PARI's commands

[CRGreathouse]

Originally Posted by science_man_88

This sounds too hard, too many things to look for.

It's pretty easy in a language like Perl, where the answer (for Pari!) would be a single, short line. But Pari has almost no string-manipulation functions, so it's rather hard to do there.

[science_man_88]

http://rosettacode.org/wiki/LZW_compression is described both on wikipedia and a book I own. Doesn't sound too hard, Of course I said that about most things lol.

[CRGreathouse]

Go for it, then.

[science_man_88]

Originally Posted by CRGreathouse

Go for it, then.

according to my book it doesn't concat anything into the library/dictionary until it hits it the second time in a string, also it talks of a simplified case of all upcase I'll have to make the starting dictionary twice as long for both cases and 52 variables to track them through 2 times i guess I must be over complicating things, doh.

[CRGreathouse]

Maybe try something easier first, then? Perhaps

http://rosettacode.org/wiki/Non-decimal_radices/Input
or
http://rosettacode.org/wiki/Write_fl...to_a_text_file

[science_man_88]

Originally Posted by CRGreathouse

Maybe try something easier first, then? Perhaps

http://rosettacode.org/wiki/Non-decimal_radices/Input

My plan for this is no less complex now.



1)Create 2, 36 index vectors then shorten them for the respective bases.
2)Read a Vec of the number to convert and use the index to add into a decimal form variable.
3)take that decimal form and use variable%b2 to get each index to return later.
4) return it.

[science_man_88]

trouble on number 1)

***   syntax error, unexpected ',': ...mb1,b1,b2)=b1digs["0","1","2","3","4","5","6"
                                                                ^--------------------

never mind forgot the equals sign.

[3.14159]

Wow. I see we've collected 2000 posts.

[science_man_88]

Originally Posted by 3.14159

Wow. I see we've collected 2000 posts.

I bet if we took out all of the useless ones we'd not even have half that lol.

convert(numb1,b1,b2)= b1digs=["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"];b1digs=vector(b1,n,b1digs[n]);b2digs=["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"];b2digs=vector(b2,n,b2digs[n]);numb1=Vec(Str(numb1));a=0;c=[];forstep(y=#numb1,1,-1,for(x=1,#b1digs,if(numb1[y]==b1digs[x],a=a+(x-1)*b1^(#numb1-y))));until(floor(a/b2)==0,c=concat(b2digs[a%b2+1],c));c;

most useless Pari code ever lol.

[CRGreathouse]

Originally Posted by science_man_88

convert(numb1,b1,b2)= b1digs=["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"];b1digs=vector(b1,n,b1digs[n]);b2digs=["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"];b2digs=vector(b2,n,b2digs[n]);numb1=Vec(Str(numb1));a=0;c=[];forstep(y=#numb1,1,-1,for(x=1,#b1digs,if(numb1[y]==b1digs[x],a=a+(x-1)*b1^(#numb1-y))));until(floor(a/b2)==0,c=concat(b2digs[a%b2+1],c));c;

most useless Pari code ever lol.

You're defining the same (long) vector twice; why not avoid that?

convert(numb1,b1,b2)={
  my(B=["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"],b1digs=vector(b1,n,B[n]),b2digs=vector(b2,n,B[n]),a=0,c=[]);
  numb1=Vec(Str(numb1));
  forstep(y=#numb1,1,-1,
    for(x=1,#b1digs,
      if(numb1[y]==b1digs[x],
        a=a+(x-1)*b1^(#numb1-y)
      )
    )
  );
  until(floor(a/b2)==0,
    c=concat(b2digs[a%b2+1],c)
  );
  c
};

It looks like you could avoid using both b1digs and b2digs by simply pulling from the general vector B each time. Usual comments on a = a + foo vs. a += foo and floor(a/foo) vs. a\foo appy.

[science_man_88]

Originally Posted by CRGreathouse

You're defining the same (long) vector twice; why not avoid that?

convert(numb1,b1,b2)={
  my(B=["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"],b1digs=vector(b1,n,B[n]),b2digs=vector(b2,n,B[n]),a=0,c=[]);
  numb1=Vec(Str(numb1));
  forstep(y=#numb1,1,-1,
    for(x=1,#b1digs,
      if(numb1[y]==b1digs[x],
        a=a+(x-1)*b1^(#numb1-y)
      )
    )
  );
  until(floor(a/b2)==0,
    c=concat(b2digs[a%b2+1],c)
  );
  c
};

It looks like you could avoid using both b1digs and b2digs by simply pulling from the general vector B each time. Usual comments on a = a + foo vs. a += foo and floor(a/foo) vs. a\foo appy.

apply* only if you only go up to index b1, or b2 for each.