Parsing XML through PHP

[kexxcream]

I am trying to extract all <p></p> tags from XML-files that is zipped. I use the following code to extract and read the XML-file (works without any problem):

//Get the zipfile class
require "zip.class.php";

//Create an object
$zipfile = new zipfile;

//Read the zip file
$zipfile->read_zip("export-03-04-2011.zip");

Then I use the following to loop through every XML-file (works without any problem):

foreach($zipfile->files as $stream)
{

Then I use the following to extract only <p></p> tags:

preg_match_all('/<p>(.*?)<\/p>/', utf8_decode($stream['data']), $ProcessedArray);

This however also includes <p></p> that literally are not empty (space or white space included).

The problem: How do I first get rid of the empty <p>-tags? Is there a regular expression that can only extract <p>-tags that have at least 2 characters?

[bradgrafelman]

Welcome to PHPBuilder!

kexxcream wrote:

Is there a regular expression that can only extract <p>-tags that have at least 2 characters?

Of course! Look at this part of your regexp:

.*?

That says, in English: "match any character zero or more times (non-greedy; stop at the nearest boundary)." Specifically, the '*' modifier is the 'zero or more' part.

Other choices would include the '+' modifier, which means 'one or more', or you could use the {M,N} syntax which literally specifies "between M and N number of times." Omitting the 'N' (but not the comma!) means "M or more times", so in that manner you could change the dot to match at least two characters.

[kexxcream]

Hello!

Excellent educational reply, thank you for that. But it does not fully solve my issue although I believe we are on the right path.

So the XML-file looks something like this:

<?xml version="1.0" encoding="utf-8"?>
<assessmentItem xsi:schemaLocation="http://www.imsglobal.org/xsd/imscp_v1p1 http://www.imsglobal.org/xsd/imscp_v1p1.xsd http://www.imsglobal.org/xsd/imsmd_v1p2 http://www.imsglobal.org/xsd/imsmd_v1p2p2.xsd http://www.imsglobal.org/xsd/imsqti_v2p1 http://www.imsglobal.org/xsd/imsqti_v2p1.xsd" identifier="item45790" title="MultipleChoice" adaptive="false" timeDependent="false" toolName="itslearning" toolVersion="3.3" xmlns="http://www.imsglobal.org/xsd/imsqti_v2p1" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:qti="http://www.imsglobal.org/xsd/imsqti_v2p1">
    <responseDeclaration cardinality="single" baseType="identifier" identifier="RESPONSE">
        <correctResponse>
            <value>A3</value>
        </correctResponse>
        <mapping lowerBound="-1" upperBound="1" defaultValue="0">
            <mapEntry mapKey="A3" mappedValue="1" />
            <mapEntry mapKey="A0" mappedValue="-0.33333333333333331" />
            <mapEntry mapKey="A1" mappedValue="-0.33333333333333331" />
            <mapEntry mapKey="A2" mappedValue="-0.33333333333333331" />
        </mapping>
    </responseDeclaration>
    <outcomeDeclaration cardinality="single" baseType="identifier" identifier="FEEDBACK" view="" normalMaximum="0" normalMinimum="0" masteryValue="0" />
    <outcomeDeclaration cardinality="single" baseType="identifier" identifier="INTEGRATEDFEEDBACK" view="" normalMaximum="0" normalMinimum="0" masteryValue="0" />
    <itemBody>
        <p>What is the capital of USA?</p>
        <choiceInteraction responseIdentifier="RESPONSE" shuffle="false" maxChoices="1" minChoices="0">
            <simpleChoice fixed="false" showHide="show" identifier="A0">
                <p>Washington DC</p>
            </simpleChoice>
            <simpleChoice fixed="false" showHide="show" identifier="A1">
                <p>Miami</p>
                <p> </p>
            </simpleChoice>
            <simpleChoice fixed="false" showHide="show" identifier="A2">
                <p>Los Angeles</p>
            </simpleChoice>
            <simpleChoice fixed="false" showHide="show" identifier="A3">
                <p>New York</p>
            </simpleChoice>
        </choiceInteraction>
    </itemBody>
    <responseProcessing />
</assessmentItem>

Notice how the option Miami contains another <p>-tag but it's empty. I use now the following php code:

preg_match_all('/<p>(.{2,200}?)<\/p>/', utf8_decode($stream['data']), $ProcessedArray);

And when I print it out I do the following:

//Replace <p>-tags with <span>-tags for question headline
$patterns = array('<p>', '</p>');
$replacements = array('<span>', '</span>');

$FormalQuestion = str_replace($patterns, $replacements, $ProcessedArray[0][0]);

//Print question headline	
echo '<strong>Fråga ' . $i . '. ' . $FormalQuestion . '</strong>' . "\n";

//Print question answers
echo $ProcessedArray[0][1] . "\n";
echo $ProcessedArray[0][2] . "\n";
echo $ProcessedArray[0][3] . "\n";
echo $ProcessedArray[0][4] . "\n";

However, the source code generates the following (another question):

Debatten omkring Rorschach's validitet kommer från tvetydiga resultat i forskning omkring:</span>

<p>Rorschachtestets prediktionsförmåga</p>
<p>Rorschachtestets förmåga att differentiera mellan olika kliniska grupper</p>
<p> </p></simpleChoice><simpleChoice fixed="false" showHide="show" identifier="A2"><p>Tolkningen av svaren i Rorschachtest</p>
<p> </p></simpleChoice><simpleChoice fixed="false" showHide="show" identifier="A3"><p>Alla ovanstående</p>

1. Why does it include XML-code?
2. How can we only extract <p>-tags in the XML that has content?

[NogDog]

Your "{2,200}" repetition modifier is requiring a minimum of 2 characters after the "<p>" tag. Since you have one space character before the following "</p>" closing tag in the empty ones, the "<" of the closing tag is being "eaten" by the ".{2,200}", so it does not find a terminating "</p>" until the end of the next "p" element. You could simply use ".*?" instead to make an "ungreedy" match on any number of characters.

In the long run, though, you might find a more robust solution would be to use the [man]DOM[/man] extension instead of using regular expressions, which only work until something changes in the source material that screws you up. 🙂

[kexxcream]

Many thanks for your explanation. I looked through the DOM document and it looked interesting. I tried a few combinations but didn't get far with it.

I hope it's not too much to get some guidance that could lead me in the right direction. DOM seems to be a good solution.

[NogDog]

Here's a fairly simple example of just grabbing all "<p>" elements:

<?php
$text = <<<EOD
<?xml version="1.0" encoding="utf-8"?>
<assessmentItem xsi:schemaLocation="http://www.imsglobal.org/xsd/imscp_v1p1 http://www.imsglobal.org/xsd/imscp_v1p1.xsd http://www.imsglobal.org/xsd/imsmd_v1p2 http://www.imsglobal.org/xsd/imsmd_v1p2p2.xsd http://www.imsglobal.org/xsd/imsqti_v2p1 http://www.imsglobal.org/xsd/imsqti_v2p1.xsd" identifier="item45790" title="MultipleChoice" adaptive="false" timeDependent="false" toolName="itslearning" toolVersion="3.3" xmlns="http://www.imsglobal.org/xsd/imsqti_v2p1" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:qti="http://www.imsglobal.org/xsd/imsqti_v2p1">
    <responseDeclaration cardinality="single" baseType="identifier" identifier="RESPONSE">
        <correctResponse>
            <value>A3</value>
        </correctResponse>
        <mapping lowerBound="-1" upperBound="1" defaultValue="0">
            <mapEntry mapKey="A3" mappedValue="1" />
            <mapEntry mapKey="A0" mappedValue="-0.33333333333333331" />
            <mapEntry mapKey="A1" mappedValue="-0.33333333333333331" />
            <mapEntry mapKey="A2" mappedValue="-0.33333333333333331" />
        </mapping>
    </responseDeclaration>
    <outcomeDeclaration cardinality="single" baseType="identifier" identifier="FEEDBACK" view="" normalMaximum="0" normalMinimum="0" masteryValue="0" />
    <outcomeDeclaration cardinality="single" baseType="identifier" identifier="INTEGRATEDFEEDBACK" view="" normalMaximum="0" normalMinimum="0" masteryValue="0" />
    <itemBody>
        <p>What is the capital of USA?</p>
        <choiceInteraction responseIdentifier="RESPONSE" shuffle="false" maxChoices="1" minChoices="0">
            <simpleChoice fixed="false" showHide="show" identifier="A0">
                <p>Washington DC</p>
            </simpleChoice>
            <simpleChoice fixed="false" showHide="show" identifier="A1">
                <p>Miami</p>
                <p> </p>
            </simpleChoice>
            <simpleChoice fixed="false" showHide="show" identifier="A2">
                <p>Los Angeles</p>
            </simpleChoice>
            <simpleChoice fixed="false" showHide="show" identifier="A3">
                <p>New York</p>
            </simpleChoice>
        </choiceInteraction>
    </itemBody>
    <responseProcessing />
</assessmentItem>
EOD;

libxml_use_internal_errors(true); // so that loadHTML does not spew warnings
try {
   $dom = new DOMDocument('1.0', 'UTF-8');
   $dom->loadHTML($text);
   $p = $dom->getElementsByTagName('p');
   foreach($p as $node) {
      $str = trim($node->textContent);
      if(!empty($str)) {
         echo "<p>$str</p>\n";
      }
   }
}
catch(Exception $e) {
   die((string) $e);
}

[kexxcream]

For some reason that solution didn't work because I am parsing XML-files in a zip-file according to the following.

//Get the zipfile class
require "zip.class.php";

//Create an object
$zipfile = new zipfile;

//Read the zip file
$zipfile->read_zip("export-03-04-2011.zip");

To read the content I use the following:

$xml = new SimpleXMLElement($stream['data']);

Then I managed to try on my own using previous recommendations and came up with this solution:

echo '<h4>Question ' . $i . ': ' . utf8_decode($xml->itemBody->p) . '</h4>' . "\n";
echo utf8_decode($xml->itemBody->choiceInteraction->simpleChoice[0]->p) . '<br />' . "\n";
echo utf8_decode($xml->itemBody->choiceInteraction->simpleChoice[1]->p) . '<br />' . "\n";
echo utf8_decode($xml->itemBody->choiceInteraction->simpleChoice[2]->p) . '<br />' . "\n";
echo utf8_decode($xml->itemBody->choiceInteraction->simpleChoice[3]->p) . '<br />' . "\n";

It looks fine in the output, but is it really a good and sustainable solution? What happens if an empty <p> tag comes before the answer?