So, just for my own perl coding practice, I wrote the following code that will do what you want (without explicit debugging though...). This programme will take in one file, where the first line is the sequence and the second line is the structure, and generate the required output

#!/usr/bin/perl
use warnings;
use strict;
open FILE, "<", $ARGV[0] or die $!;
my $count = 0;
my $seq;
my  $structString;
while (<FILE>) {
    if($count == 0){
        $seq = ($_);
        $count = 1;
    }elsif($count==1){
        $structString = ($_);
        last;
    }
}
close FILE;
my @seqCharacter = split //, $seq;
my @structChar = split //, $structString;
my $loopCount=1;
my $stemCount=1;
my $currentLoop="";
my @stemStore;
my @stemName;
my $prevChar="";

for(my $i = 0; $i < @structChar; $i=$i+1){
    if($prevChar eq ""){ #The first character
        if($structChar[$i] eq "."){
            $currentLoop = $seqCharacter[$i];
            $prevChar = ".";
        }elsif($structChar[$i] eq "("){
            $stemStore[0]=$seqCharacter[$i];
            $stemName[0] = "stem".$stemCount;
            $stemCount=$stemCount+1;
            $prevChar="(";
        }else{
            print "Undefined input, structure must start with either ( or .\n";
            die;
        }
    }elsif($prevChar eq "." && $structChar[$i] eq "."){ #Continuation
        $currentLoop=$currentLoop.$seqCharacter[$i];
    }elsif($prevChar eq "." && $structChar[$i] eq "("){ #Encounter stem
        print "loop".$loopCount.": ".$currentLoop."\n";
        $loopCount = $loopCount+1;
        $currentLoop = "";
        $prevChar = "(";
        my $size = @stemStore;
        $stemStore[$size]=$seqCharacter[$i]; #we have a new stem
        $stemName[$size] = "stem".$stemCount;
        $stemCount = $stemCount+1;
    }elsif($prevChar eq "." && $structChar[$i] eq ")"){
        if(@stemStore == 0){
            print "Unmatched ) please check your input\n";
            die;
        }else{
            print "loop".$loopCount.": ".$currentLoop."\n";
            $loopCount = $loopCount+1;
            $currentLoop = "";

            $stemStore[@stemStore-1]=$stemStore[@stemStore-1].$seqCharacter[$i];    #Add this to the last loop so we can close it
            $prevChar = ")";
        }
    }elsif($prevChar eq "(" && $structChar[$i] eq "("){
        $stemStore[@stemStore-1]=$stemStore[@stemStore-1].$seqCharacter[$i]; #Continue
    }elsif($prevChar eq "(" && $structChar[$i] eq "."){
        $prevChar = ".";
        $currentLoop = $seqCharacter[$i];

    }elsif($prevChar eq "(" && $structChar[$i] eq ")"){
        $prevChar = ")";
        $stemStore[@stemStore-1]=$stemStore[@stemStore-1].$seqCharacter[$i];
    }elsif($prevChar eq ")" && $structChar[$i] eq ")"){
        $stemStore[@stemStore-1]=$stemStore[@stemStore-1].$seqCharacter[$i]; #Add to the last loop
    }elsif($prevChar eq ")" && $structChar[$i] eq "."){
        $prevChar = ".";
        print $stemName[@stemName-1].": ".$stemStore[@stemStore-1]."\n";
        pop(@stemName);
        pop(@stemStore);
    }elsif($prevChar eq ")" && $structChar[$i] eq "("){
        $prevChar = ")";
        print $stemName[@stemName-1].": ".$stemStore[@stemStore-1]."\n";
        pop(@stemName);
        pop(@stemStore);
    }
}
#output remaining
if($currentLoop ne ""){
    print "loop".$loopCount.": ".$currentLoop."\n";
}elsif(@stemName != 0 && $prevChar eq ")"){
    print $stemName[@stemName-1].": ".$stemStore[@stemStore-1]."\n";
    pop(@stemName);
    pop(@stemStore);
}else{
    print "There are unmatched stem remaining\n";
}

forgi (github page) is a Python package for working with RNA secondary structures. I've used it in the past for tasks like this.

Here's an example. By the way, might want to check your expected output against the example sequence you provided.

import forgi.graph.bulge_graph as fgb
bg = fgb.BulgeGraph()
bg.from_fasta(
"""\
>demo
UGCCCUAGUACGAGAGGACCGGAGUG
...(((........))).........
""")
for d in bg.defines:
    print d, bg.get_define_seq_str(d)

Here's the output. Notice it keeps the two sides of the stem separate, and the 5' and 3' "stems" are listed separately.

f1 ['UGC']
h0 ['AGUACGAG']
s0 ['CCU', 'AGG']
t1 ['ACCGGAGUG']

Edit: forna by the same author is useful for visualizing structures and is helpful for debugging and troubleshooting this sort of analysis.

Edit 2: In response to the comment asking about providing a string variable to from_fasta:

It's not documented, but after playing around it looks like this method requires 1) the record must start with ">" -- not even a newline is allowed, and 2) a trailing newline at the end of the structure line. So if you have your strings sequence and structure, you can build a string that from_fasta will accept like this:

text = '\n'.join(['> ', sequence, structure, '\n'])
bg.from_fasta(text)

or if you want to include a sequence ID:

text = '\n'.join(['\n', '> ' + seqid, sequence, structure, '\n'])
bg.from_fasta(text)