how to parse binary files in Clojure

Question

What is the cleanest way to parse binary data in clojure? I need to be able to read/write equally cleanly to a file or a socket.

something like:

  (read-data source-of-data) 
  => { :index 42 , :block-size 4 , data-size: 31415, :data (1 2 3 4 ...)}

and the reverse for putting data back. It would be really great to somehow define the structure once and have the read and write functions use the same definition.

Answer 1

Gloss makes it easy to define binary formats at the byte level for both reading and writing.

(defcodec example-codec
  [:id       :uint32
   :msg-type (enum :byte {:a \A, :b \B})
   :status   (string :ascii :length 11)])

(def buffer (byte-array 16))

(.read (input-stream "filename.bin") buffer)
(decode example-codec buffer)

(encode example-codec {:id 42, :msg-type :a, :status "A-OKAY"})

The bit-map function allows bit level formats, but the number of bits defined must be divisible by 8 so the bytes still line up.

There's also byte-spec.

Answer 2

Since Clojure can use native Java functions, why not use those? A quick Googling along those lines gives: http://gnuvince.wordpress.com/2009/01/29/reading-binary-data-in-clojure/