$ jb name=json.bash creates=JSON

json.bash is a command-line tool and bash library that creates JSON.

$ jb name=json.bash creates=JSON dependencies:[,]=Bash,Grep
{"name":"json.bash","creates":"JSON","dependencies":["Bash","Grep"]}

$ # Values are strings unless explicitly typed
$ jb id=42 size:number=42 surname=null data:null
{"id":"42","size":42,"surname":"null","data":null}

$ # Reference variables with @name
$ id=42 date=2023-06-23 jb @id created@date modified@date
{"id":"42","created":"2023-06-23","modified":"2023-06-23"}

$ # Pull data from files
$ printf hunter2 > /tmp/password; jb @/tmp/password
{"password":"hunter2"}

$ # Pull data from shell pipelines
$ jb sizes:number[]@<(seq 1 4)
{"sizes":[1,2,3,4]}

$ # Nest jb calls
$ jb type=club members:json[]@<(jb name=Bob; jb name=Alice)
{"type":"club","members":[{"name":"Bob"},{"name":"Alice"}]}

$ # The Bash API can reference arrays and create JSON efficiently — without forking
$ source json.bash
$ out=people json name=Bob; out=people json name=Alice; sizes=(42 91 2)
$ id="abc.123" json @id @sizes:number[] @people:json[]
{"id":"abc.123","sizes":[42,91,2],"people":[{"name":"Bob"},{"name":"Alice"}]}

json.bash's one thing is to get shell-native data (environment variables, files, program output) to somewhere else, using JSON encapsulate it robustly.

Creating JSON from the command line or a shell script can be useful when:

• You need some ad-hoc JSON to interact with a JSON-consuming application
• You need to bundle up some data to share or move elsewhere. JSON can a good alternative to base64-encoding or a file archive.

It does no transformation or filtering itself, instead it pulls data from things you already know how to use, like files, command-line arguments, environment variables, shell pipelines and shell scripts. It glues together data from these sources, giving it enough structure to make the data easy to consume reliably in downstream programs.

It's something like a reverse tee — it pulls together data sources, using JSON to represent the aggregation. It's not an alternative to to data-processing tools like jq, rather it helps assemble JSON to send into JSON-consuming tools like jq.

Check the README on GitHub for more details & install instructions.