sloppy-json is a JSON parser that allows some slop:
- comma characters "
," aren't required in objects or arrays - not all strings need to be quoted
- new string form - back-tick strings prefixed with "
`" - comments are allowed
\o/
specifically:
- the sloppy-json parser will parse completely legal JSON correctly
- in those places you would require a comma character "
," in a JSON string, you may omit it - strings which are valid JavaScript identifiers (in the traditional sense,
not you "
π") don't need quotes at all - in addition to the double quote "
"" used by JSON, you can also use a single quote "'" - a new string form is introduced: back-tick strings (see below)
- comments are started with a "
#" or "//" prefix, and continue to the end of the line
Turns out there are a lot of cases where I need to enter some basic JSON stuff at the command-line, but it's always a pain what with all the ceremony of tradition JSON. Screw that, let's make it easier.
Strings are supported in a couple of fashions:
-
traditional double-quote strings using "
"" -
alternatively, using single-quote strings using "
'" -
unquoted strings, which are valid JavaScript identifiers, matching the following:
/^[A-Za-z_$][A-Za-z0-9_$]*$/are treated as their string value; eg
{ abc: def } === { "abc": "def" } -
back-tick strings: strings preceded by a back-tick character "
`" will take their value from all the characters to the next whitespace character; eg{ `a:b : `2019-07-03 } == { "a:b": "2019-07-03" }Back-tick strings also do no escaping; eg
{ `a:b : `\n } == { "a:b": "\\n" } // two chars - `\` and `n`, not linefeed
Double- and single-quoted strings are treated the same as in JSON with respect to escape characters and what-not.
Care should be taken using back-tick strings to ensure you have white-space terminating the string, eg
{ `a:b: `2019-07-03 } throws an error; the first token is "a:b:"
Numbers are treated the same as in JSON.
Strings and numbers are hard, so we make use of JSON.parse() to help.
For single- and double-quoted strings, a buffer is constructed with the contents
of the string, the string contents are wrapped in double-quotes, and then
passed to JSON.parse(). Bare double-quote characters used in single-quoted
strings will be replaced with the escaped form '\"'.
Back-tick strings are not processed like this, their verbatim contents are used as the string value.
For numbers, rather than deal with the arcanity, a number is recognized as
a token that starts with a digit, "-", or "+". Subsequent characters that
can be used in a number are then collected into the eventual string to
evaluate; that includes digits, "-", "+", ".", "e", "E". All these
characters are collected, and then passed to JSON.parse() to do the heavy
lifing. This means such oddities as 123.456.eee+-8+9 are treated as potential
number, but obviously will not parse as a valid value. But they aren't valid
JSON either, so it's more of an error reporting issue than anything.
In addition, the underscore character is just ignored in numbers, eg
123_456 === 123456
npm install pmuellr/sloppy-json
$ sloppy-json -E < package.json
{
name: "sloppy-json"
version: "1.0.0"
private: true
description: "deals with sloppy json"
license: MIT
author: "Patrick Mueller <pmuellr@apache.org> (https://github.com/pmuellr)"
homepage: "https://github.com/pmuellr/sloppy-json"
main: "sloppy-json.js"
bin: {
"sloppy-json": "./lib/cli.js"
}
scripts: {
build: "echo npm run build: TBD"
"deps-check": "node tools/deps-check.js"
"ts-check": "echo 'running ts-check'; tsc --allowJs --checkJs --noEmit --target ES5 --resolveJsonModule *.js"
"ts-check-skip": "echo 'skipping ts-check for the moment'"
ncu: "ncu -u --packageFile package.json"
serve: serve
standard: "echo 'running standard'; standard --verbose"
test: "npm run -s deps-check && npm run -s standard && npm run -s ts-check && npm run -s test:unit"
"test:unit": "echo 'running unit tests'; jest --colors"
watch: "nodemon --exec 'npm run -s test; node tools/log-time finished'"
}
repository: {
type: git
url: "https://github.com/pmuellr/sloppy-json.git"
}
bugs: {
url: "https://github.com/pmuellr/sloppy-json/issues"
}
standard: {
ignore: [
"tmp/**/*"
"docs/**/*"
]
}
jest: {
testEnvironment: node
verbose: true
}
nodemonConfig: {
ext: "js,ts,json"
}
dependencies: {}
devDependencies: {
"@types/jest": "^24.0.15"
"@types/node": "^12.0.10"
"dependency-check": "^3.3.0"
jest: "^24.8.0"
nodemon: "^1.19.1"
"random-seed": "^0.3.0"
serve: "^11.0.2"
standard: "^12.0.1"
typescript: "^3.5.2"
uuid: "^3.3.2"
}
}$ sloppy-json
sloppy-json 1.0.0
usage:
sloppy-json -d <value> - decode the value
sloppy-json -D <value> - decode the value indented
sloppy-json -e <value> - encode the value
sloppy-json -E <value> - encode the value indented
sloppy-json -h - print this help
sloppy-json -v - print the version
<value> should be a string to be encoded or decoded. If it is not
provided the value will be read from stdin.
Decoding will return a JSON respresentation of the sloppy JSON
passed as a parameter.
Encoding will return a sloopy JSON respresentation of the sloppy json
passed as a parameter.const SloppyJSON = require('sloppy-json')
const object = SloppyJSON.parse('{a:b c:d}') // { a: "b", c: "d" }
const boring = JSON.stringify(object) // { "a": "b", "c": "d" }
const sloppy = SloppyJSON.stringify(object) // "{a:b c:d}"
Like JSON.parse(), but accepts some slop, and no other arguments besides the
string to parse.
Like JSON.stringify(), but the second argument is ignored.
This package is licensed under the MIT license. See the LICENSE.md file for more information.
Awesome! We're happy that you want to contribute.
Please read the CONTRIBUTING.md file for more information.