JSON implementation for Ruby

Description

This is a implementation of the JSON specification according to RFC 7159 http://www.ietf.org/rfc/rfc7159.txt . Starting from version 1.0.0 on there will be two variants available:

A pure ruby variant, that relies on the iconv and the stringscan extensions, which are both part of the ruby standard library.
The quite a bit faster native extension variant, which is in parts implemented in C or Java and comes with its own unicode conversion functions and a parser generated by the ragel state machine compiler http://www.complang.org/ragel/ .

Both variants of the JSON generator generate UTF-8 character sequences by default. If an :ascii_only option with a true value is given, they escape all non-ASCII and control characters with \uXXXX escape sequences, and support UTF-16 surrogate pairs in order to be able to generate the whole range of unicode code points.

All strings, that are to be encoded as JSON strings, should be UTF-8 byte sequences on the Ruby side. To encode raw binary strings, that aren't UTF-8 encoded, please use the to_json_raw_object method of String (which produces an object, that contains a byte array) and decode the result on the receiving endpoint.

Installation

It's recommended to use the extension variant of JSON, because it's faster than the pure ruby variant. If you cannot build it on your system, you can settle for the latter.

Just type into the command line as root:

# rake install

The above command will build the extensions and install them on your system.

# rake install_pure

# ruby install.rb

will just install the pure ruby implementation of JSON.

If you use Rubygems you can type

# gem install json

instead, to install the newest JSON version.

There is also a pure ruby json only variant of the gem, that can be installed with:

# gem install json_pure

Compiling the extensions yourself

If you want to create the parser.c file from its parser.rl file or draw nice graphviz images of the state machines, you need ragel from: http://www.complang.org/ragel/

Usage

To use JSON you can

require 'json'

to load the installed variant (either the extension 'json' or the pure variant 'json_pure'). If you have installed the extension variant, you can pick either the extension variant or the pure variant by typing

require 'json/ext'

require 'json/pure'

Now you can parse a JSON document into a ruby data structure by calling

JSON.parse(document)

If you want to generate a JSON document from a ruby data structure call

JSON.generate(data)

You can also use the pretty_generate method (which formats the output more verbosely and nicely) or fast_generate (which doesn't do any of the security checks generate performs, e. g. nesting deepness checks).

There are also the JSON and JSON[] methods which use parse on a String or generate a JSON document from an array or hash:

document = JSON 'test'  => 23 # => "{\"test\":23}"
document = JSON['test' => 23] # => "{\"test\":23}"

and

data = JSON '{"test":23}'  # => {"test"=>23}
data = JSON['{"test":23}'] # => {"test"=>23}

You can choose to load a set of common additions to ruby core's objects if you

require 'json/add/core'

After requiring this you can, e. g., serialise/deserialise Ruby ranges:

JSON JSON(1..10) # => 1..10

To find out how to add JSON support to other or your own classes, read the section "More Examples" below.

To get the best compatibility to rails' JSON implementation, you can

require 'json/add/rails'

Both of the additions attempt to require 'json' (like above) first, if it has not been required yet.

More Examples

To create a JSON document from a ruby data structure, you can call JSON.generate like that:

json = JSON.generate [1, 2, {"a"=>3.141}, false, true, nil, 4..10]
# => "[1,2,{\"a\":3.141},false,true,null,\"4..10\"]"

To get back a ruby data structure from a JSON document, you have to call JSON.parse on it:

JSON.parse json
# => [1, 2, {"a"=>3.141}, false, true, nil, "4..10"]

Note, that the range from the original data structure is a simple string now. The reason for this is, that JSON doesn't support ranges or arbitrary classes. In this case the json library falls back to call Object#to_json, which is the same as #to_s.to_json.

It's possible to add JSON support serialization to arbitrary classes by simply implementing a more specialized version of the #to_json method, that should return a JSON object (a hash converted to JSON with #to_json) like this (don't forget the *a for all the arguments):

 class Range
   def to_json(*a)
     {
       'json_class'   => self.class.name, # = 'Range'
       'data'         => [ first, last, exclude_end? ]
     }.to_json(*a)
   end
 end

The hash key json_class is the class, that will be asked to deserialise the JSON representation later. In this case it's Range, but any namespace of the form A::B or ::A::B will do. All other keys are arbitrary and can be used to store the necessary data to configure the object to be deserialised.

If a the key json_class is found in a JSON object, the JSON parser checks if the given class responds to the json_create class method. If so, it is called with the JSON object converted to a Ruby hash. So a range can be deserialised by implementing Range.json_create like this:

 class Range
   def self.json_create(o)
     new(*o['data'])
   end
 end

Now it possible to serialise/deserialise ranges as well:

 json = JSON.generate [1, 2, {"a"=>3.141}, false, true, nil, 4..10]
 # => "[1,2,{\"a\":3.141},false,true,null,{\"json_class\":\"Range\",\"data\":[4,10,false]}]"
 JSON.parse json
 # => [1, 2, {"a"=>3.141}, false, true, nil, 4..10]

JSON.generate always creates the shortest possible string representation of a ruby data structure in one line. This is good for data storage or network protocols, but not so good for humans to read. Fortunately there's also JSON.pretty_generate (or JSON.pretty_generate) that creates a more readable output:

 puts JSON.pretty_generate([1, 2, {"a"=>3.141}, false, true, nil, 4..10])
 [
   1,
   2,
   {
     "a": 3.141
   },
   false,
   true,
   null,
   {
     "json_class": "Range",
     "data": [
       4,
       10,
       false
     ]
   }
 ]

There are also the methods Kernel#j for generate, and Kernel#jj for pretty_generate output to the console, that work analogous to Core Ruby's p and the pp library's pp methods.

The script tools/server.rb contains a small example if you want to test, how receiving a JSON object from a webrick server in your browser with the javasript prototype library http://www.prototypejs.org works.

Speed Comparisons

I have created some benchmark results (see the benchmarks/data-p4-3Ghz subdir of the package) for the JSON-parser to estimate the speed up in the C extension:

 Comparing times (call_time_mean):
  1 ParserBenchmarkExt#parser   900 repeats:
        553.922304770 (  real) ->   21.500x
          0.001805307
  2 ParserBenchmarkYAML#parser  1000 repeats:
        224.513358139 (  real) ->    8.714x
          0.004454078
  3 ParserBenchmarkPure#parser  1000 repeats:
         26.755020642 (  real) ->    1.038x
          0.037376163
  4 ParserBenchmarkRails#parser 1000 repeats:
         25.763381731 (  real) ->    1.000x
          0.038814780
            calls/sec (  time) ->    speed  covers
            secs/call

In the table above 1 is JSON::Ext::Parser, 2 is YAML.load with YAML compatbile JSON document, 3 is is JSON::Pure::Parser, and 4 is ActiveSupport::JSON.decode. The ActiveSupport JSON-decoder converts the input first to YAML and then uses the YAML-parser, the conversion seems to slow it down so much that it is only as fast as the JSON::Pure::Parser!

If you look at the benchmark data you can see that this is mostly caused by the frequent high outliers - the median of the Rails-parser runs is still overall smaller than the median of the JSON::Pure::Parser runs:

 Comparing times (call_time_median):
  1 ParserBenchmarkExt#parser   900 repeats:
        800.592479481 (  real) ->   26.936x
          0.001249075
  2 ParserBenchmarkYAML#parser  1000 repeats:
        271.002390644 (  real) ->    9.118x
          0.003690004
  3 ParserBenchmarkRails#parser 1000 repeats:
         30.227910865 (  real) ->    1.017x
          0.033082008
  4 ParserBenchmarkPure#parser  1000 repeats:
         29.722384421 (  real) ->    1.000x
          0.033644676
            calls/sec (  time) ->    speed  covers
            secs/call

I have benchmarked the JSON-Generator as well. This generated a few more values, because there are different modes that also influence the achieved speed:

 Comparing times (call_time_mean):
  1 GeneratorBenchmarkExt#generator_fast    1000 repeats:
        547.354332608 (  real) ->   15.090x
          0.001826970
  2 GeneratorBenchmarkExt#generator_safe    1000 repeats:
        443.968212317 (  real) ->   12.240x
          0.002252414
  3 GeneratorBenchmarkExt#generator_pretty  900 repeats:
        375.104545883 (  real) ->   10.341x
          0.002665923
  4 GeneratorBenchmarkPure#generator_fast   1000 repeats:
         49.978706968 (  real) ->    1.378x
          0.020008521
  5 GeneratorBenchmarkRails#generator       1000 repeats:
         38.531868759 (  real) ->    1.062x
          0.025952543
  6 GeneratorBenchmarkPure#generator_safe   1000 repeats:
         36.927649925 (  real) ->    1.018x 7 (>=3859)
          0.027079979
  7 GeneratorBenchmarkPure#generator_pretty 1000 repeats:
         36.272134441 (  real) ->    1.000x 6 (>=3859)
          0.027569373
            calls/sec (  time) ->    speed  covers
            secs/call

In the table above 1-3 are JSON::Ext::Generator methods. 4, 6, and 7 are JSON::Pure::Generator methods and 5 is the Rails JSON generator. It is now a bit faster than the generator_safe and generator_pretty methods of the pure variant but slower than the others.

To achieve the fastest JSON document output, you can use the fast_generate method. Beware, that this will disable the checking for circular Ruby data structures, which may cause JSON to go into an infinite loop.

Here are the median comparisons for completeness' sake:

 Comparing times (call_time_median):
  1 GeneratorBenchmarkExt#generator_fast    1000 repeats:
        708.258020939 (  real) ->   16.547x
          0.001411915
  2 GeneratorBenchmarkExt#generator_safe    1000 repeats:
        569.105020353 (  real) ->   13.296x
          0.001757145
  3 GeneratorBenchmarkExt#generator_pretty  900 repeats:
        482.825371244 (  real) ->   11.280x
          0.002071142
  4 GeneratorBenchmarkPure#generator_fast   1000 repeats:
         62.717626652 (  real) ->    1.465x
          0.015944481
  5 GeneratorBenchmarkRails#generator       1000 repeats:
         43.965681162 (  real) ->    1.027x
          0.022745013
  6 GeneratorBenchmarkPure#generator_safe   1000 repeats:
         43.929073409 (  real) ->    1.026x 7 (>=3859)
          0.022763968
  7 GeneratorBenchmarkPure#generator_pretty 1000 repeats:
         42.802514491 (  real) ->    1.000x 6 (>=3859)
          0.023363113
            calls/sec (  time) ->    speed  covers
            secs/call