Day 23 - json-next Gem - Read Next Generation JSON Versions (HanSON, SON, JSONX/JSON11, Etc) with Comments, Unquoted Keys, Multi-Line Strings, Trailing Commas, Optional Commas, and More
Written by {% avatar geraldb %} Gerald Bauer
A code monkey and enthusiastic collector (and publisher) of open football and beer data. Skills include Ruby, SQLite and CSV. Spec lead of CSV <3 JSON.
Let's start with a quiz:
Q: What's your favorite text ("human"/non-binary) data exchange format?
- (A) JSON
- (B) XML
- (C) HTML
- (D) YAML
- (E) TOML/INI
- (F) SQL
- (G) CSV
- (H) Other, please tell!
If you picked (A) JSON - let's continue with a question:
- Comments, Comments, Comments
- Unquoted Keys
- Multi-Line Strings
- a) Folded -- Folds Newlines
- b) Unfolded
- Trailing Commas in Arrays and Objects
More:
- Date/DateTime/Timestamp Type
- Optional Commas
- Optional Unquoted String Values
- "Raw" String (e.g.
''
instead of""
)- No need to escape
\
or"
etc. To escape'
use'''
e.g.''''Henry's Themes'''
- No need to escape
- More Data Types (
set
,map
,symbol
, etc.) - And much more
Fixing JSON. Request for comments, please!
We can easily agree on what’s wrong with JSON, and I can't help wondering if it'd be worth fixing it.
-- Tim Bray (Fixing JSON)
XML already does everything JSON does! And there's no way to differentiate between nodes and attributes! And there are no namespaces! And no schemas! What's the point of JSON?
-- Anonymous
We need to fix engineers that try to 'fix JSON', absolutely nothing is broken with JSON.
-- Anonymous
And what's your take? Enter the json-next gem.
The json-next gem lets you convert and read (parse) next generation json versions
including: HanSON e.g. HANSON.parse
, SON e.g. SON.parse
, JSONX e.g. JSONX.parse
.
HanSON - JSON for Humans by Tim Jansen et al
HanSON is an extension of JSON with a few simple additions to the spec:
- quotes for strings are optional if they follow JavaScript identifier rules.
- you can alternatively use backticks, as in ES6's template string literal, as quotes for strings. A backtick-quoted string may span several lines and you are not required to escape regular quote characters, only backticks. Backslashes still need to be escaped, and all other backslash-escape sequences work like in regular JSON.
- for single-line strings, single quotes (
''
) are supported in addition to double quotes (""
) - you can use JavaScript comments, both single line (
//
) and multi-line comments (/* */
), in all places where JSON allows whitespace. - Commas after the last list element or object property will be ignored.
Example:
{
listName: "Sesame Street Monsters", // note that listName needs no quotes
content: [
{
name: "Cookie Monster",
/* Note the template quotes and unescaped regular quotes in the next string */
background: `Cookie Monster used to be a
monster that ate everything, especially cookies.
These days he is forced to eat "healthy" food.`
}, {
// You can single-quote strings too:
name: 'Herry Monster',
background: `Herry Monster is a furry blue monster with a purple nose.
He's mostly retired today.`
}, // don't worry, the trailing comma will be ignored
]
}
Use HANSON.convert
to convert HanSON text to ye old' JSON text:
{
"listName": "Sesame Street Monsters",
"content": [
{ "name": "Cookie Monster",
"background": "Cookie Monster used to be a\n ... to eat \"healthy\" food."
},
{ "name": "Herry Monster",
"background": "Herry Monster is a furry blue monster with a purple nose.\n ... today."
}
]
}
Use HANSON.parse
instead of JSON.parse
to parse text to ruby hash / array / etc.:
{
"listName" => "Sesame Street Monsters",
"content" => [
{ "name" => "Cookie Monster",
"background" => "Cookie Monster used to be a\n ... to eat \"healthy\" food."
},
{ "name" => "Herry Monster",
"background" => "Herry Monster is a furry blue monster with a purple nose.\n ... today."
}
]
}
SON - Simple Object Notation by Aleksander Gurin et al
Simple data format similar to JSON, but with some minor changes:
- comments starts with
#
sign and ends with newline (\n
) - comma after an object key-value pair is optional
- comma after an array item is optional
JSON is compatible with SON in a sense that JSON data is also SON data, but not vise versa.
Example:
{
# Personal information
"name": "Alexander Grothendieck"
"fields": "mathematics"
"main_topics": [
"Etale cohomology"
"Motives"
"Topos theory"
"Schemes"
]
"numbers": [1 2 3 4]
"mixed": [1.1 -2 true false null]
}
Use SON.convert
to convert SON text to ye old' JSON text:
{
"name": "Alexander Grothendieck",
"fields": "mathematics",
"main_topics": [
"Etale cohomology",
"Motives",
"Topos theory",
"Schemes"
],
"numbers": [1, 2, 3, 4],
"mixed": [1.1, -2, true, false, null]
}
Use SON.parse
instead of JSON.parse
to parse text to ruby hash / array / etc.:
{
"name" => "Alexander Grothendieck",
"fields" => "mathematics",
"main_topics" =>
["Etale cohomology", "Motives", "Topos theory", "Schemes"],
"numbers" => [1, 2, 3, 4],
"mixed" => [1.1, -2, true, false, nil]
}
JSON with Extensions or JSON v1.1 (a.k.a. JSON11 or JSON XI or JSON II)
Includes all JSON extensions from HanSON:
- quotes for strings are optional if they follow JavaScript identifier rules.
- you can alternatively use backticks, as in ES6's template string literal, as quotes for strings. A backtick-quoted string may span several lines and you are not required to escape regular quote characters, only backticks. Backslashes still need to be escaped, and all other backslash-escape sequences work like in regular JSON.
- for single-line strings, single quotes (
''
) are supported in addition to double quotes (""
) - you can use JavaScript comments, both single line (
//
) and multi-line comments (/* */
), in all places where JSON allows whitespace. - Commas after the last list element or object property will be ignored.
Plus all JSON extensions from SON:
- comments starts with
#
sign and ends with newline (\n
) - comma after an object key-value pair is optional
- comma after an array item is optional
Plus some more extra JSON extensions:
- unquoted strings following the JavaScript identifier rules can use the dash (
-
) too e.g. allows common keys such ascore-js
,babel-preset-es2015
,eslint-config-jquery
and others
Example:
{
# use shell-like (or ruby-like) comments
listName: "Sesame Street Monsters" # note: comments after key-value pairs are optional
content: [
{
name: "Cookie Monster"
// note: the template quotes and unescaped regular quotes in the next string
background: `Cookie Monster used to be a
monster that ate everything, especially cookies.
These days he is forced to eat "healthy" food.`
}, {
// You can single-quote strings too:
name: 'Herry Monster',
background: `Herry Monster is a furry blue monster with a purple nose.
He's mostly retired today.`
}, /* don't worry, the trailing comma will be ignored */
]
}
Use JSONX.convert
(or JSONXI.convert
or JSON11.convert
or JSONII.convert
) to convert JSONX text to ye old' JSON text:
{
"listName": "Sesame Street Monsters",
"content": [
{ "name": "Cookie Monster",
"background": "Cookie Monster used to be a\n ... to eat \"healthy\" food."
},
{ "name": "Herry Monster",
"background": "Herry Monster is a furry blue monster with a purple nose.\n ... today."
}
]
}
Use JSONX.parse
(or JSONXI.parse
or JSON11.parse
or JSONII.parse
) instead of JSON.parse
to parse text to ruby hash / array / etc.:
{
"listName" => "Sesame Street Monsters",
"content" => [
{ "name" => "Cookie Monster",
"background" => "Cookie Monster used to be a\n ... to eat \"healthy\" food."
},
{ "name" => "Herry Monster",
"background" => "Herry Monster is a furry blue monster with a purple nose.\n ... today."
}
]
}
require 'json/next'
text1 =<<TXT
{
listName: "Sesame Street Monsters", // note that listName needs no quotes
content: [
{
name: "Cookie Monster",
/* Note the template quotes and unescaped regular quotes in the next string */
background: `Cookie Monster used to be a
monster that ate everything, especially cookies.
These days he is forced to eat "healthy" food.`
}, {
// You can single-quote strings too:
name: 'Herry Monster',
background: `Herry Monster is a furry blue monster with a purple nose.
He's mostly retired today.`
}, // don't worry, the trailing comma will be ignored
]
}
TXT
pp HANSON.parse( text1 ) # note: is the same as JSON.parse( HANSON.convert( text ))
resulting in:
{
"listName" => "Sesame Street Monsters",
"content" => [
{ "name" => "Cookie Monster",
"background" => "Cookie Monster used to be a\n ... to eat \"healthy\" food."
},
{ "name" => "Herry Monster",
"background" => "Herry Monster is a furry blue monster with a purple nose.\n ... today."
}
]
}
and
text2 =<<TXT
{
# Personal information
"name": "Alexander Grothendieck"
"fields": "mathematics"
"main_topics": [
"Etale cohomology"
"Motives"
"Topos theory"
"Schemes"
]
"numbers": [1 2 3 4]
"mixed": [1.1 -2 true false null]
}
TXT
pp SON.parse( text2 ) # note: is the same as JSON.parse( SON.convert( text ))
resulting in:
{
"name" => "Alexander Grothendieck",
"fields" => "mathematics",
"main_topics" =>
["Etale cohomology", "Motives", "Topos theory", "Schemes"],
"numbers" => [1, 2, 3, 4],
"mixed" => [1.1, -2, true, false, nil]
}
and
text3 =<<TXT
{
# use shell-like (or ruby-like) comments
listName: "Sesame Street Monsters" # note: comments after key-value pairs are optional
content: [
{
name: "Cookie Monster"
// note: the template quotes and unescaped regular quotes in the next string
background: `Cookie Monster used to be a
monster that ate everything, especially cookies.
These days he is forced to eat "healthy" food.`
}, {
// You can single-quote strings too:
name: 'Herry Monster',
background: `Herry Monster is a furry blue monster with a purple nose.
He's mostly retired today.`
}, /* don't worry, the trailing comma will be ignored */
]
}
TXT
pp JSONX.parse( text3 ) # note: is the same as JSON.parse( JSONX.convert( text ))
pp JSONXI.parse( text3 ) # note: is the same as JSON.parse( JSONXI.convert( text ))
pp JSON11.parse( text3 ) # note: is the same as JSON.parse( JSON11.convert( text ))
pp JSONII.parse( text3 ) # note: is the same as JSON.parse( JSONII.convert( text ))
resulting in:
{
"listName" => "Sesame Street Monsters",
"content" => [
{ "name" => "Cookie Monster",
"background" => "Cookie Monster used to be a\n ... to eat \"healthy\" food."
},
{ "name" => "Herry Monster",
"background" => "Herry Monster is a furry blue monster with a purple nose.\n ... today."
}
]
}
See the Awesome JSON - What's Next? collection / page.
- Home :: github.com/json-next/json-next
- Gem :: json-next
- Docs :: json-next
- Talk Notes :: JSON - What's Next for the No 2. Universal (Non-Binary/Human) Data Exchange Format