CODEWITHSUNDEEP
phpregexparsingsyntaxmarkup
SmallhillCZ
SmallhillCZ

Reputation: 171

How to parse heterogenous markup with PHP?

I have a string with custom markup for saving songs with chords, tabulatures, notes etc. It contains

things in various brackets: \[.+?\], \[[.+?\]], \(.+?\)
arrows: <-{3,}>, \-{3,}>, <\-{3,}
and so on...

Sample text might be 

Text Text [something]
--->
Text (something 021213)

Now I wish to parse the markup into array of tokens, objects of corresponding classes, which would look like (matched parts in brackets)

ParsedBlock_Text ("Text Text ")
ParsedBlock_Chord ("something")
ParsedBlock_Text (" ")
ParsedBlock_NewColumn
ParsedBlock_Text (" text ")
ParsedBlock_ChordDiagram ("something 021213")

I know how to match them, but either I must match each different pattern, and save offsets to properly sort the array, or I match them at once and I don't know which one has been matched.

Thanks, MK

Upvotes: 2

Views: 116

Answers (1)

Markus Jarderot
Markus Jarderot

Reputation: 89171

Assuming you do not try to nest these structures, this will tokenize your text:

function ParseText($text) {
    $re = '/\[\[(?P<DoubleBracket>.*?)]]|\[(?P<Bracket>.*?)]|\((?P<Paren>.*?)\)|(?<Arrow><---+>?|---+>)/s';
    $keys = array('DoubleBracket', 'Bracket', 'Paren', 'Arrow');
    $result = array();
    $lastStart = 0;
    if (preg_match_all($re, $text, $matches, PREG_SET_ORDER | PREG_OFFSET_CAPTURE)) {
        foreach ($matches as $match) {
            $start = $match[0][1];
            $prefix = substr($text, $lastStart, $start - $lastStart);
            $lastStart = $start + strlen($match[0][0]);
            if ($prefix != '' && !ctype_space($prefix)) {
                $result []= array('Text', trim($prefix));
            }
            foreach ($keys as $key) {
                if (isset($match[$key]) && $match[$key][1] >= 0) {
                    $result []=  array($key, $match[$key][0]);
                    break;
                }
            }
        }
    }
    $prefix = substr($text, $lastStart);
    if ($prefix != '' && !ctype_space($prefix)) {
        $result []= array('Text', trim($prefix));
    }
    return $result;
}

Example:

$mytext = <<<'EOT'
Text Text [something]
--->
Text (something 021213)
More Text
EOT;

$parsed = ParseText($mytext);
foreach ($parsed as $item) {
    print_r($item);
}

Output:

Array
(
    [0] => Text
    [1] => Text Text
)
Array
(
    [0] => Bracket
    [1] => something
)
Array
(
    [0] => Arrow
    [1] => --->
)
Array
(
    [0] => Text
    [1] => Text
)
Array
(
    [0] => Paren
    [1] => something 021213
)
Array
(
    [0] => Text
    [1] => More Text
)

http://ideone.com/kJQrBw

If you want to add more patterns to the regex, make sure you put longer patterns at the start, so they are not mistakenly matched as the wrong type.

Upvotes: 1

Related Questions