module.exports = grammar({
    name: 'swh_search_query_language',

    rules: {
        program: $ => repeat(
            choice(
                $.patternFilter,
                $.booleanFilter,
                $.numericFilter,
                $.unboundedListFilter,
                $.boundedListFilter,
                $.dateFilter,
            )
        ),

        patternFilter: $ => seq($.patternField, $.patternOp, $.patternVal),
        patternField: $ => token(choice('url', 'metadata')),
        patternOp: $ => choice(' : ', ' = '),
        patternVal: $ => $.string, // should be REGEX

        booleanFilter: $ => seq($.booleanField, $.booleanOp, $.booleanVal),
        booleanField: $ => choice('with_visit'),
        booleanOp: $ => choice(' = ', ' : '),
        booleanVal: $ => choice($.true, $.false),

        numericFilter: $ => seq($.numericField, $.numericOp, $.numberVal),
        numericField: $ => choice('nb_visits', 'limit'),
        numericOp : $ => $.rangeOp,
        numberVal: $ => $.number,

        boundedListFilter: $ => choice($.visitTypeFilter, $.sortByFilter),

        visitTypeFilter: $ => seq($.visitTypeField, $.visitTypeOp, $.visitTypeVal),
        visitTypeField: $ => 'visit_types',
        visitTypeOp : $ => choice(':'),
        visitTypeVal: $ => seq('[', commaSepStr($.visitTypeOptions), ']'),
        visitTypeOptions: $ => choice(
            "any", "cran", "deb", "deposit", "ftp", "hg", "git", "nixguix", "npm", "pypi", "svn", "tar"
        ),

        sortByFilter: $ => seq($.sortByField, $.sortByOp, $.sortByVal),
        sortByField: $ => 'sort_by',
        sortByOp: $ => choice(':'),
        sortByVal: $ => seq('[', commaSepStr($.sortByOptions), ']'),
        sortByOptions: $ => choice(
            'nb_visits',
            'last_visit_date',
            'last_eventful_visit_date',
            'last_revision_date',
            'last_release_date',
            'date_created',
            'date_modified',
            'date_published'
        ),

        unboundedListFilter: $ => seq($.listField, $.listOp, $.listVal),
        listField: $ => choice('programming_languages', 'licenses', 'keywords'),
        listOp: $ => token(choice('in', 'not in')),
        listVal: $ => $.array,
        

        dateFilter: $ => seq($.dateField, $.rangeOp, $.dateWithOptionalTime),
        dateField: $ => choice(
            'last_visit_date',
            'last_eventful_visit_date',
            'last_revision_date',
            'last_release_date',
            'date_created',
            'date_modified',
            'date_published'
        ),
        dateOp: $ => $.rangeOp,
        dateVal: $ => $.dateWithOptionalTime,


        rangeOp: $ => choice('<', '<=', '=', '!=', '>=', '>'),

        array: $ => seq(
            "[", commaSepStr($.string_content), "]"
        ),

        dateWithOptionalTime: $ => /\d{4}[-]\d{2}[-]\d{2} (\d{2}:\d{2}(:\d{2})*)*/,
        // ^\d{2}\/\d{2}\/\d{4}\s*(?:\d{2}:\d{2}(?::\d{2})?)?$
        // Reference : https://stackoverflow.com/questions/23786905/regex-for-validating-dd-mm-yyyy-with-optional-time
        // Matches :
        // 21/05/2014
        // 21/05/2014 15:54
        // 21/05/2014 15:54:12

        string: $ => wrapWithInvertedComma($.string_content),
        number : $ => /\d+/,
        true: $ => "true",
        false: $ => "false",
        string_content: $ => repeat1(choice(
            token.immediate(/[^\\"\n]+/),
            $.escape_sequence
        )),
        escape_sequence: $ => token.immediate(seq(
            '\\',
            /(\"|\\|\/|b|n|r|t|u)/
        )),

    }
});

function commaSep1(rule) {
    return seq(rule, repeat(seq(",", rule)))
}

function commaSep(rule) {
    return optional(commaSep1(rule))
}

function commaSepStr(rule){
    return commaSep(wrapWithInvertedComma(rule))
}

function wrapWithInvertedComma(rule) {
    return choice(
        seq("'", rule, "'"),
        seq('"', rule, '"'),
        rule,
    )
}