Skip to content

Latest commit

 

History

History
165 lines (124 loc) · 3.92 KB

README.md

File metadata and controls

165 lines (124 loc) · 3.92 KB

eq (edn query)

eq (edn query) is a command line tool for edn processing and pretty printing. It is inspired by jq.

Usage

eq reads edn data from stdin and pretty prints it to stdout

$ cat test.edn
{:foo "Lorem ipsum dolor sit amet" :baz [1 1 2 3 5 8 13] :quux #inst "2016-06-11" :foo.bar/baz 3.14159 :tagged #foo/bar [1 2] :booleans #{true false} :nested-maps {:works "fine"}}
$ cat test.edn | eq
{:foo "Lorem ipsum dolor sit amet"
 :baz [1 1 2 3 5 8 13]
 :quux #inst "2016-06-11"
 :foo.bar/baz 3.14159
 :tagged #foo/bar [1 2]
 :booleans #{true false}
 :nested-maps {:works "fine"}}

If you don't want syntax coloring you can pass the --no-colors flag

$ cat test.edn | eq --no-colors

eq is designed to work well with other unix tools like curl and the startup time should be instantaneous.

Queries

Queries are written in edn and all edn values represent some kind of query. "Atomic" values are self-evaluating and ignore the input:

$ echo '"I will be ignored"' | eq ':foo'
:foo

$ cat test.edn | eq '42'
42

The (id) query returns the input unchanged (this is the default query if none is specified):

$ echo '"Hello, world!"' | eq '(id)'
"Hello, world!"

(get key) looks up key in the input by key or index.

$ cat test.edn | eq '(get :baz)'
[1 1 2 3 5 8 13]
$ echo '[:a :b :c]' | eq '(get 2)'
:c

(-> q1 q2) pipes the output of q1 to the input of q2.

$ cat test.edn | eq '(-> (get :nested-maps) (get :works))'
"fine"

[q1 q2 ... qn] will apply each query in the vector and collect the results in an output vector. For example [:foo :bar :baz] is a query with three subqueries which all happen to evaluate to themselves.

$ cat test.edn | eq '[(get :foo) (get :baz)]'
["Lorem ipsum dolor sit amet" [1 1 2 3 5 8 13]]

$ cat test.edn | eq '[(-> (get :nested-maps) (get :works)) :literal (get :booleans)]'
["fine" :literal #{true false}]

Similarly to wrapping queries in a vector, it's also possible to wrap them in a map:

{kq1 vq1
 kq2 vq2
 ...
 kqn vqn}

Each of the sub queries are run on the input, and the result is collected in a map. Some examples:

$ cat test.edn | eq '{:foo :bar}'
{:foo :bar}

$ cat test.edn | eq '{:foo (get :foo)}'
{:foo "Lorem ipsum dolor sit amet"}

$ cat test.edn | eq '{(-> (get :nested-maps) (get :works)) (get :quux) :foo [1 2 (get :booleans)]}'
{"fine" #inst "2016-06-11"
 :foo [1 2 #{true false}]}

(map q) will apply the query q to each element of the input. The output will be separate edn objects (i.e. many objects will be printed to stdout, separated by newline):

$ cat test.edn | eq '(map (id))'
[:foo "Lorem ipsum dolor sit amet"]
[:baz [1 1 2 3 5 8 13]]
[:quux #inst "2016-06-11"]
[:foo.bar/baz 3.14159]
[:tagged #foo/bar [1 2]]
[:booleans #{true false}]
[:nested-maps {:works "fine"}]

It is often useful to collect the results produced by map in a collection. The following example shows how to find all the keys in an input map:

$ cat test.edn | eq '[(map (get 0))]'
[:foo :baz :quux :foo.bar/baz
 :tagged :booleans :nested-maps]

map will also work on vectors, lists and sets:

$ echo '[{:a 1 :b 2} {:a 3 :b 4}]' | eq '(map (get :a))'
1
3

Build instructions

No pre-built binaries are available yet. In the future I hope to distribute eq via package managers such as Homebrew but for now you need to build eq manually.

  • A recent version of opam is required. Install eq with
$ opam pin add eq https://github.com/jonase/eq.git

Package eq does not exist, create as a NEW package ? [Y/n] y
eq is now git-pinned to https://github.com/jonase/eq.git
[… Installation output excluded …]

eq is now in the opam bin directory (for example ~/.opam/4.03.0/bin).

License

MIT Licensed. Copyright © 2016 Jonas Enlund