Parses Lucene-like queries (mostly). This is ultimately only an approximation of correctness; a true grammar would get much closer.

This project includes an HTTP API, with a single endpoint, /parse?query=..., usage is as follows:

curl 'localhost:8081/parse?query=foo%20bar'
{"$and":["foo","bar"]}

curl --get 'localhost:8081/parse' --data-urlencode 'query=(foo OR bar) AND (baz OR >300)'
{"$and":[{"$or":["foo","bar"]},{"$or":["baz",{"$gt":"300"}]}]}

Supports standard operators: >=, <=, >, <, =, !.

Terms can be prepended with an operator, e.g. <400 or !="foo", and can be grouped with parentheses, e.g. (foo AND bar)

The output of the parser is a JSON format, and a few examples are:

// error OR info
{
    "$or": [
        "error",
        "info"
    ]
}

// foo OR bar AND baz
{
    "$or": [
        "foo",
        {
            "$and": [
                "bar",
                "baz"
            ]
        }
    ]
}

// !="foo"
{
    "$not": {
        "$eq": {
            "$quoted": "foo"
        }
    }
}