extract info from HTML

[thetakerfan]

I have an HTML file that I'm trying to pull some info out of. The HTML is formatted like this:

<font size=+2>Airplane Spin</font><br>
<b>Used by /b> Mike Rotunda<br>
<b>AKA /b> <br>
<b>Description /b> The attacker lifts the victim up across their
shoulders. The attacker starts spinning around quickly a few times
to dizzy the victim, then drops them to the mat. <br>
<br>

<p><font size=+2>Airplane Spin Toss</font><br>
<b>Used by /b> Al Perez, Sid Vicious, Oz<br>
<b>AKA /b> Ally-Coptor (Perez), Human Frisbee (Sid), Twister Slam
(Oz)<br>
<b>Description /b> The victim is lifted up over the attacker's
shoulder so the victim is facing upwards and their back is held
over the shoulder of the attacker. The attacker holds the victim in
place and spins around a few times, then tosses the victim into the
air dropping them back first to the mat.<br>
</p>
<p><font size=+2>Airplane Spin Toss, Face First</font><br>
<b>Used by /b> Mike Enos<br>
<b>AKA /b><br>
<b>Description /b> The attacker lifts up the victim over their
shoulder as if for a body slam. The attacker then spins around a
few times and then tosses the victim in the air dropping them to
the mat face first.</p>
<p><font size=+2>Arm Breaker</font><br>
<b>Used by /b><br>
<b>AKA /b><br>
<b>Description /b> The attacker has the victim's arm in a
wristlock. The attacker steps forward and drives the victim's arm
across their knee.</p>

What I want is to pull out the name (w/in the larger font tags) and description, but it isn't working properly. The first two work fine, then it skips the third, and then it just works sometimes, and not others. Here is my code:

$input = fopen ("E:/Programming/Docs/Reg.htm", "r" );
$output = fopen ("E:/Programming/Docs/Reg.txt", "w" );
while (!feof ($input)) {
$in_buf = fgets($input, 4096);
if (strpos($in_buf, "<font size=+2>" ) !== FALSE) {
$pos_start = strpos($in_buf, "<font size=+2>" ) + 14;
$pos_end = strpos($in_buf, "</font>" );
$string = trim(substr($in_buf, $pos_start, $pos_end - $pos_start));
//fputs($output, trim($in_buf)."\n" );
} 
elseif (strpos($in_buf, "<b>Description :</b>" ) !== FALSE) {
$pos_start = strpos($in_buf, "<b>Description :</b>" ) + 20;
$desc = 1;
$string .= ' *****'.trim(substr($in_buf, $pos_start));
} 
elseif ($desc == 1) {
if ((strpos($in_buf, "<br>" ) == FALSE) && (strpos($in_buf, "<p>" ) == FALSE)) {
$string .= trim($in_buf);
} elseif ((strpos($in_buf, "<br>" ) !== FALSE)) {
$pos_end = strpos($in_buf, "<br>" );
$string .= trim(substr($in_buf, 0, $pos_end).'*****');
fputs($output, trim($string)."\n\n" );
$desc = 0;
} elseif ((strpos($in_buf, "<p>" ) !== FALSE)) {
$pos_end = strpos($in_buf, "<p>" );
$string .= trim(substr($in_buf, 0, $pos_end)).'*****';
fputs($output, trim($string)."\n\n" );
$desc = 0;
}
}
}
fclose ($input);
fclose ($output); 

and a sample the output:

Airplane Spin *****The attacker lifts the victim up across theirshoulders. The attacker starts spinning around quickly a few timesto dizzy the victim, then drops them to the mat. *****

Airplane Spin Toss *****The victim is lifted up over the attacker'sshoulder so the victim is facing upwards and their back is heldover the shoulder of the attacker. The attacker holds the victim inplace and spins around a few times, then tosses the victim into theair dropping them back first to the mat.*****

Arm Breaker<b>Used by /b>*****

Arm Breaker, Fireman's Carry<b>Used by /b> CW Anderson*****

thanks in advance for any help

[Herbster]

I haven't played with your code. This is just from reading so I could be offbase.

You are processing the file in 4096 byte chunks, but the variable description length suggests to me that the paragraphs do not all have the same length.

I would probably read the entire file, explode it on <p> to create an array of paragraphs and loop through the array.

[thetakerfan]

I thought of that, but the problem is that the person who created the site wasn't uniform, so sometimes there are p's, and others just br's

[Herbster]

In that case, I would read the complete file and drop the paragrahs/sections with substr() as I process them.

I suspect you may split a search string operating on uniform 4096 byte chunks.