rison

README

go-rison

Go port of Rison.

This page describes Rison, a data serialization format optimized for compactness in URIs. Rison is a slight variation of JSON that looks vastly superior after URI encoding. Rison still expresses exactly the same set of data structures as JSON, so data can be translated back and forth without loss or guesswork.

Examples

import "gopkg.in/sakura-internet/go-rison.v3"

func ExampleDecode() {
	r := "(id:example,str:'string',num:100,yes:!t,nil:!n,arr:!(1,2,3))"
	v, _ := rison.Decode([]byte(r), rison.Rison)
	m := v.(map[string]interface{})
	fmt.Printf(
		"id:%v, str:%v, num:%v, yes:%v, nil:%v, arr:%v",
		m["id"], m["str"], m["num"], m["yes"], m["nil"], m["arr"],
	)
	// Output: id:example, str:string, num:100, yes:true, nil:<nil>, arr:[1 2 3]
}

// The object keys corresponding the struct fields can be
// specified in struct tag (not "rison" but) "json".
type exampleStruct struct {
	I int64       `json:"i"`
	F float64     `json:"f"`
	S string      `json:"s"`
	B bool        `json:"b"`
	P *bool       `json:"p"`
	A []int64     `json:"a"`
	X interface{} `json:"x"`
}

func ExampleUnmarshal() {
	r := "(i:1,f:2.3,s:str,b:!t,p:!n,a:!(7,8,9),x:(y:Y))"
	var v exampleStruct
	_ = rison.Unmarshal([]byte(r), &v, rison.Rison)
	fmt.Printf("%+v\n", v)
	// Output: {I:1 F:2.3 S:str B:true P:<nil> A:[7 8 9] X:map[y:Y]}
}

func ExampleMarshal() {
	v := exampleStruct{
		I: 1,
		F: 2.3,
		S: "str",
		B: true,
		P: nil,
		A: []int64{7, 8, 9},
		X: map[string]interface{}{"y": "Y"},
	}
	r, _ := rison.Marshal(&v, rison.Rison)
	fmt.Println(string(r))
	// Output: (a:!(7,8,9),b:!t,f:2.3,i:1,p:!n,s:str,x:(y:Y))
}

func ExampleToJSON() {
	r := "!(1,2.3,str,'ing',true,nil,(a:b),!(7,8,9))"
	j, _ := rison.ToJSON([]byte(r), rison.Rison)
	fmt.Printf("%s\n", string(j))
	// Output: [1,2.3,"str","ing","true","nil",{"a":"b"},[7,8,9]]
}

func ExampleQuote() {
	s := "~!*()-_.,:@$'/ \"#%&+;<=>?[\\]^`{|}"
	fmt.Println(rison.QuoteString(s))
	// Output: ~!*()-_.,:@$'/+%22%23%25%26%2B%3B%3C%3D%3E%3F%5B%5C%5D%5E%60%7B%7C%7D
}

func ExampleParseError_ErrorInLang() {
	r := "!("
	_, err := rison.ToJSON([]byte(r), rison.Rison)
	e, _ := err.(interface {
		ErrorInLang(lang string) string
		Langs() []string
	})
	fmt.Println(strings.Join(e.Langs(), ", "))
	fmt.Println(e.ErrorInLang("en"))
	fmt.Println(e.ErrorInLang("ja"))
	// Output:
	// en, ja
	// unmatched "!(" (at the end of string "!(" -> EOS)
	// "!(" が閉じていません (場所: 文字列終端: "!(" → EOS)
}

Descriptions

The following descriptions are some excerpts from the original README and the article:

Differences from JSON syntax

• no whitespace is permitted except inside quoted strings.

• almost all character escaping is left to the uri encoder.

• single-quotes are used for quoting, but quotes can and should be left off strings when the strings are simple identifiers.

• the e+ exponent format is forbidden, since + is not safe in form values and the plain e format is equivalent.

• the E, E+, and E exponent formats are removed.

• object keys should be lexically sorted when encoding. the intent is to improve url cacheability.

• uri-safe tokens are used in place of the standard json tokens:

  rison token	json token	meaning
  '	"	string quote
  !	\	string escape
  (...)	{...}	object
  !(...)	[...]	array

• the JSON literals that look like identifiers (true, false and null) are represented as ! sequences:

  rison token	json token
  !t	true
  !f	false
  !n	null

The ! character plays two similar but different roles, as an escape character within strings, and as a marker for special values. This may be confusing.

Notice that services can distinguish Rison-encoded strings from JSON-encoded strings by checking the first character. Rison structures start with ( or !(. JSON structures start with [ or {. This means that a service which expects a JSON encoded object or array can accept Rison-encoded objects without loss of compatibility.

Interaction with URI %-encoding

Rison syntax is designed to produce strings that be legible after being form- encoded for the query section of a URI. None of the characters in the Rison syntax need to be URI encoded in that context, though the data itself may require URI encoding. Rison tries to be orthogonal to the %-encoding process - it just defines a string format that should survive %-encoding with very little bloat. Rison quoting is only applied when necessary to quote characters that might otherwise be interpreted as special syntax.

Note that most URI encoding libraries are very conservative, percent-encoding many characters that are legal according to RFC 3986. For example, Javascript's builtin encodeURIComponent() function will still make Rison strings difficult to read. The rison.js library includes a more tolerant URI encoder.

Rison uses its own quoting for strings, using the single quote (**'**) as a string delimiter and the exclamation point (**!**) as the string escape character. Both of these characters are legal in uris. Rison quoting is largely inspired by Unix shell command line parsing.

All Unicode characters other than **'** and **!** are legal inside quoted strings. This includes newlines and control characters. Quoting all such characters is left to the %-encoding process.

Grammar

modified from the json.org grammar.

• object
  • ()
  • ( members )
• members
  • pair
  • pair , members
• pair
  • key : value
• array
  • !()
  • !( elements )
• elements
  • value
  • value , elements
• key
  • id
  • string
• value
  • id
  • string
  • number
  • object
  • array
  • !t
  • !f
  • !n
    　　　　────────────
• id
  • idstart
  • idstart idchars
• idchars
  • idchar
  • idchar idchars
• idchar
  • any alphanumeric ASCII character
  • any ASCII character from the set - _ . / ~
  • any non-ASCII Unicode character
• idstart
  • any idchar not in -, digit
    　　　　────────────
• string
  • ''
  • ' strchars '
• strchars
  • strchar
  • strchar strchars
• strchar
  • any Unicode character except ASCII ' and !
  • !!
  • !'
    　　　　────────────
• number
  • int
  • int frac
  • int exp
  • int frac exp
• int
  • digit
  • digit1-9 digits
  • - digit
  • - digit1-9 digits
• frac
  • . digits
• exp
  • e digits
• digits
  • digit
  • digit digits
• e
  • e
  • e-

Documentation

Overview

Package rison implements encoding and decoding of Rison. https://github.com/Nanonid/rison

Rison is a data serialization format optimized for compactness in URIs. Rison is a slight variation of JSON that looks vastly superior after URI encoding. Rison still expresses exactly the same set of data structures as JSON, so data can be translated back and forth without loss or guesswork.

Index

• func Decode(data []byte, m Mode) (interface{}, error)
• func Encode(v interface{}, m Mode) ([]byte, error)
• func FromJSON(data []byte, m Mode) ([]byte, error)
• func Marshal(v interface{}, m Mode) ([]byte, error)
• func Quote(s []byte) []byte
• func QuoteString(s string) string
• func ToJSON(data []byte, m Mode) ([]byte, error)
• func Unmarshal(data []byte, v interface{}, m Mode) error
• type ErrType
• type Mode
• type ParseError
  • func (e *ParseError) Error() string
  • func (e *ParseError) ErrorInLang(lang string) string
  • func (e *ParseError) Langs() []string
  • func (e *ParseError) Translate(lang string)

Examples

• Decode
• FromJSON
• Marshal
• ParseError.ErrorInLang
• Quote
• ToJSON
• Unmarshal

Functions

func Decode

func Decode(data []byte, m Mode) (interface{}, error)

Decode parses the Rison-encoded data and returns the result as the tree of map[string]interface{} (or []interface{} or scalar value).

Example

r := "(id:example,str:'string',num:100,yes:!t,nil:!n,arr:!(1,2,3))"
v, _ := rison.Decode([]byte(r), rison.Rison)
m := v.(map[string]interface{})
fmt.Printf(
	"id:%v, str:%v, num:%v, yes:%v, nil:%v, arr:%v",
	m["id"], m["str"], m["num"], m["yes"], m["nil"], m["arr"],
)

Output:

id:example, str:string, num:100, yes:true, nil:<nil>, arr:[1 2 3]

func Encode

func Encode(v interface{}, m Mode) ([]byte, error)

Encode is an alias of Marshal.

func FromJSON

func FromJSON(data []byte, m Mode) ([]byte, error)

FromJSON parses the JSON-encoded data and returns the Rison-encoded data that expresses the equal value.

Example

j := `[1,2.3,"str","-ing","true","nil",{"a":"b"},[7,8,9]]`
r, err := rison.FromJSON([]byte(j), rison.Rison)
if err != nil {
	panic(err)
}
fmt.Printf("%s\n", string(r))

Output:

!(1,2.3,str,'-ing',true,nil,(a:b),!(7,8,9))

func Marshal

func Marshal(v interface{}, m Mode) ([]byte, error)

Marshal returns the Rison encoding of v.

The object keys corresponding the struct fields can be specified in struct tag (not "rison" but) "json".

Example

v := exampleStruct{
	I: 1,
	F: 2.3,
	S: "str",
	B: true,
	P: nil,
	A: []int64{7, 8, 9},
	X: map[string]interface{}{"y": "Y"},
}
r, _ := rison.Marshal(&v, rison.Rison)
fmt.Println(string(r))

Output:

(a:!(7,8,9),b:!t,f:2.3,i:1,p:!n,s:str,x:(y:Y))

func Quote

func Quote(s []byte) []byte

Quote is like "net/url".QueryEscape but quotes fewer characters.

Example

s := "~!*()-_.,:@$'/ \"#%&+;<=>?[\\]^`{|}"
fmt.Println(rison.QuoteString(s))

Output:

~!*()-_.,:@$'/+%22%23%25%26%2B%3B%3C%3D%3E%3F%5B%5C%5D%5E%60%7B%7C%7D

func QuoteString

func QuoteString(s string) string

QuoteString is like "net/url".QueryEscape but quotes fewer characters.

func ToJSON

func ToJSON(data []byte, m Mode) ([]byte, error)

ToJSON parses the Rison-encoded data and returns the JSON-encoded data that expresses the equal value.

Example

r := "!(1,2.3,str,'ing',true,nil,(a:b),!(7,8,9))"
j, _ := rison.ToJSON([]byte(r), rison.Rison)
fmt.Printf("%s\n", string(j))

Output:

[1,2.3,"str","ing","true","nil",{"a":"b"},[7,8,9]]

func Unmarshal

func Unmarshal(data []byte, v interface{}, m Mode) error

Unmarshal parses the Rison-encoded data and stores the result in the value pointed to by v.

The object keys corresponding the struct fields can be specified in struct tag (not "rison" but) "json".

Example

r := "(i:1,f:2.3,s:str,b:!t,p:!n,a:!(7,8,9),x:(y:Y))"
var v exampleStruct
_ = rison.Unmarshal([]byte(r), &v, rison.Rison)
fmt.Printf("%+v\n", v)

Output:

{I:1 F:2.3 S:str B:true P:<nil> A:[7 8 9] X:map[y:Y]}

Types

type ErrType

type ErrType int

ErrType is an enum type of error

const (
	// EInternal is an error indicating an internal error occurred.
	EInternal ErrType = iota
	// EEncoding is an error indicating encoding failed.
	EEncoding
	// EEmptyString is an error indicating the string is empty.
	EEmptyString
	// EUnmatchedPair is an error indicating characters such as parentheses are not paired.
	EUnmatchedPair
	// EMissingCharacter is an error indicating necessary characters are missing.
	EMissingCharacter
	// EMissingCharacterAfterEscape is an error indicating there is no character after the escape character.
	EMissingCharacterAfterEscape
	// EExtraCharacter is an error indicating extra characters.
	EExtraCharacter
	// EExtraCharacterAfterRison is an error indicating there are extra characters after valid Rison.
	EExtraCharacterAfterRison
	// EInvalidLiteral is an error indicating an invalid literal was found.
	EInvalidLiteral
	// EInvalidCharacter is an error indicating an invalid character was found.
	EInvalidCharacter
	// EInvalidTypeOfObjectKey is an error indicating an invalid type object key was found.
	EInvalidTypeOfObjectKey
	// EInvalidStringEscape is an error indicating an invalid string escape was found.
	EInvalidStringEscape
	// EInvalidNumber is an error indicating an invalid number was found.
	EInvalidNumber
	// EInvalidLargeExp is an error indicating an upper case "E" is used as an exponent.
	EInvalidLargeExp
)

type Mode

type Mode int

Mode is an enum type to specify which Rison variation to use to encode/decode.

const (
	// Rison is the encoding/decoding mode covering all types.
	Rison Mode = iota
	// ORison is the special encoding/decoding mode for object type.
	ORison
	// ARison is the special encoding/decoding mode for array type.
	ARison
)

type ParseError

type ParseError struct {
	Child error
	Type  ErrType
	Args  []interface{}
	Src   []byte
	Pos   int
	// contains filtered or unexported fields
}

ParseError is an error type to be raised by parser

func (*ParseError) Error

func (e *ParseError) Error() string

func (*ParseError) ErrorInLang

func (e *ParseError) ErrorInLang(lang string) string

ErrorInLang returns the error message in specified language.

Example

r := "!("
_, err := rison.ToJSON([]byte(r), rison.Rison)
e, _ := err.(interface {
	ErrorInLang(lang string) string
	Langs() []string
})
fmt.Println(strings.Join(e.Langs(), ", "))
fmt.Println(e.ErrorInLang("en"))
fmt.Println(e.ErrorInLang("ja"))

Output:

en, ja
unmatched "!(" (at the end of string "!(" -> EOS)
"!(" が閉じていません (場所: 文字列終端: "!(" → EOS)

func (*ParseError) Langs

func (e *ParseError) Langs() []string

Langs returns supported languages.

func (*ParseError) Translate

func (e *ParseError) Translate(lang string)

Translate sets the language of the error message to be retrieved by Error().