to recognize '{source}'", self.name)

    def log_multiple_results(self, context, source, list_of_ret):
        pass
        # if not self.log.isEnabledFor(logging.DEBUG):
        #     return
        #
        # context.log(f"Recognized '{source}' as multiple concepts", self.name)
        # for r in list_of_ret:
        #     value = context.return_value_to_str(r)
        #     context.log(f"  Recognized '{value}'", self.name)

    def get_return_value_body(self, sheerka, source, parsed, try_parse, errors):
        """
        All parsers must return their result in a standard way
        :param sheerka:
        :param source:
        :param parsed:
        :param try_parse:
        :param errors:
        :return:
        """
        if len(errors) == 1 and isinstance(errors[0], Concept):
            return errors[0]

        if len(errors):
            if parsed is None:
                return sheerka.new(BuiltinConcepts.NOT_FOR_ME,
                                   body=source,
                                   reason=errors)
            else:
                return sheerka.new(BuiltinConcepts.ERROR,
                                   body=errors)

        return sheerka.new(BuiltinConcepts.PARSER_RESULT,
                           parser=self,
                           source=source,
                           body=parsed,
                           try_parsed=try_parse)

    @staticmethod
    def get_name(name):
        return BaseParser.PREFIX + name


class BaseParserInputParser(BaseParser):
    """
    Base parser for stateful parser where context, parser input, and error sink are part of the class
    """

    def __init__(self, name, priority: int, enabled=True, yield_eof=False):
        super(BaseParserInputParser, self).__init__(name, priority, enabled, yield_eof)

        self.error_sink = []
        self.context: ExecutionContext = None
        self.sheerka = None
        self.parser_input: ParserInput = None
        self.yield_eof = yield_eof

    def reset_parser(self, context, parser_input: ParserInput):
        self.context = context
        self.sheerka = context.sheerka
        self.parser_input = parser_input
        self.error_sink.clear()

        try:
            self.parser_input.reset(self.yield_eof)
        except LexerError as e:
            self.add_error(e, False)
            return False
        return True

    def parse(self, context, parser_input):
        pass

    def add_error(self, error, next_token=True):
        self.error_sink.append(error)
        if next_token:
            self.parser_input.next_token()
        return error

    @property
    def has_error(self):
        return len(self.error_sink) > 0

    @staticmethod
    def get_input_as_lexer_nodes(parser_input, expected_parser=None):
        """
        Extract the lexer node from the parser_input
        :param parser_input:
        :param expected_parser: returns the nodes if the parent parser is the expected one
        :return:
        """
        if not isinstance(parser_input, ParserResultConcept):
            return None

        if expected_parser and parser_input.parser != expected_parser:
            return None

        from parsers.BaseNodeParser import LexerNode
        if isinstance(parser_input.value, list):
            if len(parser_input.value) == 0:
                return None

            for node in parser_input.value:
                if not isinstance(node, LexerNode):
                    return None

            return parser_input.value

        else:
            if not isinstance(parser_input.value, LexerNode):
                return None

            return [parser_input.value]

    @staticmethod
    def get_tokens_boundaries(tokens):
        """
        Returns the first and the last valid index of the tokens
        a valid index is a token that is not a whitespace nor and EOF
        :param tokens:
        :return:
        """
        if tokens is None:
            return None

        if len(tokens) == 0:
            return 0, 0

        if tokens[0].type == TokenKind.EOF:
            return 0, 0

        start = 1 if tokens[0].type == TokenKind.WHITESPACE else 0
        end = len(tokens) - 1
        while tokens[end].type in (TokenKind.WHITESPACE, TokenKind.EOF):
            end -= 1

        return start, end