in reply to Bracketing Substring(s) in the String
I'd like to propose a different approach, but it relies on one feature of the problem that you haven't made explicit: that all the strings you're dealing with, and all the substrings, are uppercase letters only. If so, then what we can do is make each match case-insensitive and lowercase any region that's found to match. Then, simply place brackets at uppercase/lowercase boundaries, and we're set:
And this passes all four of your test cases.sub put_bracket { my ($str,$ar) = @_; foreach my $subs ( @$ar ) { my $lsub = lc($subs); s/$lsub/$lsub/ig; } # add brackets $str =~ s/(?:(?<=[A-Z])|^)(?=[a-z])/[/g; $str =~ s/(?<=[a-z])(?:(?=[A-Z])|$)/]/g; # re-uppercase $str = uc($str); print "$str\n"; return ; }
But wait! Here's a test case that this code - and all the other code posted to this thread that I've tried - fails on:
Oh dear, we didn't account for the case when the same substring overlaps itself. How are we going to handle this? Fortunately, it's not impossible. What we need to have happen is basically "repeat the substitution, but don't change anything already all in lowercase, until no changes are made". Fortunately, perl's got a syntax for that:my $s5 ='CCACCACCACCTGTC'; my @a5 = qw(CCACC); put_bracket($s5,\@a5); # should be [CCACCACCACC]TGTC
That new line says:sub put_bracket { my ($str,$ar) = @_; foreach my $subs ( @$ar ) { my $lsub = lc($subs); do {} while ($str =~ s/(?!$lsub)(?i:$lsub)/$lsub/g); } # add brackets $str =~ s/(?:(?<=[A-Z])|^)(?=[a-z])/[/g; $str =~ s/(?<=[a-z])(?:(?=[A-Z])|$)/]/g; # re-uppercase $str = uc($str); print "$str\n"; return ; }
-- @/=map{[/./g]}qw/.h_nJ Xapou cets krht ele_ r_ra/; map{y/X_/\n /;print}map{pop@$_}@/for@/
|
|---|