Re: VarStructor II -- Abbreviation tool

[jeffa]

Since you give no usage what so ever, i can only speculate just exactly what you want to output to look like. And more importantly, what the "rules" are. But i can say that you have a lot of needless code there. Try this instead:

use Data::Dumper;

my (%symbol,@old,@new);
my $max          = 10;
my $ws_replace   = '_';
my $letters_only = 1;

while (<DATA>) {
   chomp;
   push @old, $_;
   $_ = substr $_,0,10 if $max;
   s/\s+/$ws_replace/g if $ws_replace;
   s/\W+//g            if $letters_only;
   $_ .= $.            if $symbol{$_}++;
   push @new, $_;
}

print Dumper \@old, \@new;

my %compare;
@compare{@old} = @new;

print Dumper \@compare;

__DATA__
Line one
Line two
xxxxxxxxxxx
Another line
xxxxxxxxxxx
Lines end here not
xxxxxxxxxxx
Fourth line (used to be)
Line five
Lines end here
Lines end here too

And read The Dynamic Duo --or-- Holy Getopt::Long, Pod::UsageMan!

[Wassercrats]

No need to speculate, since the script was set up as a demo, as I said. If you ran it, you would see:

Line_o
Line_t
x1
A
x2
L1
x3
F
Line_f
L2
L3

Your script produces:

'Line_one',
'Line_two',
'xxxxxxxxxx',
'Another_li',
'xxxxxxxxxx5',
'Lines_end_',
'xxxxxxxxxx7',
'Fourth_lin',
'Line_five',
'Lines_end_10',
'Lines_end_11',

Which follows a different numbering scheme, using higher numbers. Higher numbers increases the chances of them having more digits, which would require eliminating more of the semantically meaningful letters if the string length is to stay the same. This probably wouldn't matter much to me, but I already have my arguably better way, which doesn't require a module.

[jeffa]

No. There is every need to speculate with your code, Wassercrats. My code was not meant to replace yours, or even reproduce 100% of the functionality. It was to show you that you can do what you are trying to do with much less code! How about this instead: it truncates the string to the first two letters (lowercased) and adds incremented numbers for each collision. And this time i won't require you to use a module!

my (%symbol,@old,@new);

while (<DATA>) {
   chomp($_ = substr lc($_),0,2);
   push @old, $_;
   $symbol{$_}++;
   push @new, $_ . $symbol{$_};
}

print "$_\n" for @new;

__DATA__
Line one
Line two
xxxxxxxxxxx
Another line
xxxxxxxxxxx
Lines end here not
xxxxxxxxxxx
Fourth line (used to be)
Line five
Lines end here
Lines end here too

Of course, while i find this to be quite silly -- it is more intuitive than what you have. Besides, you already have your arguable better way.

[Wassercrats]

I didn't try it, but it sounds like it would work fine for me. In fact, I haven't even implemented my version into the script I want to use it in, but what I'm currently using works too. It's just that what I'm currently using puts no limit on the length of the string (used as a variable name), which could look sloppy, and I tried creating something a little "smarter" than your version. But since my version needs only a slight tweak to function as Text::Abbrev (with added features), I think I'll eventually add that functionality and make it a module.

By the way, if I made my script a module and didn't include its code in the line count, my code could be even shorter than your module using code.

[jdporter]

It's not clear to me why you have the lines

Line one
Line two
Line five

reduce to

L1
L2
L3

but the lines

Lines end here not
Lines end here
Lines end here too

reduce to

Line_o
Line_t
Line_f

I mean, I understand that the latter are treated differently because they're identical within the Max_Length span, whereas the shorter ones aren't. But still... why should that matter? Wouldn't it have been acceptable -- aye, preferable -- to have the latter reduce to (e.g.)

L4
L5
L6

???

Thanks,
jdporter

[Wassercrats]

You got the output example backwards.

Line one
Line two
Line five

reduces to:

Line_o
Line_t
Line_f

and:

Lines end here not
Lines end here
Lines end here too

reduces to:

L1
L2
L3

The script is intended to create the shortest possible unique string, with a given maximum number of characters from the beginning of the string. When two strings are identical, a number must be appended and since the text is of no use in distinguishing those strings, and the shortest possible unique string contains just one (actually, zero) characters, that's how many characters are kept.

Whether sticking to that produces the most useful function is another issue, but it does produce the shortest string and puts value on the inclusion of text when the text isn't identical.

I will be adding an option to have identical text that's appended with a number contain the maximum number of characters instead of just one.