Module: Oj

Defined in:
lib/oj.rb,
lib/oj/bag.rb,
lib/oj/saj.rb,
lib/oj/error.rb,
lib/oj/mimic.rb,
lib/oj/version.rb,
lib/oj/easy_hash.rb,
lib/oj/schandler.rb,
lib/oj/active_support_helper.rb,
ext/oj/oj.c

Overview

Optimized JSON (Oj), as the name implies was written to provide speed optimized JSON handling.

Oj has several dump or serialization modes which control how Objects are converted to JSON. These modes are set with the :mode option in either the default options or as one of the options to the dump() method.

  • :strict mode will only allow the 7 basic JSON types to be serialized. Any other Object will raise and Exception.

  • :null mode replaces any Object that is not one of the JSON types is replaced by a JSON null.

  • :object mode will dump any Object as a JSON Object with keys that match the Ruby Object’s variable names without the ‘@’ character. This is the highest performance mode.

  • :compat mode is is the compatible with other systems. It will serialize any Object but will check to see if the Object implements a to_hash() or to_json() method. If either exists that method is used for serializing the Object. The to_hash() is more flexible and produces more consistent output so it has a preference over the to_json() method. If neither the to_json() or to_hash() methods exist then the Oj internal Object variable encoding is used.

Defined Under Namespace

Classes: ActiveSupportHelper, Bag, CStack, DepthError, Doc, EasyHash, Error, LoadError, MimicDumpOption, MimicError, ParseError, Saj, ScHandler, StreamWriter, StringWriter

Constant Summary collapse

VERSION =

Current version of the module.

'2.17.5'

Class Method Summary collapse

Class Method Details

.compat_load(json, options) ⇒ Object, ...

Parses a JSON document String into an Object, Hash, Array, String, Fixnum, Float, true, false, or nil. It parses using a mode that is generally compatible with other Ruby JSON parsers in that it will create objects based on the :create_id value. It is not compatible in every way to every other parser though as each parser has it’s own variations.

When used with a document that has multiple JSON elements the block, if any, will be yielded to. If no block then the last element read will be returned.

Raises an exception if the JSON is malformed or the classes specified are not valid. If the input is not a valid JSON document (an empty string is not a valid JSON document) an exception is raised.

A block can also be provided with a single argument. That argument will be the parsed JSON document. This is useful when parsing a string that includes multiple JSON documents.

Returns:

  • (Object, Hash, Array, String, Fixnum, Float, true, false, nil)

Parameters:

  • json (String|IO) —

    JSON String or an Object that responds to read()

  • options (Hash) —

    load options (same as default_options)

.default_options ⇒ Hash

Returns the default load and dump options as a Hash. The options are

  • indent: [Fixnum|String|nil] number of spaces to indent each element in an JSON document, zero or nil is no newline between JSON elements, negative indicates no newline between top level JSON elements in a stream, a String indicates the string should be used for indentation

  • circular: [true|false|nil] support circular references while dumping

  • auto_define: [true|false|nil] automatically define classes if they do not exist

  • symbol_keys: [true|false|nil] use symbols instead of strings for hash keys

  • escape_mode: [:newline|:json|:xss_safe|:ascii|nil] determines the characters to escape

  • class_cache: [true|false|nil] cache classes for faster parsing (if dynamically modifying classes or reloading classes then don’t use this)

  • mode: [:object|:strict|:compat|:null] load and dump modes to use for JSON

  • time_format: [:unix|:unix_zone|:xmlschema|:ruby] time format when dumping in :compat and :object mode

  • bigdecimal_as_decimal: [true|false|nil] dump BigDecimal as a decimal number or as a String

  • bigdecimal_load: [:bigdecimal|:float|:auto] load decimals as BigDecimal instead of as a Float. :auto pick the most precise for the number of digits.

  • create_id: [String|nil] create id for json compatible object encoding, default is ‘json_create’

  • second_precision: [Fixnum|nil] number of digits after the decimal when dumping the seconds portion of time

  • float_precision: [Fixnum|nil] number of digits of precision when dumping floats, 0 indicates use Ruby

  • use_to_json: [true|false|nil] call to_json() methods on dump, default is false

  • nilnil: [true|false|nil] if true a nil input to load will return nil and not raise an Exception

  • allow_gc: [true|false|nil] allow or prohibit GC during parsing, default is true (allow)

  • quirks_mode: [true,|false|nil] Allow single JSON values instead of documents, default is true (allow)

  • allow_invalid_unicode: [true,|false|nil] Allow invalid unicode, default is false (don’t allow)

  • indent_str: [String|nil] String to use for indentation, overriding the indent option is not nil

  • space: [String|nil] String to use for the space after the colon in JSON object fields

  • space_before: [String|nil] String to use before the colon separator in JSON object fields

  • object_nl: [String|nil] String to use after a JSON object field value

  • array_nl: [String|nil] String to use after a JSON array value

  • nan: [:null|:huge|:word|:raise|:auto] how to dump Infinity and NaN in null, strict, and compat mode. :null places a null, :huge places a huge number, :word places Infinity or NaN, :raise raises and exception, :auto uses default for each mode.

  • hash_class: [Class|nil] Class to use instead of Hash on load

  • omit_nil: [true|false] if true Hash and Object attributes with nil values are omitted

Returns:

  • (Hash)

Returns:

  • (Hash) —

    all current option settings.



244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
 
# File 'ext/oj/oj.c', line 244

static VALUE
get_def_opts(VALUE self) {
    VALUE	opts = rb_hash_new();

    if (0 == oj_default_options.dump_opts.indent_size) {
	rb_hash_aset(opts, indent_sym, INT2FIX(oj_default_options.indent));
    } else {
	rb_hash_aset(opts, indent_sym, rb_str_new2(oj_default_options.dump_opts.indent_str));
    }
    rb_hash_aset(opts, sec_prec_sym, INT2FIX(oj_default_options.sec_prec));
    rb_hash_aset(opts, circular_sym, (Yes == oj_default_options.circular) ? Qtrue : ((No == oj_default_options.circular) ? Qfalse : Qnil));
    rb_hash_aset(opts, class_cache_sym, (Yes == oj_default_options.class_cache) ? Qtrue : ((No == oj_default_options.class_cache) ? Qfalse : Qnil));
    rb_hash_aset(opts, auto_define_sym, (Yes == oj_default_options.auto_define) ? Qtrue : ((No == oj_default_options.auto_define) ? Qfalse : Qnil));
    rb_hash_aset(opts, symbol_keys_sym, (Yes == oj_default_options.sym_key) ? Qtrue : ((No == oj_default_options.sym_key) ? Qfalse : Qnil));
    rb_hash_aset(opts, bigdecimal_as_decimal_sym, (Yes == oj_default_options.bigdec_as_num) ? Qtrue : ((No == oj_default_options.bigdec_as_num) ? Qfalse : Qnil));
    rb_hash_aset(opts, use_to_json_sym, (Yes == oj_default_options.to_json) ? Qtrue : ((No == oj_default_options.to_json) ? Qfalse : Qnil));
    rb_hash_aset(opts, nilnil_sym, (Yes == oj_default_options.nilnil) ? Qtrue : ((No == oj_default_options.nilnil) ? Qfalse : Qnil));
    rb_hash_aset(opts, allow_gc_sym, (Yes == oj_default_options.allow_gc) ? Qtrue : ((No == oj_default_options.allow_gc) ? Qfalse : Qnil));
    rb_hash_aset(opts, quirks_mode_sym, (Yes == oj_default_options.quirks_mode) ? Qtrue : ((No == oj_default_options.quirks_mode) ? Qfalse : Qnil));
    rb_hash_aset(opts, allow_invalid_unicode_sym, (Yes == oj_default_options.allow_invalid) ? Qtrue : ((No == oj_default_options.allow_invalid) ? Qfalse : Qnil));
    rb_hash_aset(opts, float_prec_sym, INT2FIX(oj_default_options.float_prec));
    switch (oj_default_options.mode) {
    case StrictMode:	rb_hash_aset(opts, mode_sym, strict_sym);	break;
    case CompatMode:	rb_hash_aset(opts, mode_sym, compat_sym);	break;
    case NullMode:	rb_hash_aset(opts, mode_sym, null_sym);		break;
    case ObjectMode:
    default:		rb_hash_aset(opts, mode_sym, object_sym);	break;
    }
    switch (oj_default_options.escape_mode) {
    case NLEsc:		rb_hash_aset(opts, escape_mode_sym, newline_sym);	break;
    case JSONEsc:	rb_hash_aset(opts, escape_mode_sym, json_sym);		break;
    case XSSEsc:	rb_hash_aset(opts, escape_mode_sym, xss_safe_sym);	break;
    case ASCIIEsc:	rb_hash_aset(opts, escape_mode_sym, ascii_sym);		break;
    default:		rb_hash_aset(opts, escape_mode_sym, json_sym);		break;
    }
    switch (oj_default_options.time_format) {
    case XmlTime:	rb_hash_aset(opts, time_format_sym, xmlschema_sym);	break;
    case RubyTime:	rb_hash_aset(opts, time_format_sym, ruby_sym);		break;
    case UnixZTime:	rb_hash_aset(opts, time_format_sym, unix_zone_sym);	break;
    case UnixTime:
    default:		rb_hash_aset(opts, time_format_sym, unix_sym);		break;
    }
    switch (oj_default_options.bigdec_load) {
    case BigDec:	rb_hash_aset(opts, bigdecimal_load_sym, bigdecimal_sym);break;
    case FloatDec:	rb_hash_aset(opts, bigdecimal_load_sym, float_sym);	break;
    case AutoDec:
    default:		rb_hash_aset(opts, bigdecimal_load_sym, auto_sym);	break;
    }
    rb_hash_aset(opts, create_id_sym, (0 == oj_default_options.create_id) ? Qnil : rb_str_new2(oj_default_options.create_id));
    rb_hash_aset(opts, space_sym, (0 == oj_default_options.dump_opts.after_size) ? Qnil : rb_str_new2(oj_default_options.dump_opts.after_sep));
    rb_hash_aset(opts, space_before_sym, (0 == oj_default_options.dump_opts.before_size) ? Qnil : rb_str_new2(oj_default_options.dump_opts.before_sep));
    rb_hash_aset(opts, object_nl_sym, (0 == oj_default_options.dump_opts.hash_size) ? Qnil : rb_str_new2(oj_default_options.dump_opts.hash_nl));
    rb_hash_aset(opts, array_nl_sym, (0 == oj_default_options.dump_opts.array_size) ? Qnil : rb_str_new2(oj_default_options.dump_opts.array_nl));

    switch (oj_default_options.dump_opts.nan_dump) {
    case NullNan:	rb_hash_aset(opts, nan_sym, null_sym);	break;
    case RaiseNan:	rb_hash_aset(opts, nan_sym, raise_sym);	break;
    case WordNan:	rb_hash_aset(opts, nan_sym, word_sym);	break;
    case HugeNan:	rb_hash_aset(opts, nan_sym, huge_sym);	break;
    case AutoNan:
    default:		rb_hash_aset(opts, nan_sym, auto_sym);	break;
    }
    rb_hash_aset(opts, omit_nil_sym, oj_default_options.dump_opts.omit_nil ? Qtrue : Qfalse);
    rb_hash_aset(opts, hash_class_sym, oj_default_options.hash_class);
    
    return opts;
}

.default_options=(opts) ⇒ nil

Sets the default options for load and dump.

Parameters:

  • opts (Hash) —

    options to change

  • :indent (Fixnum|String|nil) —

    number of spaces to indent each element in a JSON document or the String to use for identation.

  • :circular (true|false|nil) —

    support circular references while dumping

  • :auto_define (true|false|nil) —

    automatically define classes if they do not exist

  • :symbol_keys (true|false|nil) —

    convert hash keys to symbols

  • :class_cache (true|false|nil) —

    cache classes for faster parsing

  • :escape (:newline|:json|:xss_safe|:ascii|nil) —

    mode encodes all high-bit characters as escaped sequences if :ascii, :json is standand UTF-8 JSON encoding, :newline is the same as :json but newlines are not escaped, and :xss_safe escapes &, <, and >, and some others.

  • :bigdecimal_as_decimal (true|false|nil) —

    dump BigDecimal as a decimal number or as a String

  • :bigdecimal_load (:bigdecimal|:float|:auto|nil) —

    load decimals as BigDecimal instead of as a Float. :auto pick the most precise for the number of digits.

  • load (:object|:strict|:compat|:null) —

    and dump mode to use for JSON :strict raises an exception when a non-supported Object is encountered. :compat attempts to extract variable values from an Object using to_json() or to_hash() then it walks the Object’s variables if neither is found. The :object mode ignores to_hash() and to_json() methods and encodes variables using code internal to the Oj gem. The :null mode ignores non-supported Objects and replaces them with a null.

  • time (:unix|:xmlschema|:ruby) —

    format when dumping in :compat mode :unix decimal number denoting the number of seconds since 1/1/1970, :unix_zone decimal number denoting the number of seconds since 1/1/1970 plus the utc_offset in the exponent , :xmlschema date-time format taken from XML Schema as a String, :ruby Time.to_s formatted String

  • :create_id (String|nil) —

    create id for json compatible object encoding

  • :second_precision (Fixnum|nil) —

    number of digits after the decimal when dumping the seconds portion of time

  • :float_precision (Fixnum|nil) —

    number of digits of precision when dumping floats, 0 indicates use Ruby

  • :use_to_json (true|false|nil) —

    call to_json() methods on dump, default is false

  • :nilnil (true|false|nil) —

    if true a nil input to load will return nil and not raise an Exception

  • :allow_gc (true|false|nil) —

    allow or prohibit GC during parsing, default is true (allow)

  • :quirks_mode (true|false|nil) —

    allow single JSON values instead of documents, default is true (allow)

  • :allow_invalid_unicode (true|false|nil) —

    allow invalid unicode, default is false (don’t allow)

  • :space (String|nil) —

    String to use for the space after the colon in JSON object fields

  • :space_before (String|nil) —

    String to use before the colon separator in JSON object fields

  • :object_nl (String|nil) —

    String to use after a JSON object field value

  • :array_nl (String|nil) —

    String to use after a JSON array value

  • :nan (:null|:huge|:word|:raise) —

    how to dump Infinity and NaN in null, strict, and compat mode. :null places a null, :huge places a huge number, :word places Infinity or NaN, :raise raises and exception, :auto uses default for each mode.

  • :hash_class (Class|nil) —

    Class to use instead of Hash on load

  • :omit_nil (true|false) —

    if true Hash and Object attributes with nil values are omitted

Returns:

  • (nil) 


357
358
359
360
361
362
363
 
# File 'ext/oj/oj.c', line 357

static VALUE
set_def_opts(VALUE self, VALUE opts) {
    Check_Type(opts, T_HASH);
    oj_parse_options(opts, &oj_default_options);

    return Qnil;
}

.dump(obj, options) ⇒ Object

Dumps an Object (obj) to a string.

Parameters:

  • obj (Object) —

    Object to serialize as an JSON document String

  • options (Hash) —

    same as default_options



928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
 
# File 'ext/oj/oj.c', line 928

static VALUE
dump(int argc, VALUE *argv, VALUE self) {
    char		buf[4096];
    struct _Out		out;
    struct _Options	copts = oj_default_options;
    VALUE		rstr;
    
    if (2 == argc) {
	oj_parse_options(argv[1], &copts);
    }
    out.buf = buf;
    out.end = buf + sizeof(buf) - 10;
    out.allocated = 0;
    out.omit_nil = copts.dump_opts.omit_nil;
    oj_dump_obj_to_json(*argv, &copts, &out);
    if (0 == out.buf) {
	rb_raise(rb_eNoMemError, "Not enough memory.");
    }
    rstr = rb_str_new2(out.buf);
    rstr = oj_encode(rstr);
    if (out.allocated) {
	xfree(out.buf);
    }
    return rstr;
}

.load(json, options) ⇒ Object, ...

Parses a JSON document String into a Object, Hash, Array, String, Fixnum, Float, true, false, or nil according to the default mode or the mode specified. Raises an exception if the JSON is malformed or the classes specified are not valid. If the string input is not a valid JSON document (an empty string is not a valid JSON document) an exception is raised.

When used with a document that has multiple JSON elements the block, if any, will be yielded to. If no block then the last element read will be returned.

This parser operates on string and will attempt to load files into memory if a file object is passed as the first argument. A stream input will be parsed using a stream parser but others use the slightly faster string parser.

A block can also be provided with a single argument. That argument will be the parsed JSON document. This is useful when parsing a string that includes multiple JSON documents.

Returns:

  • (Object, Hash, Array, String, Fixnum, Float, true, false, nil)

Parameters:

  • json (String|IO) —

    JSON String or an Object that responds to read()

  • options (Hash) —

    load options (same as default_options)



754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
 
# File 'ext/oj/oj.c', line 754

static VALUE
load(int argc, VALUE *argv, VALUE self) {
    Mode	mode = oj_default_options.mode;

    if (1 > argc) {
	rb_raise(rb_eArgError, "Wrong number of arguments to load().");
    }
    if (2 <= argc) {
	VALUE	ropts = argv[1];
	VALUE	v;

	Check_Type(ropts, T_HASH);
	if (Qnil != (v = rb_hash_lookup(ropts, mode_sym))) {
	    if (object_sym == v) {
		mode = ObjectMode;
	    } else if (strict_sym == v) {
		mode = StrictMode;
	    } else if (compat_sym == v) {
		mode = CompatMode;
	    } else if (null_sym == v) {
		mode = NullMode;
	    } else {
		rb_raise(rb_eArgError, ":mode must be :object, :strict, :compat, or :null.");
	    }
	}
    }
    switch (mode) {
    case StrictMode:
	return oj_strict_parse(argc, argv, self);
    case NullMode:
    case CompatMode:
	return oj_compat_parse(argc, argv, self);
    case ObjectMode:
    default:
	break;
    }
    return oj_object_parse(argc, argv, self);
}

.load_file(*args) ⇒ Object

call-seq: load_file(path, options) => Object, Hash, Array, String, Fixnum, Float, true, false, or nil

Parses a JSON document String into a Object, Hash, Array, String, Fixnum, Float, true, false, or nil according to the default mode or the mode specified. Raises an exception if the JSON is malformed or the classes specified are not valid. If the string input is not a valid JSON document (an empty string is not a valid JSON document) an exception is raised.

When used with a document that has multiple JSON elements the block, if any, will be yielded to. If no block then the last element read will be returned.

If the input file is not a valid JSON document (an empty file is not a valid JSON document) an exception is raised.

This is a stream based parser which allows a large or huge file to be loaded without pulling the whole file into memory.

A block can also be provided with a single argument. That argument will be the parsed JSON document. This is useful when parsing a string that includes multiple JSON documents.

Parameters:

  • path (String) —

    path to a file containing a JSON document

  • options (Hash) —

    load options (same as default_options)



819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
 
# File 'ext/oj/oj.c', line 819

static VALUE
load_file(int argc, VALUE *argv, VALUE self) {
    char		*path;
    int			fd;
    Mode		mode = oj_default_options.mode;
    struct _ParseInfo	pi;

    if (1 > argc) {
	rb_raise(rb_eArgError, "Wrong number of arguments to load().");
    }
    Check_Type(*argv, T_STRING);
    pi.options = oj_default_options;
    pi.handler = Qnil;
    pi.err_class = Qnil;
    if (2 <= argc) {
	VALUE	ropts = argv[1];
	VALUE	v;

	Check_Type(ropts, T_HASH);
	if (Qnil != (v = rb_hash_lookup(ropts, mode_sym))) {
	    if (object_sym == v) {
		mode = ObjectMode;
	    } else if (strict_sym == v) {
		mode = StrictMode;
	    } else if (compat_sym == v) {
		mode = CompatMode;
	    } else if (null_sym == v) {
		mode = NullMode;
	    } else {
		rb_raise(rb_eArgError, ":mode must be :object, :strict, :compat, or :null.");
	    }
	}
    }
    path = StringValuePtr(*argv);
    if (0 == (fd = open(path, O_RDONLY))) {
	rb_raise(rb_eIOError, "%s", strerror(errno));
    }
    switch (mode) {
    case StrictMode:
	oj_set_strict_callbacks(&pi);
	return oj_pi_sparse(argc, argv, &pi, fd);
    case NullMode:
    case CompatMode:
	oj_set_compat_callbacks(&pi);
	return oj_pi_sparse(argc, argv, &pi, fd);
    case ObjectMode:
    default:
	break;
    }
    oj_set_object_callbacks(&pi);

    return oj_pi_sparse(argc, argv, &pi, fd);
}

.mimic_JSON(*args) ⇒ Object

call-seq: mimic_JSON() => Module

Creates the JSON module with methods and classes to mimic the JSON gem. After this method is invoked calls that expect the JSON module will use Oj instead and be faster than the original JSON. Most options that could be passed to the JSON methods are supported. The calls to set parser or generator will not raise an Exception but will not have any effect. The method can also be called after the json gem is loaded. The necessary methods on the json gem will be replaced with Oj methods.

Note that this also sets the default options of :mode to :compat and :encoding to :ascii.



1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
 
# File 'ext/oj/oj.c', line 1981

static VALUE
define_mimic_json(int argc, VALUE *argv, VALUE self) {
    VALUE	ext;
    VALUE	dummy;
    VALUE	verbose;
    VALUE	json_error;
    
    // Either set the paths to indicate JSON has been loaded or replaces the
    // methods if it has been loaded.
    if (rb_const_defined_at(rb_cObject, rb_intern("JSON"))) {
	mimic = rb_const_get_at(rb_cObject, rb_intern("JSON"));
    } else {
	mimic = rb_define_module("JSON");
    }
    verbose = rb_gv_get("$VERBOSE");
    rb_gv_set("$VERBOSE", Qfalse);
    rb_define_module_function(rb_cObject, "JSON", mimic_dump_load, -1);
    if (rb_const_defined_at(mimic, rb_intern("Ext"))) {
	ext = rb_const_get_at(mimic, rb_intern("Ext"));
     } else {
	ext = rb_define_module_under(mimic, "Ext");
    }
    if (!rb_const_defined_at(ext, rb_intern("Parser"))) {
	dummy = rb_define_class_under(ext, "Parser", rb_cObject);
    }
    if (!rb_const_defined_at(ext, rb_intern("Generator"))) {
	dummy = rb_define_class_under(ext, "Generator", rb_cObject);
    }
    // convince Ruby that the json gem has already been loaded
    dummy = rb_gv_get("$LOADED_FEATURES");
    if (rb_type(dummy) == T_ARRAY) {
	rb_ary_push(dummy, rb_str_new2("json"));
	if (0 < argc) {
	    VALUE	mimic_args[1];

	    *mimic_args = *argv;
	    rb_funcall2(Oj, rb_intern("mimic_loaded"), 1, mimic_args);
	} else {
	    rb_funcall2(Oj, rb_intern("mimic_loaded"), 0, 0);
	}
    }
    rb_define_module_function(mimic, "parser=", no_op1, 1);
    rb_define_module_function(mimic, "generator=", no_op1, 1);
    rb_define_module_function(mimic, "create_id=", mimic_set_create_id, 1);
    rb_define_module_function(mimic, "create_id", mimic_create_id, 0);

    rb_define_module_function(mimic, "dump", mimic_dump, -1);
    rb_define_module_function(mimic, "load", mimic_load, -1);
    rb_define_module_function(mimic, "restore", mimic_load, -1);
    rb_define_module_function(mimic, "recurse_proc", mimic_recurse_proc, 1);
    rb_define_module_function(mimic, "[]", mimic_dump_load, -1);

    rb_define_module_function(mimic, "generate", mimic_generate, -1);
    rb_define_module_function(mimic, "fast_generate", mimic_generate, -1);
    rb_define_module_function(mimic, "pretty_generate", mimic_pretty_generate, -1);
    /* for older versions of JSON, the deprecated unparse methods */
    rb_define_module_function(mimic, "unparse", mimic_generate, -1);
    rb_define_module_function(mimic, "fast_unparse", mimic_generate, -1);
    rb_define_module_function(mimic, "pretty_unparse", mimic_pretty_generate, -1);

    rb_define_module_function(mimic, "parse", mimic_parse, -1);
    rb_define_module_function(mimic, "parse!", mimic_parse, -1);

    rb_define_method(rb_cObject, "to_json", mimic_object_to_json, -1);

    rb_gv_set("$VERBOSE", verbose);

    create_additions_sym = ID2SYM(rb_intern("create_additions"));	rb_gc_register_address(&create_additions_sym);
    symbolize_names_sym = ID2SYM(rb_intern("symbolize_names"));		rb_gc_register_address(&symbolize_names_sym);

    if (rb_const_defined_at(mimic, rb_intern("JSONError"))) {
	rb_funcall(mimic, rb_intern("remove_const"), 1, ID2SYM(rb_intern("JSONError")));
    }
    json_error = rb_define_class_under(mimic, "JSONError", rb_eStandardError);
    if (rb_const_defined_at(mimic, rb_intern("ParserError"))) {
	rb_funcall(mimic, rb_intern("remove_const"), 1, ID2SYM(rb_intern("ParserError")));
    }
    json_parser_error_class = rb_define_class_under(mimic, "ParserError", json_error);

    if (!rb_const_defined_at(mimic, rb_intern("State"))) {
        rb_define_class_under(mimic, "State", rb_cObject);
    }

    oj_default_options = mimic_object_to_json_options;
    oj_default_options.to_json = Yes;

    return mimic;
}

.mimic_loaded(mimic_paths = []) ⇒ Object 



32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
 
# File 'lib/oj/mimic.rb', line 32

def self.mimic_loaded(mimic_paths=[])
  $LOAD_PATH.each do |d|
    next unless File.exist?(d)

    jfile = File.join(d, 'json.rb')
    $LOADED_FEATURES << jfile unless $LOADED_FEATURES.include?(jfile) if File.exist?(jfile)
    
    Dir.glob(File.join(d, 'json', '**', '*.rb')).each do |file|
      $LOADED_FEATURES << file unless $LOADED_FEATURES.include?(file)
    end
  end
  mimic_paths.each { |p| $LOADED_FEATURES << p }
  $LOADED_FEATURES << 'json' unless $LOADED_FEATURES.include?('json')

  if Object.const_defined?('OpenStruct')
    OpenStruct.class_eval do
      # Both the JSON gem and Rails monkey patch as_json. Let them battle it out.
      unless defined?(self.as_json)
        def as_json(*)
          name = self.class.name.to_s
          raise JSON::JSONError, "Only named structs are supported!" if 0 == name.length
          { JSON.create_id => name, 't' => table }
        end
      end
      def self.json_create(h)
        new(h['t'])
      end
    end
  end

  Range.class_eval do
    # Both the JSON gem and Rails monkey patch as_json. Let them battle it out.
    unless defined?(self.as_json)
      def as_json(*)
        {JSON.create_id => 'Range', 'a' => [first, last, exclude_end?]}
      end
    end
    def self.json_create(h)
      new(h['a'])
    end
  end

  Rational.class_eval do
    # Both the JSON gem and Rails monkey patch as_json. Let them battle it out.
    unless defined?(self.as_json)
      def as_json(*)
        {JSON.create_id => 'Rational', 'n' => numerator, 'd' => denominator }
      end
    end
    def self.json_create(h)
      Rational(h['n'], h['d'])
    end
  end

  Regexp.class_eval do
    # Both the JSON gem and Rails monkey patch as_json. Let them battle it out.
    unless defined?(self.as_json)
      def as_json(*)
        {JSON.create_id => 'Regexp', 'o' => options, 's' => source }
      end
    end
    def self.json_create(h)
      new(h['s'], h['o'])
    end
  end

  Struct.class_eval do
    # Both the JSON gem and Rails monkey patch as_json. Let them battle it out.
    unless defined?(self.as_json)
      def as_json(*)
        name = self.class.name.to_s
        raise JSON::JSONError, "Only named structs are supported!" if 0 == name.length
        { JSON.create_id => name, 'v' => values }
      end
    end
    def self.json_create(h)
      new(h['v'])
    end
  end

  Symbol.class_eval do
    # Both the JSON gem and Rails monkey patch as_json. Let them battle it out.
    unless defined?(self.as_json)
      def as_json(*)
        {JSON.create_id => 'Symbol', 's' => to_s }
      end
    end
    def self.json_create(h)
      h['s'].to_sym
    end
  end

  Time.class_eval do
    # Both the JSON gem and Rails monkey patch as_json. Let them battle it out.
    unless defined?(self.as_json)
      def as_json(*)
        {JSON.create_id => 'Symbol', 's' => to_s }
        nsecs = [ tv_usec * 1000 ]
        nsecs << tv_nsec if respond_to?(:tv_nsec)
        nsecs = nsecs.max
        { JSON.create_id => 'Time', 's' => tv_sec, 'n' => nsecs }
      end
    end
    def self.json_create(h)
      if usec = h.delete('u')
        h['n'] = usec * 1000
      end
      if instance_methods.include?(:tv_nsec)
        at(h['s'], Rational(h['n'], 1000))
      else
        at(h['s'], h['n'] / 1000)
      end
    end
  end

  JSON.module_eval do
    def self.dump_default_options
      Oj::MimicDumpOption.new
    end

    def self.dump_default_options=(h)
      m = Oj::MimicDumpOption.new
      h.each do |k,v|
        m[k] = v
      end
    end
  end

end

.object_load(json, options) ⇒ Object, ...

Parses a JSON document String into an Object, Hash, Array, String, Fixnum, Float, true, false, or nil. In the :object mode the JSON should have been generated by Oj.dump(). The parser will reconstitute the original marshalled or dumped Object. The :auto_define and :circular options have meaning with this parsing mode.

When used with a document that has multiple JSON elements the block, if any, will be yielded to. If no block then the last element read will be returned.

Raises an exception if the JSON is malformed or the classes specified are not valid. If the input is not a valid JSON document (an empty string is not a valid JSON document) an exception is raised.

Note: Oj is not able to automatically deserialize all classes that are a subclass of a Ruby Exception. Only exception that take one required string argument in the initialize() method are supported. This is an example of how to write an Exception subclass that supports both a single string intializer and an Exception as an argument. Additional optional arguments can be added as well.

The reason for this restriction has to do with a design decision on the part of the Ruby developers. Exceptions are special Objects. They do not follow the rules of other Objects. Exceptions have ‘mesg’ and a ‘bt’ attribute. Note that these are not ‘@mesg’ and ‘@bt’. They can not be set using the normal C or Ruby calls. The only way I have found to set the ‘mesg’ attribute is through the initializer. Unfortunately that means any subclass that provides a different initializer can not be automatically decoded. A way around this is to use a create function but this example shows an alternative.

A block can also be provided with a single argument. That argument will be the parsed JSON document. This is useful when parsing a string that includes multiple JSON documents.

Returns:

  • (Object, Hash, Array, String, Fixnum, Float, true, false, nil)

Parameters:

  • json (String|IO) —

    JSON String or an Object that responds to read()

  • options (Hash) —

    load options (same as default_options)

.register_odd(clas, create_object, create_method, *members) ⇒ Object

Registers a class as special. This is useful for working around subclasses of primitive types as is done with ActiveSupport classes. The use of this function should be limited to just classes that can not be handled in the normal way. It is not intended as a hook for changing the output of all classes as it is not optimized for large numbers of classes.

Parameters:

  • clas (Class|Module) —

    Class or Module to be made special

  • create_object (Object) —

    object to call the create method on

  • create_method (Symbol) —

    method on the clas that will create a new instance of the clas when given all the member values in the order specified.

  • members (Symbol|String) —

    methods used to get the member values from instances of the clas



1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
 
# File 'ext/oj/oj.c', line 1014

static VALUE
register_odd(int argc, VALUE *argv, VALUE self) {
    if (3 > argc) {
	rb_raise(rb_eArgError, "incorrect number of arguments.");
    }
    switch (rb_type(*argv)) {
    case T_CLASS:
    case T_MODULE:
	break;
    default:
	rb_raise(rb_eTypeError, "expected a class or module.");
	break;
    }
    Check_Type(argv[2], T_SYMBOL);
    if (MAX_ODD_ARGS < argc - 2) {
	rb_raise(rb_eArgError, "too many members.");
    }
    oj_reg_odd(argv[0], argv[1], argv[2], argc - 3, argv + 3, false);

    return Qnil;
}

.register_odd_raw(clas, create_object, create_method, dump_method) ⇒ Object

Registers a class as special and expect the output to be a string that can be included in the dumped JSON directly. This is useful for working around subclasses of primitive types as is done with ActiveSupport classes. The use of this function should be limited to just classes that can not be handled in the normal way. It is not intended as a hook for changing the output of all classes as it is not optimized for large numbers of classes. Be careful with this option as the JSON may be incorrect if invalid JSON is returned.

Parameters:

  • clas (Class|Module) —

    Class or Module to be made special

  • create_object (Object) —

    object to call the create method on

  • create_method (Symbol) —

    method on the clas that will create a new instance of the clas when given all the member values in the order specified.

  • dump_method (Symbol|String) —

    method to call on the object being serialized to generate the raw JSON.



1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
 
# File 'ext/oj/oj.c', line 1054

static VALUE
register_odd_raw(int argc, VALUE *argv, VALUE self) {
    if (3 > argc) {
	rb_raise(rb_eArgError, "incorrect number of arguments.");
    }
    switch (rb_type(*argv)) {
    case T_CLASS:
    case T_MODULE:
	break;
    default:
	rb_raise(rb_eTypeError, "expected a class or module.");
	break;
    }
    Check_Type(argv[2], T_SYMBOL);
    if (MAX_ODD_ARGS < argc - 2) {
	rb_raise(rb_eArgError, "too many members.");
    }
    oj_reg_odd(argv[0], argv[1], argv[2], 1, argv + 3, true);

    return Qnil;
}

.safe_load(doc) ⇒ Hash|Array|String|Fixnum|Bignum|BigDecimal|nil|True|False

Loads a JSON document in strict mode with :auto_define and :symbol_keys turned off. This function should be safe to use with JSON received on an unprotected public interface.

Parameters:

  • doc (String|IO) —

    JSON String or IO to load

Returns:

  • (Hash|Array|String|Fixnum|Bignum|BigDecimal|nil|True|False) 


882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
 
# File 'ext/oj/oj.c', line 882

static VALUE
safe_load(VALUE self, VALUE doc) {
    struct _ParseInfo	pi;
    VALUE		args[1];

    pi.err_class = Qnil;
    pi.options = oj_default_options;
    pi.options.auto_define = No;
    pi.options.sym_key = No;
    pi.options.mode = StrictMode;
    oj_set_strict_callbacks(&pi);
    *args = doc;

    return oj_pi_parse(1, args, &pi, 0, 0, 1);
}

.saj_parse ⇒ Object

.sc_parse ⇒ Object

.strict_load(json, options) ⇒ Hash, ...

Parses a JSON document String into an Hash, Array, String, Fixnum, Float, true, false, or nil. It parses using a mode that is strict in that it maps each primitive JSON type to a similar Ruby type. The :create_id is not honored in this mode. Note that a Ruby Hash is used to represent the JSON Object type. These two are not the same since the JSON Object type can have repeating entries with the same key and Ruby Hash can not.

When used with a document that has multiple JSON elements the block, if any, will be yielded to. If no block then the last element read will be returned.

Raises an exception if the JSON is malformed or the classes specified are not valid. If the input is not a valid JSON document (an empty string is not a valid JSON document) an exception is raised.

A block can also be provided with a single argument. That argument will be the parsed JSON document. This is useful when parsing a string that includes multiple JSON documents.

Returns:

  • (Hash, Array, String, Fixnum, Float, true, false, nil)

Parameters:

  • json (String|IO) —

    JSON String or an Object that responds to read()

  • options (Hash) —

    load options (same as default_options)

.to_file(file_path, obj, options) ⇒ Object

Dumps an Object to the specified file.

Parameters:

  • file_path (String) —

    file path to write the JSON document to

  • obj (Object) —

    Object to serialize as an JSON document String

  • options (Hash) —

    formating options

  • :indent (Fixnum) —

    format expected

  • :circular (true|false) —

    allow circular references, default: false



964
965
966
967
968
969
970
971
972
973
974
975
 
# File 'ext/oj/oj.c', line 964

static VALUE
to_file(int argc, VALUE *argv, VALUE self) {
    struct _Options	copts = oj_default_options;
    
    if (3 == argc) {
	oj_parse_options(argv[2], &copts);
    }
    Check_Type(*argv, T_STRING);
    oj_write_obj_to_file(argv[1], StringValuePtr(*argv), &copts);

    return Qnil;
}

.to_stream(io, obj, options) ⇒ Object

Dumps an Object to the specified IO stream.

Parameters:

  • io (IO) —

    IO stream to write the JSON document to

  • obj (Object) —

    Object to serialize as an JSON document String

  • options (Hash) —

    formating options

  • :indent (Fixnum) —

    format expected

  • :circular (true|false) —

    allow circular references, default: false



986
987
988
989
990
991
992
993
994
995
996
 
# File 'ext/oj/oj.c', line 986

static VALUE
to_stream(int argc, VALUE *argv, VALUE self) {
    struct _Options	copts = oj_default_options;
    
    if (3 == argc) {
	oj_parse_options(argv[2], &copts);
    }
    oj_write_obj_to_stream(argv[1], *argv, &copts);

    return Qnil;
}