How to preg_match_all on German Umlauts [äöü]? - php

If a text has German Umlauts [äöü] the result of preg_match_all has wrong offsets (it seems each Umlaut extend the offset by 1)
I need the position of each word, because they will be replaced by other strings. With this tool https://regex101.com/r/UosqVD/2 it worked, the matches have the correct start value.
$pattern = "~\b\w+\b~u";
$text = "Käthe würde gerne wählen.";
if (preg_match_all($pattern, $text, $matches, PREG_OFFSET_CAPTURE)) {
foreach ($matches[0] as $m) {
echo $m[0]."; ".$m[1]."; ".mb_strlen($m[0], "utf-8")."<br />";
}
}
Text; Start, Length<br>
Käthe; 0; 5<br>
würde; 7; 5<br>
gerne; 14; 5<br>
wählen; 20; 6<br>

The PHP documentation contains a function mb_preg_match_all() written by a user that seems to fit your needs :
function mb_preg_match_all($ps_pattern, $ps_subject, &$pa_matches, $pn_flags = PREG_PATTERN_ORDER, $pn_offset = 0, $ps_encoding = NULL) {
// WARNING! - All this function does is to correct offsets, nothing else:
//
if (is_null($ps_encoding))
$ps_encoding = mb_internal_encoding();
$pn_offset = strlen(mb_substr($ps_subject, 0, $pn_offset, $ps_encoding));
$ret = preg_match_all($ps_pattern, $ps_subject, $pa_matches, $pn_flags, $pn_offset);
if ($ret && ($pn_flags & PREG_OFFSET_CAPTURE))
foreach($pa_matches as &$ha_match)
foreach($ha_match as &$ha_match)
$ha_match[1] = mb_strlen(substr($ps_subject, 0, $ha_match[1]), $ps_encoding);
//
// (code is independent of PREG_PATTER_ORDER / PREG_SET_ORDER)
return $ret;
}

Related

Php search replacer

I have a search String: $str (Something like "test"), a wrap string: $wrap (Something like "|") and a text string: $text (Something like "This is a test Text").
$str is 1 Time in $text. What i want now is a function that will wrap $str with the wrap defined in $wrap and output the modified text (even if $str is more than one time in $text).
But it shall not output the whole text but just 1-2 of the words before $str and then 1-2 of the words after $str and "..." (Only if it isn`t the first or last word). Also it should be case insensitive.
Example:
$str = "Text"
$wrap = "<span>|</span>"
$text = "This is a really long Text where the word Text appears about 3 times Text"
Output would be:
"...long <span>Text</span> where...word <span>Text</span> appears...times <span>Text</span>"
My Code (Obviusly doesnt works):
$tempar = preg_split("/$str/i", $text);
if (count($tempar) <= 2) {
$result = "... ".substr($tempar[0], -7).$wrap.substr($tempar[1], 7)." ...";
} else {
$amount = substr_count($text, $str);
for ($i = 0; $i < $amount; $i++) {
$result = $result.".. ".substr($tempar[$i], -7).$wrap.substr($tempar[$i+1], 0, 7)." ..";
}
}
If you have a tipp or a solution dont hesitate to let me know.
I have taken your approach and made it more flexible. If $str or $wrap changes you could have escaping issues within the regex pattern so I have used preg_quote.
Note that I added $placeholder to make it clearer, but you can use $placeholder = "|" if you don't like [placeholder].
function wrapInString($str, $text, $element = 'span') {
$placeholder = "[placeholder]"; // The string that will be replaced by $str
$wrap = "<{$element}>{$placeholder}</{$element}>"; // Dynamic string that can handle more than just span
$strExp = preg_quote($str, '/');
$matches = [];
$matchCount = preg_match_all("/(\w+\s+)?(\w+\s+)?({$strExp})(\s+\w+)?(\s+\w+)?/i", $text, $matches);
$response = '';
for ($i = 0; $i < $matchCount; $i++) {
if (strlen($matches[1][$i])) {
$response .= '...';
}
if (strlen($matches[2][$i])) {
$response .= $matches[2][$i];
}
$response .= str_replace($placeholder, $matches[3][$i], $wrap);
if (strlen($matches[4][$i])) {
$response .= $matches[4][$i];
}
if (strlen($matches[5][$i]) && $i == $matchCount - 1) {
$response .= '...';
}
}
return $response;
}
$text = "text This is a really long Text where the word Text appears about 3 times Text";
string(107) "<span>text</span> This...long <span>text</span> where...<span>text</span> appears...times <span>text</span>"
To make the replacement case insensitive you can use the i regex option.
If I understand your question correct, just a little bit of implode and explode magic needed
$text = "This is a really long Text where the word Text appears about 3 times Text";
$arr = explode("Text", $text);
print_r(implode('<span>Text</span>', $arr));
If you specifically need to render the span tags using HTML, just write it that way
$arr = explode("Text", $text);
print_r(implode('<span>Text</span>', $arr));
Use patern below to get your word and 1-2 words before and after
/((\w+\s+){1,2}|^)text((\s+\w+){1,2}|$)/i
demo
In PHP code it can be:
$str = "Text";
$wrap = "<span>|</span>";
$text = "This is a really long Text where the word Text appears about 3 times Text";
$temp = str_replace('|', $str, $wrap); // <span>Text</span>
// find patern and 1-2 words before and after
// (to make it casesensitive, delete 'i' from patern)
if(preg_match_all('/((\w+\s+){1,2}|^)text((\s+\w+){1,2}|$)/i', $text, $match)) {
$res = array_map(function($x) use($str, $temp) { return '... '.str_replace($str, $temp, $x) . ' ...';}, $match[0]);
echo implode(' ', $res);
}

filtering bad words from text

This function filer the email from text and return matched pattern
function parse($text, $words)
{
$resultSet = array();
foreach ($words as $word){
$pattern = 'regex to match emails';
preg_match_all($pattern, $text, $matches, PREG_OFFSET_CAPTURE );
$this->pushToResultSet($matches);
}
return $resultSet;
}
Similar way I want to match bad words from text and return them as $resultSet.
Here is code to filter badwords
TEST HERE
$badwords = array('shit', 'fuck'); // Here we can use all bad words from database
$text = 'Man, I shot this f*ck, sh/t! fucking fu*ker sh!t f*cking sh\t ;)';
echo "filtered words <br>";
echo $text."<br/>";
$words = explode(' ', $text);
foreach ($words as $word)
{
$bad= false;
foreach ($badwords as $badword)
{
if (strlen($word) >= strlen($badword))
{
$wordOk = false;
for ($i = 0; $i < strlen($badword); $i++)
{
if ($badword[$i] !== $word[$i] && ctype_alpha($word[$i]))
{
$wordOk = true;
break;
}
}
if (!$wordOk)
{
$bad= true;
break;
}
}
}
echo $bad ? 'beep ' : ($word . ' '); // Here $bad words can be returned and replace with *.
}
Which replaces badwords with beep
But I want to push matched bad words to $this->pushToResultSet() and returning as in first code of email filtering.
can I do this with my bad filtering code?
Roughly converting David Atchley's answer to PHP, does this work as you want it to?
$blocked = array('fuck','shit','damn','hell','ass');
$text = 'Man, I shot this f*ck, damn sh/t! fucking fu*ker sh!t f*cking sh\t ;)';
$matched = preg_match_all("/(".implode('|', $blocked).")/i", $text, $matches);
$filter = preg_replace("/(".implode('|', $blocked).")/i", 'beep', $text);
var_dump($filter);
var_dump($matches);
JSFiddle for working example.
Yes, you can match bad words (saving for later), replace them in the text and build the regex dynamically based on an array of bad words you're trying to filter (you might store it in DB, load from JSON, etc.). Here's the main portion of the working example:
var blocked = ['fuck','shit','damn','hell','ass'],
matchBlocked = new RegExp("("+blocked.join('|')+")", 'gi'),
text = $('.unfiltered').text(),
matched = text.match(matchBlocked),
filtered = text.replace(matchBlocked, 'beep');
Please see the JSFiddle link above for the full working example.

PHP How to split a string with 2 delimiters with php starting by the right

I need delimiter a string (it's html) and i need extract a substring, but i only have a function to delimiter by the left:
function ddd($str, $sLeft, $sRight)
{
$pleft = strpos($str, $sLeft, 0);
$pright = strpos($str, $sRight, $pleft + strlen($sLeft));
return (substr($str, $pleft + strlen($sLeft), ($pright - ($pleft + strlen($sLeft)))));
}
but i need start on the right ($sRight) and end at $sLeft.
Thanks!
If I understand you correctly, you need to search the string by the left side, and also by the right side. If that is the case you can use this function:
PHP
<?php
// Function from php.net, see the link above.
// Starts searching the value from the right side
// of the string, returning the pos
function backwardStrpos($haystack, $needle, $offset = 0){
$length = strlen($haystack);
$offset = ($offset > 0)?($length - $offset):abs($offset);
$pos = strpos(strrev($haystack), strrev($needle), $offset);
return ($pos === false) ? (false) : ($length - $pos - strlen($needle));
}
$string = 'My string is this, and it is a good dialog';
echo backwardStrpos($string, "i"); // outputs 37
echo strpos($string, "i", 0); // outputs 6
?>
ILLUSTRATED VERSION OF THE OUTPUT:
EDITED
You've just placed a comment related to what you are trying to do.
For that you can use PHP strip_tags:
<?php
$str = '<div><span class="some style bla bla assd dfsdf">ImportantText</span>';
echo strip_tags($str); //outputs: ImportantText
?>
EDITED
To extract text between HTML tags:
<?php
function getTextBetweenTags($string, $tagname) {
$pattern = "/<$tagname ?.*>(.*)<\/$tagname>/";
preg_match($pattern, $string, $matches);
return $matches[1];
}
$str = '<div><span class="some style bla bla assd dfsdf">ImportantText</span>';
$txt = getTextBetweenTags($str, "span");
echo $txt; // outputs: ImportantText
?>
This came from a fellow stackoverflow! link

Convert lib_string to string w/o Regex

I need to convert lib_someString to someString inside a block of text using str_replace [not regex].
Here's an example to give an exact sense what I mean: lib_12345 => 12345. I need to do this for a bunch of instances in a block of text.
Below is my attempt. Problem I'm getting is that my function is not doing anything (I just get lib_id returned).
function extractLibId($val){ // function to get the "12345" in the above example
$lclRetVal = substr($val, 5, strlen($val));
return $lclRetVal;
}
function Lib($text){ // does the replace for all lib_ instances in the text
$lclVar = "lib_";
$text = str_replace($lclVar, "<a href='".extractLibId($lclVar)."'>".extractLibId($lclVar)."</a>", $text);
return $text;
}
Regexp gonna be faster and more clear, you will have no need to call your function for every possible 'lib_' string:
function Lib($text) {
$count = null;
return preg_replace('/lib_([0-9]+)/', '$1', $text, -1, $count);
}
$text = 'some text lib_123123 goes here lib_111';
$text = Lib($text);
Without regexp, but every time Lib2 will be called somewhere will die cute kitten:
function extractLibId($val) {
$lclRetVal = substr($val, 4);
return $lclRetVal;
}
function Lib2($text) {
$count = null;
while (($pos = strpos($text, 'lib_')) !== false) {
$end = $pos;
while (!in_array($text[$end], array(' ', ',', '.')) && $end < strlen($text))
$end++;
$sub = substr($text, $pos, $end - $pos);
$text = str_replace($sub, ''.extractLibId($sub).'', $text);
}
return $text;
}
$text = 'some text lib_123123 goes here lib_111';
$text = Lib2($text);
Use preg_replace.
Although it is possible to do what you need without regular expressions, you say you don't want to use them because of performance reasons. I doubt the other solution will be faster, so here is a simple regex to benchmark against:
echo preg_replace("/lib_(\w+)/", '$1', $str);
As shown here: http://codepad.org/xGj78r9r
Ignoring how ridiculous area of optimizing this is, even the simplest implementation with minimal validation already takes only 33% less time than a regex
<?php
function uselessFunction( $val ) {
if( strpos( $val, "lib_" ) !== 0 ) {
return $val;
}
$str = substr( $val, 4 );
return "{$str}";
}
$l = 100000;
$now = microtime(TRUE);
while( $l-- ) {
preg_replace( '/^lib_(.*)$/', "$1", 'lib_someString' );
}
echo (microtime(TRUE)-$now)."\n";
//0.191093
$l = 100000;
$now = microtime(TRUE);
while( $l-- ) {
uselessFunction( "lib_someString" );
}
echo (microtime(TRUE)-$now);
//0.127598
?>
If you're restricted from using a regex, you're going to have difficult time searching for a string you describe as "someString", i.e. not precisely known in advance. If you know the string is exactly lib_12345, for example, then set $lclVar to that string. On the other hand, if you don't know the exact string in advance, you'll have to use a regex via preg_replace() or a similar function.

How do I return a part of text with a certain word in the middle?

If this is the input string:
$input = 'In biology (botany), a "fruit" is a part of a flowering
plant that derives from specific tissues of the flower, mainly one or
more ovaries. Taken strictly, this definition excludes many structures
that are "fruits" in the common sense of the term, such as those
produced by non-flowering plants';
And now I want to perform a search on the word tissues and consequently return only a part of the string, defined by where the result is, like this:
$output = '... of a flowering plant that derives from specific tissues of the flower, mainly one or more ovaries ...';
The search term may be in the middle.
How do I perform the aforementioned?
An alternative to my other answer using preg_match:
$word = 'tissues'
$matches = array();
$found = preg_match("/\b(.{0,30}$word.{0,30})\b/i", $string, $matches);
if ($found == 0) {
// string not found
} else {
$output = $matches[1];
}
This may be better as it uses word boundaries.
EDIT: To surround the search term with a tag, you'll need to slightly alter the regex. This should do it:
$word = 'tissues'
$matches = array();
$found = preg_match("/\b(.{0,30})$word(.{0,30})\b/i", $string, $matches);
if ($found == 0) {
// string not found
} else {
$output = $matches[1] . "<strong>$word</strong>" . $matches[2];
}
User strpos to find the location of the word and substr to extract the quote. For example:
$word = 'tissues'
$pos = strpos($string, $word);
if ($pos === FALSE) {
// string not found
} else {
$start = $pos - 30;
if ($start < 0)
$start = 0;
$output = substr($string, $start, 70);
}
Use stripos for case insensitive search.

Categories