json

import "json" as json

A pure-Argon JSON encoder/decoder (see stdlib/json/init.ar).

json.stringify(value, pretty=false)

Converts an Argon value into a JSON string.

term.log(json.stringify(null))                 # "null"
term.log(json.stringify(true))                  # "true"
term.log(json.stringify("hello"))               # "\"hello\""
term.log(json.stringify([1, 2, 3]))              # "[1,2,3]"
term.log(json.stringify({"a": 1, "b": 2}))       # "{\"a\":1,\"b\":2}"
term.log(json.stringify({"a": [1, 2, 3]}))       # "{\"a\":[1,2,3]}"

Dictionary keys that aren’t strings are coerced to their string form:

term.log(json.stringify({1: "a"}))       # {"1":"a"}
term.log(json.stringify({true: "a"}))    # {"true":"a"}

Non-ASCII characters are escaped using \uXXXX (including surrogate pairs for characters outside the Basic Multilingual Plane, like emoji):

term.log(json.stringify("🇬🇧"))   # "\"\\uD83C\\uDDEC\\uD83C\\uDDE7\""

Circular references raise a ValueError:

let a = []
a.append(a)
json.stringify(a)   # throws ValueError("circular JSON serialisation")

json.parse(s)

Parses a JSON string into Argon values (null/true/false, numbers, strings, arrays, and dictionaries):

term.log(json.parse("null"))                  # null
term.log(json.parse("[1,2,3]"))               # [1, 2, 3]
term.log(json.parse("{\"a\":1,\"b\":2}"))     # {a: 1, b: 2}

Whitespace around tokens is tolerated:

term.log(json.parse(" [ 1 , 2 , 3 ] "))   # [1, 2, 3]

All standard JSON string escapes are supported when parsing, including \n, \t, \\, \/, \b, \f, \r, and \uXXXX (with correct surrogate-pair handling for characters outside the Basic Multilingual Plane).

Errors

Malformed JSON raises a ValueError, for example:

json.parse("")                 # ValueError — empty input
json.parse("{")                # ValueError — unterminated object
json.parse("[1,2,]")           # ValueError — trailing comma
json.parse("\"unterminated")   # ValueError — unterminated string
json.parse("tru")              # ValueError — truncated keyword
json.parse("1 2")              # ValueError — trailing content after value
json.parse("\"\\uD800\"")      # ValueError — lone (unpaired) surrogate

Round-tripping

json.parse(json.stringify(value)) reproduces value for all JSON-representable data — this is exercised extensively in tests/json_test.ar, which is also a good reference for exact expected output formatting.

Next

• file