regex case insensitive and with/without whitespace - php

Question

Not being that knowledgable in regex patterns and after reading all wikis and references I found I'm having problems altering a pattern for word detection and higlighting.
I found a function on another stackoverflow answer that did everything it was needed but now I found out it misses out on a few things
The function is:
function ParserGlossario($texto, $termos) {
$padrao = '\1\2\3';
if (empty($termos)) {
return $texto;
}
if (is_array($termos)) {
$substituir = array();
$com = array();
foreach ($termos as $key => $value) {
$key = $value;
$value = $padrao;
// $key = '([\s])(' . $key . ')([\s\.\,\!\?\<])';
$key = '([\s])(' . $key . ')([\s\.\,\!\?\<])';
$substituir[] = '|' . $key . '|ix';
$com[] = empty($value) ? $padrao : $value;
}
return preg_replace($substituir, $com, $texto);
} else {
$termos = '([\s])(' . $termos . ')([\s])';
return preg_replace('|'.$termos.'|i', $padrao, $texto);
}
}
Some words are not being highlighted (the ones marked with red arrows):
And I don't know if it helps, but here is the array of "terms" that is used to search the text:
EDIT. The string being searched is just plain text:
Abaxial Xxxxx acaule Acaule xxxxxx xxx; xxxxx xxx Abaxial esporos.
abaxial
EDIT. Added PHP code fiddle
http://phpfiddle.org/main/code/079ad24318f554d9f2ba
Any help? I really don't know much about regexes...

Answer 1

try
$key = '(^|\b)(' . $key . ')\b';
insetad of
$key = '([\s])(' . $key . ')([\s\.\,\!\?\<])';
should help. Your matches still will be in the second group but there will be no third and I think the first should not be touched, so I believe this
$padrao = '\1\2\3';
is better to be as
$padrao = '$2';
and forgot (sorry):
change
$substituir[] = '|' . $key . '|ix';
to
$substituir[] = '#' . $key . '#ix';
And also I would use a string
$com = empty($value) ? $padrao : $value;
instead of array, it's not needed in this case.

Answer 2

Let us look together on value of $key for example for array element acaule.
([\s])(acaule)([\s\.\,\!\?\<])
There are 3 marking groups defined by 3 pairs of (...).
The first marking group matches any whitespace character. If there is no whitespace character like for Abaxial at beginning of the string, the word is ignored.
Putting \s into a character class, i.e. within [...] is not really needed here as \s is itself a character class. ([\s]) and (\s) are equal.
The second marking group matches just the word from array.
The third marking group matches
either any whitespace character,
or a period,
or a comma,
or an exclamation mark,
or a question mark, i.e. the standard punctuation marks,
or a left angle bracket (from an HTML or XML tag).
A semicolon or colon is not matched and other non word characters are also ignored for a positive match.
If there is none of those characters like for abaxial at end of the string, the search is negative.
By the way: ([\s.,!?<]) is equal ([\s\.\,\!\?\<]) as only \ and ] (always) and - (depending on position) must be escaped with a backslash within a character class definition to be interpreted as literal character. Well, [ should be also escaped with a backslash within [...] for easier reading.
So it is clear why Abaxial at beginning of string and abaxial at end of the string are not matched.
But why is Acaule not matched?
Well, there is left to this word acaule with a space left and a space right as required for a positive match. So the space right of acaule was already taken for this positive match. Therefore for Acaule there is no whitespace character anymore left to this word.
There is \b which means word boundary not matching any character which might be used together with \W*? instead of ([\s]) and instead of ([\s\.\,\!\?\<]) to avoid matching substrings within a word.
Possible would be something like
$key = '(\W*?)(\b' . $key . '\b)(\W*?)';
\W*? means any non word character 0 or more times non-greedy.
\W? means any non word character 0 or 1 times and could be also used in first and third capturing group if that is better for the replace.
But what is the right search string depends on what you want as result of the replace.
I don't have a PHP interpreter installed at all and therefore can't try it out what your PHP code does on replace and therefore what you would like to see after replace done on provided example string.