SDJSONPrettyPrint - The Slow JSON Serializer

SDJSONPrettyPrint transforms a compatible Foundation-based object tree into a JSON string. It is not designed to be a fast, production-ready JSON serializer, for that you probably want JSONKit. SDJSONPrettyPrint focuses on producing the most human-friendly JSON possible, which you can use for logging or debugging purposes.

Purpose

There exist multitudes of JSON serializers for Objective-C which - including a very good built-in one that produces perfectly valid JSON strings, primarily for the consumption of machines. Most of these solutions ignore - or have otherwise limited solutions for - producing output designed to be read by humans. This is an important feature for logging since in the production of a large scale system with a complex API, during the course of development many thousands of API requests and responses will be made. Many of these will be checked by hand by the programmer to ensure the application is interacting with the API correctly. This project aims to produce JSON output that's considerably easier to read, is formatted in a pleasing way and presents the data in a way that's least likely to frustrate.

Example

Consider the following JSON sample, formatted for machines:

{"function":null,"numbers":[4,8,15,16,23,42],"y_index":2,"x_index":12,"z_index":5,"arcs":[{"p2":[22.1,50],"p1":[10.5,15.5],"radius":5},{"p2":[23.1,40],"p1":[11.5,15.5],"radius":10},{"p2":[23.1,30],"p1":[12.5,15.5],"radius":3},{"p2":[24.1,20],"p1":[13.5,15.5],"radius":2},{"p2":[25.1,10],"p1":[14.5,15.5],"radius":8},{"p2":[26.1,0],"p1":[15.5,15.5],"radius":2}],"label":"my label"}

If you run this through NSJSONSerialization with NSJSONWritingPrettyPrinted enabled, it produces:

{
  "function" : null,
  "numbers" : [
    4,
    8,
    15,
    16,
    23,
    42
  ],
  "y_index" : 2,
  "x_index" : 12,
  "z_index" : 5,
  "arcs" : [
    {
      "p2" : [
        22.1,
        50
      ],
      "p1" : [
        10.5,
        15.5
      ],
      "radius" : 5
    },
    {
      "p2" : [
        23.1,
        40
      ],
      "p1" : [
        11.5,
        15.5
      ],
      "radius" : 10
    },
    {
      "p2" : [
        23.1,
        30
      ],
      "p1" : [
        12.5,
        15.5
      ],
      "radius" : 3
    },
    {
      "p2" : [
        24.1,
        20
      ],
      "p1" : [
        13.5,
        15.5
      ],
      "radius" : 2
    },
    {
      "p2" : [
        25.1,
        10
      ],
      "p1" : [
        14.5,
        15.5
      ],
      "radius" : 8
    },
    {
      "p2" : [
        26.1,
        0
      ],
      "p1" : [
        15.5,
        15.5
      ],
      "radius" : 2
    }
  ],
  "label" : "my label"
}

This is slightly better, but really all it did was add a heap of whitespace to fill up vertical space. It indisciminately adds newlines for every element in an array, and the order in which it prints out keys in JSON objects seems fairly arbitrary. The label key is - surprisingly - underneath the giant array of objects arcs which means it could very easily be missed.

JSONKit is a highly respected JSON parsing and serializing library and is used in production by many apps. Similarly to NSJSONSerialization, it has a pretty printing mode. When this example JSON is run through it, it produces:

{
  "arcs": [
    {
      "p1": [
        10.5,
        15.5
      ],
      "p2": [
        22.100000000000001,
        50
      ],
      "radius": 5
    },
    {
      "p1": [
        11.5,
        15.5
      ],
      "p2": [
        23.100000000000001,
        40
      ],
      "radius": 10
    },
    {
      "p1": [
        12.5,
        15.5
      ],
      "p2": [
        23.100000000000001,
        30
      ],
      "radius": 3
    },
    {
      "p1": [
        13.5,
        15.5
      ],
      "p2": [
        24.100000000000001,
        20
      ],
      "radius": 2
    },
    {
      "p1": [
        14.5,
        15.5
      ],
      "p2": [
        25.100000000000001,
        10
      ],
      "radius": 8
    },
    {
      "p1": [
        15.5,
        15.5
      ],
      "p2": [
        26.100000000000001,
        0
      ],
      "radius": 2
    }
  ],
  "function": null,
  "label": "my label",
  "numbers": [
    4,
    8,
    15,
    16,
    23,
    42
  ],
  "x_index": 12,
  "y_index": 2,
  "z_index": 5
}

This is slightly better since it at least orders the keys in a predictable - if slightly useless - way: alphabetically. It still has the problem of indiscriminate newlines and placing large values which have large JSON structures before simple values, making it easy to overlook data.

The following is the output from SDJSONPrettyPrint.

{
  "label": "my label",
  "x_index": 12,
  "y_index": 2,
  "z_index": 5,
  "numbers": [4, 8, 15, 16, 23, 42],
  "function": null,
  "arcs": [
    {
      "radius": 5,
      "p1": [10.5, 15.5],
      "p2": [22.1, 50]
    },
    {
      "radius": 10,
      "p1": [11.5, 15.5],
      "p2": [23.1, 40]
    },
    {
      "radius": 3,
      "p1": [12.5, 15.5],
      "p2": [23.1, 30]
    },
    {
      "radius": 2,
      "p1": [13.5, 15.5],
      "p2": [24.1, 20]
    },
    {
      "radius": 8,
      "p1": [14.5, 15.5],
      "p2": [25.1, 10]
    },
    {
      "radius": 2,
      "p1": [15.5, 15.5],
      "p2": [26.1, 0]
    }
  ]
}

Performance

"The Slow JSON Serializer", just how slow is it anyway?

Very slow.

What SDJSONPrettyPrint is

• A JSON string serializer that focuses on human-readable output.
• Good for debugging and logging.
• A JSON visualisation mechanism.

What SDJSONPrettyPrint is not

• A production-ready JSON serializer. It should probably be disabled in your production builds, but since it contains no private APIs, it's perfectly App-Store safe.
• Fast, reliable or strict.
• A JSON validator. You're much better off using NSJSONSerialization for this. In fact, SDJSONPrettyPrint even uses NSJSONSerialization internally to sanity check the JSON object, so you're absolutely no better off using this class for that purpose.
• A JSON parser. Again, there are plenty of fantastic JSON parsers out there, and I certainly couldn't do a better job than they already do.

Requirements

• Automatic Reference Counting.
• Mac OS X >= 10.7 or iOS >= 5.0. It will probably run under older versions if you install AnyJSON, but I haven't tested this.
• Optional: CocoaPods.

Installation

1. Recommended: CocoaPods.

   pod 'SDJSONPrettyPrint'
   

2. Manually. Copy SDJSONPrettyPrint.h and SDJSONPrettyPrint.m files into your project. It is unintrusive, requires no frameworks other than Foundation and it implements no categories.

License

SDJSONPrettyPrint is licensed under the permissive MIT License.

Copyright (c) 2013, Tyrone Trevorrow

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.