shimatta-kenkyusho/shimatta_kenkyusho/shimatta_modules/ShimattaSearchLanguage.py

import re
from django.db.models import Q
from sly import Lexer, Parser
from shimatta_modules.EngineeringNumberConverter import EngineeringNumberConverter

class ShimattaSearchConstants():
    '''
    Just a bunch of conversions and regular expression stored here.
    '''

    # convert the prefixes from the engineering number converter be used in the lexer
    PREFIX_DICT = {k: v for k, v in EngineeringNumberConverter.prefixes if k}
    PREFIX_DICT.update({k: v for k, v in EngineeringNumberConverter.it_prefixes if k})
    PREFIX_RULE = r'(' + r'|'.join([rf'(?:{p})' for p in PREFIX_DICT.keys()]) + r')?'

    TEXTUAL_REGEX = r'(?:([^"\s]+))|(?:"([^"]*)")'
    VALUE_REGEX = rf'(\-?\d+(?:\.\d+)?){PREFIX_RULE}'

    TEXTUAL_PATTERN = re.compile(TEXTUAL_REGEX)
    VALUE_PATTERN = re.compile(VALUE_REGEX)

class ShimattaSearchLexer(Lexer):
    '''
    Stupid lexer to tokenize a search string.
    '''

    tokens = {GT, LT, GTE, LTE, EQ, NEQ, AND, OR, LPAREN, RPAREN, NUMBER, TEXTUAL}

    # ignore whitespace only characters as well as newlines
    ignore = ' \t\n'

    # Regular expression rules for simple tokens
    GT        = r'>'
    LT        = r'<'
    GTE       = r'>='
    LTE       = r'<='
    EQ        = r'=='
    NEQ       = r'!='
    AND       = r'&'
    OR        = r'\|'
    LPAREN    = r'\('
    RPAREN    = r'\)'

    def __init__(self):
        self.errors = []
        super().__init__()

    @_(ShimattaSearchConstants.VALUE_REGEX)
    def NUMBER(self, t):
        '''
        Parse numbers with engineering unit prefixes
        '''
        match = ShimattaSearchConstants.VALUE_PATTERN.match(t.value)
        t.value = float(match.group(1))
        prefix = match.group(2)
        if prefix:
            t.value *= ShimattaSearchConstants.PREFIX_DICT[prefix]
        return t

    @_(ShimattaSearchConstants.TEXTUAL_REGEX)
    def TEXTUAL(self, t):
        '''
        Find texts with or without param_ prefix (used to filter for parameters)
        '''
        match = ShimattaSearchConstants.TEXTUAL_PATTERN.match(t.value)

        # strip the quotation marks
        value = match.group(1)
        if match.group(2):
            value = match.group(2)

        t.value = value
        return t

    def error(self, t):
        self.errors.append(f'Line {self.lineno}: Bad character {t.value}')
        self.index += 1

class ShimattaSearchParser(Parser):
    # Get the token list from the lexer (required)
    tokens = ShimattaSearchLexer.tokens

    def __init__(self):
        self.errors = []
        super().__init__()

    @staticmethod
    def _get_filter(key, value, compare_suffix='', invert=False):
        '''
        Assemble a filter to grep data from the relational database structure
        '''

        # filter for params - stored in two separate tables
        if key.startswith('param_'):
            key = key[len('param_'):]

            key_query = Q(Q(**{f'componentparameter__parameter_type__parameter_name': key})| \
                        Q(**{f'package__packageparameter__parameter_type__parameter_name': key}))

            if isinstance(value, str):
                query = Q(Q(**{f'componentparameter__text_value{compare_suffix}': value})| \
                        Q(**{f'package__packageparameter__text_value{compare_suffix}': value}))&key_query
            else:
                query = Q(Q(**{f'componentparameter__value{compare_suffix}': value})| \
                        Q(**{f'package__packageparameter__value{compare_suffix}': value}))&key_query
        # filter for direct attributes - or whatever the user throws into the search input
        else:
            query = Q(**{f'{key}{compare_suffix}': value})

        if invert:
            query = ~query
        return Q(query)

    # ruleset

    @_('expression : textual GT number')
    def expression(self, p):
        return self._get_filter(p.textual.strip(), p.number, '__gt', False)

    @_('expression : textual LT number')
    def expression(self, p):
        return self._get_filter(p.textual.strip(), p.number, '__lt', False)

    @_('expression : textual GTE number')
    def expression(self, p):
        return self._get_filter(p.textual.strip(), p.number, '__gte', False)

    @_('expression : textual LTE number')
    def expression(self, p):
        return self._get_filter(p.textual.strip(), p.number, '__lte', False)

    @_('expression : textual EQ number')
    def expression(self, p):
        return self._get_filter(p.textual.strip(), p.number, '', False)

    @_('expression : textual NEQ number')
    def expression(self, p):
        return self._get_filter(p.textual.strip(), p.number, '', True)

    @_('expression : textual EQ textual')
    def expression(self, p):
        return self._get_filter(p.textual0.strip(), p.textual1, '', False)

    @_('expression : textual NEQ textual')
    def expression(self, p):
        return self._get_filter(p.textual0.strip(), p.textual1, '', True)

    @_('TEXTUAL')
    def textual(self, p):
        return p.TEXTUAL

    @_('NUMBER')
    def number(self, p):
        return p.NUMBER

    @_('expression : LPAREN expression RPAREN')
    def expression(self, p):
        return Q(p.expression)

    @_('expression : expression AND expression')
    def expression(self, p):
        return p.expression0&p.expression1

    @_('expression : expression OR expression')
    def expression(self, p):
        return p.expression0|p.expression1

    @_('expression')
    def expression(self, p):
        return p.expression

    # Error rule for syntax errors
    def error(self, p):
        self.errors.append(f'Syntax error in input {p}!')

class ShimattaSearchLanguage():

    def __init__(self):
        self.lexer = ShimattaSearchLexer()
        self.parser = ShimattaSearchParser()

    def search_for_components(self, search_string):
        query = self.parser.parse(self.lexer.tokenize(search_string))
        return query, self.lexer.errors + self.parser.errors