NAYA (Not Another Yet Another)

A fast streaming JSON parser for Python

NAYA is designed to parse JSON quickly and efficiently in pure Python 3 with no dependencies.

NAYA is different from other JSON parsers in that it can be used to stream a JSON array, even if the entire array is not yet available.

Usage

stream_array

Given the following JSON which will be coming through a file like object:

[
{
    "id": 1,
    "type": "message",
    "content": "Hello!"
},
{
    "id": 2,
    "type": "query",
    "user_name": "tarzan",
    "password": "not_jane"
},
{
    "id": 3,
    "type": "command",
    "action": "swing"
}
]

Even if each message will be sent one at a time on the file like object (called fp)

messages = stream_array(tokenize(fp))
for message in messages:
    handle_message(message)

That is, stream_array returns an iterator, which will block until the next array item is available.

handle_message will be called three times with the following parameters:

handle_message({
    "id": 1,
    "type": "message",
    "content": "Hello!"
})
handle_message({
    "id": 2,
    "type": "query",
    "user_name": "tarzan",
    "password": "not_jane"
})
handle_message({
    "id": 3,
    "type": "command",
    "action": "swing"
})

parse

In addition to streaming array, NAYA also supports standard parsing, like the built in library:

obj = parse(fp)

This method will block until the entire object is read. There is also the convenience method

obj = parse_string('{"key": "value"}')

Related Projects

Yajl-Py

Yajl-Py is a wrapper around the Yajl JSON library that can be used to generate SAX style events while parsing JSON. It also has a pure python implementation.

UltraJSON

UltraJSON is a super fast JSON Parser written in C with python bindings.

danielyule/naya