I need a regex to solve the following situation.
A satisfactory answer will be PHP code that takes a string as an input
and returns the same string with problem resolved.

I have a block of html that has a group of list elements (<li>Bullet
Text</li>) which are added using an html text editor.  The text editor
is not adding the required surrounded <ul> and </ul>.  The regex
should therefore search a string for sets of list elements that do not
have surrounding <ul> tags and add them.
NB: 
-if there are <ul> tags then no more should be added.
-there may be more than one set of lists
-sets of <ol> tags should be ignored
-<li> tags may or may not be on seperate lines 
-the code will be tested using the following string: note the line line breaks


<P ALIGN="LEFT">this is some text<br/><br/>
<h2>Stay Safe</h2><br/>
<B>Residential care</B><br/><br/>this is some text<br/></P>
<LI>list item 1</LI><LI>list item 2</LI><LI>list item
3</LI><LI></LI><P ALIGN="LEFT">Our staff are blah blah</P><LI>list
item 1</LI>

<LI>list item 2</LI><LI>list item 3</LI>some more text might start straight away

- the code should return the following string

<P ALIGN="LEFT">this is some text<br/><br/>
<h2>Stay Safe</h2><br/>
<B>Residential care</B><br/><br/>this is some text<br/></P>
<UL><LI>list item 1</LI><LI>list item 2</LI><LI>list item
3</LI><LI></LI></UL><P ALIGN="LEFT">Our staff are blah
blah</P><UL><LI>list item 1</LI>

<LI>list item 2</LI><LI>list item 3</LI></UL>some more text might
start straight away

---

Clarification of Question by larking-ga on 10 Dec 2005 01:27 PST

when i say <ol> lists should be ignored i mean to say they should be
left intact and nothing should be changed about them

The following solution works if you can live with whitespace being
altered between <LI> tags:

$input = preg_replace("/(<\/li>)\s*(<li>)/i","$1$2",$input);
$input = preg_replace("/(?<!li>)<li>/i","<UL><LI>",$input);
$input = preg_replace("/<\/li>(?!<li>)/i","</LI></UL>",$input);

Step 1 closes the whitespace between <LI> tags
Step 2 matches any <LI> tag which is not preceeded by a closing </LI>
tag and inserts a start <UL> through replacement.
Step 3 does the reverse, adding closing <UL> tags.

Running the above creates the string:

<P ALIGN="LEFT">this is some text<br/><br/>
<h2>Stay Safe</h2><br/>
<B>Residential care</B><br/><br/>this is some text<br/></P>
<UL><LI>list item 1</LI><LI>list item 2</LI><LI>list item
3</LI><LI></LI></UL><P ALIGN="LEFT">Our staff are blah blah</P><UL><LI>list
item 1</LI><LI>list item 2</LI><LI>list item 3</LI></UL>some more text
might start straight away

Which matches your test case (but removes whitespace between LI).  Let
me know if I can improve this solution for you.

Thanks, that is perfect. Works like a charm :D