Encoding/decoding with custom character sets

[calh]

Hello,

This is probably a dumb/simple question, but I can't figure out how to decode a string with a custom character set. Here's a quick example:

1.9.3-p545 :001 > BASE62_ALPHABET = (0..9).to_a + ('a'..'z').to_a + ('A'..'Z').to_a
 => [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, "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", "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"] 
1.9.3-p545 :002 > 809145531659995.b(62).to_s(BASE62_ALPHABET[0,62])
 => "3HLszsQsP" 
1.9.3-p545 :003 > "3HLszsQsP".b(BASE62_ALPHABET[0,62])
NoMethodError: undefined method `<' for #<Array:0x00000006559a78>
1.9.3-p545 :008 > "3HLszsQsP".b(62).to_i
 => 716130955509941 

What am I doing wrong here?

Thank you!

[trans]

Ha! That's a damn good question!

Well, this works at least:

b10 = Radix::Base.new(Radix::BASE::B10)
b10.convert("3HLszsQsP", BASE62_ALPHABET)
=> "809145531659995"

But you are right the convenience extensions should allow it too. I'll have to fix that.

[trans]

Ah, I missed the very last part -- you are getting a incorrect result in addition to the undefined method error. The incorrect result is b/c the first 10 entries in your custom character set are not strings. Change to:

  BASE62_ALPHABET = (0..9).map{|x| x.to_s} + ('a'..'z').to_a + ('A'..'Z').to_a

And I think that fixes that issue. Not sure why it doesn't work without doing that. It may be treating the numbers as ASCII codes. Will have to look into it and decided what the proper behavior should be. If you have any suggestions in that regard, please let me know.

With regards to the undefined method error, I think I figure out the fix, so the next release should have that taken care of at least.

[calh]

Hmm, it seems to still do the same thing:

1.9.3-p545 :001 > BASE62_ALPHABET = (0..9).map{|x| x.to_s} + ('a'..'z').to_a + ('A'..'Z').to_a
 => ["0", "1", "2", "3", "4", "5", "6", "7", "8", "9", "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", "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"] 
1.9.3-p545 :004 > 809145531659995.b(62).to_s(BASE62_ALPHABET[0,62])
 => "3HLszsQsP" 
1.9.3-p545 :005 > "3HLszsQsP".b(BASE62_ALPHABET[0,62])
NoMethodError: undefined method `<' for #<Array:0x00000001d33a78>

However, your other suggestion does seem to work fine:

1.9.3-p545 :002 > b10 = Radix::Base.new(Radix::BASE::B10)
 => #<Radix::Base:0x00000002371bd8 @chars=["0", "1", "2", "3", "4", "5", "6", "7", "8", "9"], @base=10, @values={"0"=>0, "1"=>1, "2"=>2, "3"=>3, "4"=>4, "5"=>5, "6"=>6, "7"=>7, "8"=>8, "9"=>9}> 
1.9.3-p545 :004 > b10.convert("3HLszsQsP", BASE62_ALPHABET)
 => "809145531659995" 

So I'm fine with this answer, but if you want to fix this bug that I found that would be super cool.

Thanks!

[calh]

As a side note... I think the real thing to blame is that there seems to be no standard for the Base 62 alphabet! And it's a crapshoot on the glyph order for each library's implementation.

[trans]

I just pushed the fix about 10 minutes ago. Might want to try again. I'll release the new gem here in a few minutes.

Btw, you can remove a small bit of redundancy in your example with:

809145531659995.b(BASE62_ALPHABET[0,62]).to_s

Also, there is a standard base-62 built-in as Radix::BASE::B62. It is exactly the same array that you used expect the capital letters come before the lowercase letters b/c ASCII order. (Not quite a universal standard but it helps.)

[trans]

The gem has been pushed. Hopefully this is now all fixed and can be closed. Let me know. And thanks for reporting it!

[calh]

Works perfectly! Thank you!