A command-line utility for generating simple JSON documents without having to worry about escaping or quoting the inputs.

Without any inputs, jgen generates a blank JSON document:

[jcmcken@localhost]$ jgen
{}

Creating a simple document:

[jcmcken@localhost]$ jgen foo=bar
{"foo": "bar"}

Creating a nested document:

[jcmcken@localhost]$ jgen foo.bar=baz
{"foo": {"bar": "baz"}}

Creating a more complex nested document:

[jcmcken@localhost]$ jgen foo.bar=baz foo.foo=bar
{"foo": {"foo": "bar", "bar": "baz"}}

Working with arrays:

[jcmcken@localhost]$ jgen foo.bar=a,b,c
{"foo": {"bar": ["a", "b", "c"]}}

Use non-strinrg data types (see section below for details):

[jcmcken@localhost]$ jgen foo=true bar=null baz=3.141
{"baz": 3.141, "foo": true, "bar": null}

Some additional command-line options exist for jgen.

• -p/--pretty-print : "Pretty-print" the resulting document
• -y/--yaml : If PyYAML library is available, print the document as YAML instead of JSON
• --prefix : Prepend a prefix to each top-level key in the resulting data
• -s/--stringify : Disable all type coercion (see the Data Types section below)

Currently, jgen does the following type conversions on data values (but not keys!):

• If a value is either of the strings "true" or "false", the resulting value will be the boolean true or false (respectively).
• If a value is the string "null", the resulting value will be the JSON null type.
• If a value is "integer-like" (that is if the mapping lambda x: str(int(x)) is the identity mapping), the int function will be applied to the value.
• If a value is "float-like" (that is if the mapping lambda x: str(float(x.rstrip("0"))) is the identity mapping), the float function will be applied to the value.

Note that the JSON specification does not allow data keys to be anything except the string type.

This type coercion behavior can be disabled by passing the -s/--stringify command line option.

Because jgen relies on recursive hash merges to build up the JSON document, you need to be careful about how you specify the data.

For example, if the input string is foo.bar=baz foo=baz, jgen will take foo.bar=baz and generate the document {"foo": {"bar": "baz"}}. Next it will take the string foo=baz and generate the document {"foo": "baz"}. Since the recursive merge of these two documents is {"foo": "baz"}, some of the data you specified at the command line is lost.