YYJSON.jl

YYJSON is a convenient wrapper around yyjson library for reading and parsing JSON documents and provides its own parser implementation.

Feature	Description
Performance	Able to read gigabytes of JSON document per second.
Flexibility	The library wraps a lot of methods of the original library, which allows you to make your own implementation of JSON reader/parser/(de)serializer.
Parser	Provides its own parser implementation using yyjson tools.

Installation

To install YYJSON, simply use the Julia package manager:

] add YYJSON

Usage

A small example of parsing the returned result from a ticker request:

using YYJSON

json = """
{
  "retCode":0,
  "retMsg":"OK",
  "result":{
    "ap":0.6636,
    "bp":0.6634,
    "h":0.6687,
    "l":0.6315,
    "lp":0.6633,
    "o":0.6337,
    "qv":1.1594252877069e7,
    "s":"ADAUSDT",
    "t":"2024-03-25T19:05:35.491000064",
    "v":1.780835204e7
  },
  "retExtInfo":{},
  "time":"2024-03-25T19:05:38.912999936"
}
"""

julia> parse_json(json)
Dict{String, Any} with 5 entries:
  "retExtInfo" => Dict{String, Any}()
  "time"       => "2024-03-25T19:05:38.912999936"
  "retCode"    => 0
  "retMsg"     => "OK"
  "result"     => Dict{String, Any}("v"=>1.78084e7, "ap"=>0.6636, "o"=>0.6337, ...)

Lazy parsing enables more efficient value retrieval compared to regular parsing:

using YYJSON

json = read("assets/exchange_info.json", String)

function test_full_parse(json)
    parsed = parse_json(json)
    return parsed["symbols"][1]["filters"][1]["filterType"]
end

function test_lazy_parse(json)
    parse_lazy_json(json) do lazy_parsed
        return lazy_parsed["symbols"][1]["filters"][1]["filterType"]
    end
end

julia> @time test_full_parse(json)
  0.000245 seconds (2.89 k allocations: 203.727 KiB)
"PRICE_FILTER"

julia> @time test_lazy_parse(json)
  0.000041 seconds (10 allocations: 448 bytes)
"PRICE_FILTER"

Contributing

Contributions to YYJSON are welcome! If you encounter a bug, have a feature request, or would like to contribute code, please open an issue or a pull request on GitHub.