From: Joseph McDonald <joe@...>
Date: 2001-10-25T08:49:54+09:00
Subject: [ruby-talk:23239] Re: splitting a string with nested elements


>> string = "start {{blah {{ }} outside\n{{ {{ }}\n}} bye"

>> Is there a way to do such a thing with a regex, or do I need to resort
>> to a scanning technique, keeping track of the nested elements?

DT> You can do it with regexps fairly efficiently, but it's a tad
DT> messy. What I do it flag matching delimiters first, using something
DT> like:

DT>     count = "0000";

DT>     1 while @content.gsub!(/\{([^\{\}]*)\}/m) {
DT>       count = count.succ
DT>       "#{MAGIC}:#{count}:#$1#{MAGIC1}:#{count}:"
DT>     }

DT>     @content.gsub!(/#{MAGIC}/,  '{')
DT>     @content.gsub!(/#{MAGIC1}/, '}')

DT> The pattern matches paired braces that don't themselves contain
DT> braces, so it works progressively from the deeper nestings out. It
DT> replaces braces with two magic characters (I use ^A and ^B), followed
DT> by a count. It then replaces the magic characters with braces, just
DT> to I can read the string when debugging.
[...]

Thanks Dave, Robert, Jimmy, etc...

I came up with a slightly different solution which I think should be
pretty quick:  create 2 arrays that have the locations of the
beginning and ending delimiters, then just walk those arrays to create
the ranges of balanced delimiters, then merge it with the stuff
outside the delimiters.  I'm sure this could be made much better, so
if anyone wants to point out better ways of doing this, feel free.

def asplit_text(pre_str, post_str, string)

  results = [];  s = []; e = []; ar = []
  slen = pre_str.length; elen = post_str.length
  pos = 0
  s << pos while pos = string.index(pre_str,pos+slen)
  pos = 0
  e << pos while pos = string.index(post_str,pos+elen)

  # get the ranges of the parts with delimiters
  start = s[0]; len = s.length-1
  1.upto(len) do |x|
    if s[x] > e[x-1]
      ar << [start,e[x-1]]
      start = s[x]
    end
  end
  ar << [start,e[-1]] # catch the last one...

  # merge it with the parts outside the delimiters
  last = -elen
  ar.each do |range|
    if range[0] > last
      results << string[last+elen..range[0]-1]
      last = range[1]
      results << string[range[0]..last+elen-1]
    end
  end

  # catch the trailing string if needed
  if ar[-1][1]+elen-1 < string.length-1
    results << string[ar[-1][1]+elen..string.length]
  end

  puts string
  p ar
  p results
end

asplit_text("{{", "}}", "start {{blah {{ {{  }} }} }} {{ asdf }} out {{ {{ }} }} bye")
puts "--------------------"
asplit_text("(", ")", "Hello ( ( World ) ) () (  ) ((())) Goodbye ()" )

===========================================================
prints:
start {{blah {{ {{  }} }} }} {{ asdf }} out {{ {{ }} }} bye
[[6, 26], [29, 37], [44, 53]]
["start ", "{{blah {{ {{  }} }} }}", " ", "{{ asdf }}", " out ", "{{ {{ }} }}", " bye"]
--------------------
Hello ( ( World ) ) () (  ) ((())) Goodbye ()
[[6, 18], [20, 21], [23, 26], [28, 33], [43, 44]]
["Hello ", "( ( World ) )", " ", "()", " ", "(  )", " ", "((()))", " Goodbye ", "()"]

thanks,
-joe