Trouble getting <img> alt text with regular expression

[Greaser9780]

Here is what I have:

$regexp = "<img\s[^>]*alt=(\"??)([^\" >]*?)\\1[^>]*>"; 
if(preg_match_all("/$regexp/siU", $input, $img, PREG_SET_ORDER)) { 
foreach($img as $alt){
echo "$alt[2] <br />";
}}

 

I am a reg expression noob. This presents empty results. Can anyone find the error?

[PHP_PhREEEk]

Before going through all of that expression logic, can you enlighten us on what you're trying to accomplish? The bigger picture, that is... there might be a much easier way to get the end result you want.

 

PhREEEk

[Greaser9780]

I'm in the midst of building a search engine. I am going to rank sites for each search on things like % of query words to total words for various groups(alt text, description words, title, keywords) and so on. I did however get the following to work:

$regexp = "<img src=\"(.*)\" alt=\"(.*)\">"; 
if(preg_match_all("/$regexp/siU", $input, $img, PREG_SET_ORDER)) { 
foreach($img as $alt){
$alt_words = adv_count_words($alt[2]);
echo "$alt[2]  -  $alt_words<br />";

 

So far I can get the words and word count for the following:

title

meta description

meta keywords

link text

link href

heading text

bold text

underlined text

img alt text

[Greaser9780]

Now I'm starting to wonder how to put it all in the db. Since the links, alt text,headings, and bold type are all listed by line. I hate to input it with a foreach loop since this will cause many small queries, but maybe this is better than one ginormous query.

    Then there is the problem when a search is performed. How exactly should the user query to keyword % get checked?

1. Check for all keywords

  if all are found what is the percentage

  if all are not found break string into individual words and check for %

or

2. Break user query into individual words and check percentage this way?

[SammyGunnz]

This should work...

 

<?php

  preg_match_all('/\<img .* alt=\"([aA-zZ]+)\"/',$str, $imgalt);
  
  print_r($imgalt[1]);


?>

[Greaser9780]

Thank you. That's pretty much what I ended up using.