[Resolved] preg_replace puzzle

[trube]

I have a function that bolds query words for search results like in Google. But the problem is that sometimes it bolds matches that are inside HTML <a> tag. For example, if I have a string:

<a href=events.html>Events</a>

It will bold it like this:

<a href=<b>events</b>.html><b>Events</b></a>

What I want is:

<a href=events.html><b>Events</b></a>

Can someone help me fix my function so it only bolds query word matches that aren't inside a tag?

	// case-insensitive bolding of query words

function EmphasizeQueryWords($text, $queryWords, $beginstr='<b>', $endstr='</b>')
{
	$text = " ".$text." ";
	if (is_array($queryWords)) {
		// loop through all query words
		reset($queryWords);
		while(list($queryWordKey,$queryWord)=each($queryWords)) {
			if (stristr($text,$queryWord)) {
				$lowercase = stripslashes(strtolower($queryWord));
				$uppercase = stripslashes(ucwords($queryWord));
				$allcaps = stripslashes(strtoupper($queryWord));
				$patterns = "/(^|[^A-Za-z0-9])(".$lowercase."|".$uppercase."|".$allcaps.")+([^A-Za-z0-9]|$)/";
				$replace = "\\1".$beginstr."\\2".$endstr."\\3";
				$text = preg_replace($patterns, $replace, $text);
			}
		}
	}	
	$text = trim($text);
	return $text;

} // end func

thanks!

[tsinka]

Hi,

do you just want to manipulate the queryWords in a link/between HTML tags or in the complete document ?

To just replace it in tags:

    $theString = "<a href=event.php>event</a><<<";
    $queryWord = "Event";
    $queryWord = stripslashes($queryWord);
    $theString = preg_replace("/>.*(.$queryWord)+.*</i","<b>\\\\1</b>",$theString);
    echo $theString;

In a link:

<?PHP
    $theString = "<a href=event.php>Event</a><<<";
    $queryWord = "Event";
    $queryWord = stripslashes($queryWord);
    $theString = preg_replace("/(<a href=.*>.*)($queryWord)+(.*<\\/a>)/i","\\\\1<b>\\\\2</b>\\\\3",$theString);
    echo $theString;
?>

(without the white space after </a>

[trube]

Thanks for replying and for your help. Unfortunately, I could not get your suggestions to work. However, I altered the regexp and now it works. Here is the working function for anyone who may benefit from it.

	// case-insensitive bolding of query words

function EmphasizeQueryWords($text, $queryWords, $beginstr='<b>', $endstr='</b>') {
	$text = " ".$text." ";
	$replace = "\\\\1".$beginstr."\\\\2".$endstr."\\\\3";
	if (is_array($queryWords)) {
		// loop through all query words
		reset($queryWords);
		while(list($queryWordKey,$queryWord)=each($queryWords)) {
			if (stristr($text,$queryWord)) {
				// throw all letters into an array
				$lowercase = stripslashes(strtolower($queryWord));
				$word_length = strlen($lowercase);
				$letter = array();
				for ($i=0; $i<$word_length; $i++) {
					$letter[$i] = substr($lowercase,$i,1);
				}
				$regexp_bit = array();
				foreach ($letter as $key => $val) {
					$regexp_bit[] = "[".$val.strtoupper($val)."]";
				}
				$patterns = "/([\\s\\$-'\\"])(".implode("",$regexp_bit).")([\\s,-\\.;:!\\?'\\"])/";					
				$text = preg_replace($patterns, $replace, $text);
			}
		}
	}	
	$text = trim($text);
	return $text;

} // end func

[trube]

As a bonus, here's my other function that replaces stretches of text without matches with an elipsis, like in Google search results.

function CutDownLength($text, $max=200, $beginstr='<b>', $endstr='</b>', $wordspacing=5) {

	$debug = false;
	$debug2 = false;

	// $beginstr .. beginning of query words (i.e., already emboldened)
	// $endstr .. end of query words (i.e., already emboldened)
	// $wordspacing .. how many words before and after the query word to leave

	$b_strlen = strlen($beginstr);
	$e_strlen = strlen($endstr);

	$bold_strlen = $b_strlen + $e_strlen;

	// REDUCES STRING TO $max CHARACTERS OR LESS WITHOUT BREAKING UP WORDS
	$text = trim($text);

	// reduce string if it's greater than max

	if (strlen($text) > $max) {
		$times = substr_count($text, $beginstr);
		// if there is only one query word

		switch ($times) {
			case 0:
				break;
			case 1:

				$i=0;

				// loop while $text is longer than max allowed
				$thismax = $max+$bold_strlen;

				while ((strlen($text) > $thismax)){

					if ($debug) echo $text . "<br>";

					// get the query word along with $beginstr and $endstr

					$begin_pos = strpos($text,$beginstr);
					$end_pos = strpos($text,$endstr);
					$query_word = substr($text, $begin_pos, ($end_pos - $begin_pos + strlen($endstr)));
					if ($debug) echo "text strlen: " . strlen($text) . '<br>';
					if ($debug) echo "begin_pos: " . $begin_pos . '<br>';
					if ($debug) echo "end_pos: " . $end_pos . '<br>';
					if ($debug) echo "query word: " . $query_word . '<br>';

					$outer_len = strlen($text) - strlen($query_word);
					$before = explode($beginstr,$text);
					$after = explode($endstr,$text);
					$before_len = strlen(trim($before[0]));
					$before_ratio = $before_len / $outer_len;
					$after_len = strlen(trim($after[1]));
					$after_ratio = $after_len / $outer_len;

					$total_len_to_chop = strlen($text) - $thismax;
					$len_to_chop_before = round($before_ratio * $total_len_to_chop) + 1;
					$len_to_chop_after = round($after_ratio * $total_len_to_chop) + 1;
					if ($debug) echo "total_len_to_chop: " . $total_len_to_chop . '<br>';
					if ($debug) echo "before_ratio: " . $before_ratio . '<br>';
					if ($debug) echo "after_ratio: " . $after_ratio . '<br>';
					if ($debug) echo "len_to_chop_before: " . $len_to_chop_before . '<br>';
					if ($debug) echo "len_to_chop_after: " . $len_to_chop_after . '<br>';

					// this is so weird
					// need to trim the before or after part for the count of the resulting array to decrement

					// if query word is not at beginning, remove one word from the beginning
					if ($begin_pos <> 0) {

						$before_query_word_array = explode(" ", trim($before[0]));
						if (is_array($before_query_word_array)) {
							$num_before_words = count($before_query_word_array);
							if ($num_before_words > $wordspacing) {

								if ($i==0) {
									// based on how long the string is, guess how many words to chop
									$ave_before_word_len = $before_len / $num_before_words;
									$numWordChop = number_format($len_to_chop_before / $ave_before_word_len, 0, '.', '');
									if (($num_before_words - $numWordChop) < $wordspacing) {
										$numWordChop = $num_before_words - $wordspacing;
									}

									if ($debug) echo "Before word count: " . count($before_query_word_array) . '<br>';
									$before_query_word_array = array_slice($before_query_word_array, $numWordChop);
								} else {
									unset($before_query_word_array[0]);
								}

								$before_word_removed = true;
							}
							$before_query_word = implode(" ", $before_query_word_array);
							unset($before_query_word_array);
						} else {
							$before_query_word = $before[0];
						}
					}


					// if query word is not at end, remove one word from the end
					if ($end_pos <> (strlen($text)-(strlen($endstr))-1)) {
						$after_query_word_array = explode(" ", trim($after[1]));
						if (is_array($after_query_word_array)) {
							$num_after_words = count($after_query_word_array);
							if ($num_after_words > $wordspacing) {

								if ($i==0) {
									// based on how long the string is, guess how many words to chop
									$ave_after_word_len = $after_len / $num_after_words;
									$numWordChop = number_format($len_to_chop_after / $ave_after_word_len, 0, '.', '');
									if (($num_after_words - $numWordChop) < $wordspacing) {
										$numWordChop = $num_after_words - $wordspacing;
									}

									if ($debug) echo "after word count: " . count($after_query_word_array) . '<br>';
									$after_query_word_array = array_reverse($after_query_word_array);
									$after_query_word_array = array_slice($after_query_word_array, $numWordChop);
								} else {
									unset($after_query_word_array[0]);
								}
								$after_word_removed = true;
								$after_query_word_array = array_reverse($after_query_word_array);
							}
							$after_query_word = implode(" ", $after_query_word_array);
							unset($after_query_word_array);
						} else {
							$after_query_word = $after[1];
						}

					}

					$text = '';
					if ($before_query_word) {

						$text .= $before_query_word;

						// if the last character is alphanumeric then put a space
						if (preg_match("/([A-Za-z0-9.,;)])$/", $before_query_word)) {
							$text .= " ";
						}

					}
					$text .= $query_word;
					if ($after_query_word) {

						// if the first character is alphanumeric then put a space
						if (preg_match("/^([A-Z-a-z0-9&(])/", $after_query_word)) {
							$text .= " ";
						}

						$text .= $after_query_word;
					}
					unset($before_query_word, $after_query_word, $after, $before);

					if ($i >20) break;
					$i++;

					if ($debug) echo "<br>";


				} // loop while $text is longer than max allowed

				if ($before_word_removed) {
					$text = "... " . $text;
				}
				if ($after_word_removed) {
					$text .= " ...";
				}

				break;

			default:

				// this is for when there is more than one occurence query word(s) in the text

				if ($debug2) echo $text . '<br>';
				$thismax = $max + ($bold_strlen * $times);
				if ($debug2) echo "thismax: " . $thismax . '<br>';

				// identify and remove both ends

				$s = " " . $text;
				// explode based on beginstr
				$a = explode($beginstr,$s);
				$strbeforefirst = $a[0];
				if ($debug2) echo "strbeforefirst: " . $strbeforefirst . '<br>';

				// remove str before first occurence
				$a = array_slice($a,1);

				// put back into string
				$s = implode($beginstr,$a) . " ";

				// explode based on endstr
				$a = explode($endstr,$s);
				$a = array_reverse($a);
				// remove str after last occurence
				$strafterlast = $a[0];
				if ($debug2) echo "strafterlast: " . $strafterlast . '<br>';

				$a = array_slice($a,1);
				$a = array_reverse($a);

				// put back into string
				$s = $beginstr.implode($endstr,$a).$endstr;

				$length = strlen(strip_tags($s));

				if ($length > $thismax) {
				if ($debug2) echo "<i>length too long need to cut down</i><br>";

					// then will need to split up the string
					// look for the longest break between occurences

//					$s = " $s";
					// explode based on beginstr
					$a = explode($beginstr,$s);
					while(list($key1,$val1)=each($a)) {
						$sectionLen = strlen($val1)-$e_strlen;
						if ($debug2) echo "sectionLen: " . $sectionLen . '<br>';

						if ($sectionLen > ($thismax/3)) {
							$b = explode(" ", $val1);
							if (is_array($b)) {
								$numBwords = count($b); // num words in this section
								if ($numBwords > $wordspacing) {
									$d = array_slice($b,($numBwords-$wordspacing));
									$e = array_reverse($b);
									$c = array_slice($e,($numBwords-$wordspacing));
									$c = array_reverse($c);
									$a[$key1] = implode(" ", $c) . " ... " .implode(" ", $d);
								}
							}
						}
					}

					$s =  implode($beginstr,$a);

				} 

				// output the whole thing with a little before and after
				$a = explode(" ",$strbeforefirst);
				if (is_array($a)) {
					$numWords = count($a);
					$a = array_slice($a,($numWords-3));
					$beforestrclip = "... " . trim(implode(" ", $a));

						// if the last character is alphanumeric then put a space
						if (preg_match("/([A-Za-z0-9.,;)&])$/", $beforestrclip)) {
							$beforestrclip .= " ";
						}

					if (strlen($strbeforefirst) < strlen($beforestrclip)) {
						$beforestrclip = $strbeforefirstv;
					}
				}

				$a = explode(" ",$strafterlast);
				if (is_array($a)) {
					$numWords = count($a);
					$a = array_reverse($a);
					$a = array_slice($a,($numWords-$wordspacing));
					$a = array_reverse($a);
					$afterstrclip =  trim(implode(" ", $a)) . " ...";

						// if the first character is alphanumeric then put a space
						if (preg_match("/^([A-Z-a-z0-9&(])/", $afterstrclip)) {
							$afterstrclip = " " . $afterstrclip;
						}
					if (strlen($strafterlast) < strlen($afterstrclip)) {
						$afterstrclip = $strafterlast;
					}
				}

				$text = $beforestrclip . $s . $afterstrclip;

				if ($debug2) echo "$text<br>";
				if ($debug2) echo "resulting legnth: " . strlen($text) . "<br>";
				if ($debug2) echo "<hr>";

				break;

		}

	} 

	return $text;
}

The only problem with this function is that if there are tons of matches to the query words then, it doesn't cut it down to specified size. If anyone expands on the function to accomodate that, please share it with me. Thanks.