jsongrep 0.1.1

Examples

Let’s start with a real-world example.

Let’s grab the even tweets from the last 10 pulled from Twitter’s JSON feed http://apiwiki.twitter.com/Twitter-REST-API-Method:-statuses-public_timeline

$ curl -s 'http://twitter.com/statuses/public_timeline.json' | jsongrep '[02468].text'
ARGHHHHHH. facebook is being gay
5-5 in the darts between Barney and Whitlock. Amazing. #darts
I wonder if I'm still located on 5th ave?
Estou de volta  a internet .... Essa chuva n para !   Estou de boa com a minha familia .

Yeah, that’s just about what I expected.

Now suppose you have a JSON document like the one in tests/ongz.json which looks like this:

{
    "bah" : {
        "feh" : true,
        "foo" : 3,
        "lah" : "songz"
    },
    "blah" : {
        "lol" : "gongz"
    },
    "arr" : [
        "a", "b", "c", "d", "e",
        "f", "g", "h", "i", "j",
        "k", "l", "m", "n", "o",
        "p", "q", "r", "s", "t",
        "u", "v", "w", "x", "y",
        "z"
    ]
}

jsongrep will let you match structural patterns, where . (dot) separates nested properties.

Let’s glob on property names:

$ jsongrep 'b*.l*' tests/ongz.json
gongz
songz

Works on arrays, too:

$ jsongrep 'arr.?' tests/ongz.json
a
b
c
d
e
f
g
h
i
j

Note that we’re still globbing, even though arrays have numeric indices:

$ jsongrep 'arr.2?' tests/ongz.json
u
v
w
x
y
z

If you specify a JSON subtree, that’s what you get back:

$ jsongrep 'bah' tests/ongz.json
{"foo": 3, "lah": "songz", "feh": true}

Syntax

jsongrep currently supports normal shell glob patterns within property names:

?       Matches any one character
*       Matches any number of characters within a field
[seq]   Matches any of the characters in seq
[!seq]  Matches any of the characters not in seq

Dot is the field separator.

Usage

Usage: jsongrep [options] [PATTERN | -e PATTERN [-e PATTERN ...]] [FILE]

Parses JSON data structurally to select a subset of data.

Options:
  --version             show program's version number and exit
  -h, --help            show this help message and exit
  -e PATTERNS, --pattern=PATTERNS
                        Additional patterns to match.
  -D, --detect-encoding
                        Attempts to detect the character encoding of input if
                        LC_TYPE and file.encoding provide insufficient hints.
                        (Slow) [default: False]

TODO

• Support star-star (**) non-greedy matches of spanning subgraphs

• Support unicode, escapes in patterns

• Support extended regexps

• Support no-pattern-matching lookup

• Options:

  • Property separator in patterns (. by default)

  • Output separator (newline by default)

  • Quote string values in output?

  • 1/0 vs true/false for bool values in output?

Feedback

Open a ticket at http://github.com/dsc/jsongrep , or send me an email at dsc@less.ly .