This code is really confusing, please help me out!

Hi Everyone!

I am new to Elixir. I followed ElixirSchool and Elixir course at Pragmatic Studio. I also developed some toy Phoenix application. I mean I am familiar with the syntax (except the typespecs).

Right now i’m following The little Elixir & OTP Guidebook by @bentanweihao

The following code is really confusing. The writer has tried to explain it, but I didn’t get it, even if I am familiar with the Elixir syntax.

defmodule ID3Parser do

  def parse(file_name) do
    case File.read(file_name) do

      {:ok, mp3} ->
        mp3_byte_size = byte_size(mp3) - 128
       
        # from here
        << _ :: binary-size(mp3_byte_size), id3_tag :: binary >> = mp3

        << "TAG", title   :: binary-size(30),
                  artist  :: binary-size(30),
                  album   :: binary-size(30),
                  year    :: binary-size(4),
                  _rest   :: binary >>       = id3_tag
        # up to here


        IO.puts "#{artist} - #{title} (#{album}, #{year})"

      _ ->
        IO.puts "Couldn't open #{file_name}"

    end
  end
end

I have marked the points with comment. The code which is confusing.
Would someone explain it in more detail?

Also when I ran the program on a sample mp3 file, it didn’t give me the result as

Some Artist - Some Song (Some Album, 2020) 
:ok

but as

Some Artist^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@ - Some Song^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@ (Some Album^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@, 2020) 
:ok

(The sample file I had didn’t have metadata, so I added the metadata (title, artist, album and year) through VLC).

Thank you!

Hello and welcome,

This strange code is binaries pattern matching. You might learn more here…

It is very powerful when You are dealing with external protocols, because You can easily decode.

More on this…

BTW this code is working for me…

3 Likes

Welcome @kkhan,

If you do an IO.inspect on the id3_tag, you would be getting the following value(Some random MP3):

<<84, 65, 71, 67, 108, 111, 115, 101, 32, 84, 111, 32, 77, 101, 0, 0, 0, 0, 0,
  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 69, 108, 108, 105, 101, 32, 71, 111,
  117, 108, 100, 105, 110, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
  66, 105, 108, 108, 98, 111, 97, 114, 100, 32, 72, 111, 116, 32, 49, 48, 48,
  32, 83, 105, 110, 103, 108, 101, 115, 32, 67, 104, 97, 114, 50, 48, 49, 57,
  80, 77, 69, 68, 73, 65, 32, 78, 69, 84, 87, 79, 82, 75, 0, 0, 0, 0, 0, 0, 0,
  0, 0, 0, 0, 0, 0, 0, 0, 28, 255>>

We are doing a pattern matching of the metadata for n sized binaries (binary-size(n)), the zero’s in the above metadata gets printed as ^@. You can check that using IO.puts(<<0>>).

3 Likes

You have already got some answers which point you to the point in the docs and throw some concept names at you.

<< _ :: binary-size(mp3_byte_size), id3_tag :: binary >> = mp3

This line splits the input binary into 2 parts, the sound data (which instantly gets dropped) and the id3-tag data which gets bound to id3_tag.

<< "TAG", title   :: binary-size(30),
          artist  :: binary-size(30),
          album   :: binary-size(30),
          year    :: binary-size(4),
          _rest   :: binary >>       = id3_tag

This splits the tag data up even further.

It matches in order to the sequence of bytes resembling the UTF-8 sequence “TAG”, 3 30 byte sequences and a 4 byte long sequence. The remainder of the tag data is then dropped.

The fixed length sequences are bound to title, artist, album and year.

As those individual fields are “fixed length”, they have to be filled with something in the binary blob. Probably due to having its origin in C the 0 byte has been choosen.

You have to strip it using String.trim_trailing/2, eg. String.trim_trailing(title, <<0>>)

3 Likes