parse simple search expression to sqlalchemy-filters

gistfile1.txt

#!/usr/bin/env python3
  
import logging
from pprint import pprint
from lark import Lark, Transformer


"""
Boolean expression definition widely recognizes basic building blocks as
'terms' and 'factors', do not confuse them with parser's terminals. Also note
that EBNF expresses logic operators precedence.

https://unnikked.ga/how-to-build-a-boolean-expression-evaluator-518e9e068a65
https://cs.stackexchange.com/questions/10558/grammar-for-describing-boolean-expressions-with-and-or-and-not/10622#10622
"""
search_grammar = r"""
        ?start: expression

        expression: term ("OR" term)*
        term: _factor ("AND" _factor)*
        _factor: criteria | "(" expression ")"

        criteria: FIELD OP VALUE
        FIELD: /[a-z]+/
        OP: "==" | "!="
        VALUE: ESCAPED_STRING

        %import common.ESCAPED_STRING
        %import common.WS
        %ignore WS
"""

class TreeToSAFilter(Transformer):
        def expression(self, args):
                if len(args) > 1:
                        return {"or": args}
                else:
                        return args[0]

        def term(self, args):
                if len(args) > 1:
                        return {"and": args}
                else:
                        return args[0]

        def criteria(self, args):
                return {'field': args[0].value, 'op': args[1].value, 'value': args[2].value}


search_parser = Lark(search_grammar, parser='lalr', lexer='standard', transformer=TreeToSAFilter())


def test(testcase, expected):
        output = search_parser.parse(testcase)
        print('testcase: %s outputs %s' % (testcase, output))
        assert output == expected


if __name__ == '__main__':

        test('a=="a"', {'field': 'a', 'op': '==', 'value': '"a"'})
        test('a=="a" AND b=="b"', {'and': [{'field': 'a', 'op': '==', 'value': '"a"'}, {'field': 'b', 'op': '==', 'value': '"b"'}]})
        test('a=="a" AND b=="b" AND c=="c"', {'and': [{'field': 'a', 'op': '==', 'value': '"a"'}, {'field': 'b', 'op': '==', 'value': '"b"'}, {'field': 'c', 'op': '==', 'value': '"c"'}]})

        test('a=="a" OR b=="b"', {'or': [{'field': 'a', 'op': '==', 'value': '"a"'}, {'field': 'b', 'op': '==', 'value': '"b"'}]})
        test('a=="a" OR b=="b" OR c=="c"', {'or': [{'field': 'a', 'op': '==', 'value': '"a"'}, {'field': 'b', 'op': '==', 'value': '"b"'}, {'field': 'c', 'op': '==', 'value': '"c"'}]})

        test('a=="a" OR b=="b" AND c!="c"', {'or': [{'field': 'a', 'op': '==', 'value': '"a"'}, {'and': [{'field': 'b', 'op': '==', 'value': '"b"'}, {'field': 'c', 'op': '!=', 'value': '"c"'}]}]})
        test('a=="a" OR b=="b" OR c!="c" AND d=="d"', {'or': [{'field': 'a', 'op': '==', 'value': '"a"'}, {'field': 'b', 'op': '==', 'value': '"b"'}, {'and': [{'field': 'c', 'op': '!=', 'value': '"c"'}, {'field': 'd', 'op': '==', 'value': '"d"'}]}]})