Smarty

[toma]

Hey all. I recently created this page,
http://smarty.incutio.com/?page=modifier.google_highlight
but I'm hoping someone has a better way to parse the search string than I use:

Currently I pull out all the double quoted strings before exploding the search term on spaces, but I think there's probably a way to do this using preg_split() along the lines of
/\".*\"|:space:/
but I know for sure that doesn't work and I'm not regexguru enough to figure it out.

Any help is appreciated!
Tom

[mohrt]

You're probably safest to leave this in two passes, first pull out the quoted blocks, then split by spaces.

(untested)

// get all double quoted strings out
preg_match_all('!".*?"!', $string, $_match);

// split by spaces
preg_split('!\s+!', $string, PREG_SPLIT_NO_EMPTY);


Monte

[toma]

preg_match_all works well but doesn't leave me with an easily parseable string for the non-quoted terms.
preg_split('!".*?"!', $search);

This is what I have now:

// get all double quoted strings then terms
preg_match_all('!".*?"!', $search, $_match);
$terms = explode(' ', implode('', preg_split('!".*?"!', $search)));
if (sizeof($_match[0])) {
$terms = array_merge($_match[0], $terms);
}


But it can still contain empty elements.

[boots]

mohrt's regex does not handle escaped quotes, so careful!

[toma]

Thanks boots. That got me started. I changed how you did it (and added a ? to your regex) and got it down to two lines.

However, I just found a bug w/ the replacing for this. A search term of 'style' will make a hell of a mess because previous 'style's inserted by other terms are replaced. I really don't like the idea of tokenizing the text but I don't see any other solution.

Do you?

Thanks again,
Tom

[boots]

To be honest, I haven't looked at your page at the wiki yet, so I'm not sure how your processing is occuring For example, I don't see why you say that styles will replace previous styles.

There are other ways of doing this without using a string stack. I find it convenient to use a stack, but if you don't like it, I also sometimes use preg_replace_callback which allows for more sophisticated inline handling.

See: http://www.php.net/manual/en/function.preg-replace-callback.php

A few things about callback: only the matched array is passed to the callback function--if you need to keep static data or use data from elsewhere, you will have to co-ordinate that prior to using preg_replace_callback. You can register an arbitrary function in any object by using the array($object, 'function') notation.

You can also use while (preg_match (...) ) to process the strings without creating an intervening stack. You can use a variation of this to allow you to abort processing mid-way through the scan if you come across a syntax error in the input.

Questions:

[toma]

>>To be honest, I haven't looked at your page at the wiki yet, so I'm not sure how your processing is occuring For example, I d

I don't know what to say to this comment except, perhaps, check it out?

I've updated the wiki as I've refined the processing. The search string parsing is to a point I'm happy w/ it but my problem lies in the
$text = preg_replace("/($val)/i", $style . '$1' . '</b>', $text);
for each search term. Previously found search terms will have <b style=... in them and a search for 'style' later on will find that code.

I think I need a regex like
/[^<b style...]($val)[^</b>]/
(think of this notation as phoenetic regex

Tom

Edit:
As for .*? .* etc, my final code is
preg_match_all( '/"(.*?)"/', $search, $_quotes);
$_terms = array_merge((array)$_quotes[1], explode(' ', preg_replace( '/".*?"/', ' ', $search )));

Using just '/".*"/' resulted in missing multiple quote-enclosed search terms. "blah and" stuff "dna halb"
was found as
blah and" stuff "dna halb
.*? fixed it.

[boots]

[mohrt]

[boots]

[mohrt]

[boots]

@mohrt: Thanks for the reference! I currently only have a (ragged) printed copy of the material from the PHP manual and though there is a lot in there, there is not a lot of practical information to springboard from.

For those interested, the O'Reilly page for the book is at http://www.oreilly.com/catalog/regex/chapter/index.html and has a link to a sample chapter Looks good!!

ps. good show on version 2.4!

[mohrt]

btw, that O'Reilly book talks about lookahead and negative lookahead, but says negative lookbehind is not supported. However, Perl 5+ supports negative lookbehind under the condition that it is static.

example:

(?<!ABC)D

will match D only if preceeded by ABC. The ABC part must be a static value, you can't look for a regex.

In PHP if you want to match a double quote not preceeded by a backslash:

preg_match('/(?<!\\\\)"/', $foo);

You must be careful to escape the escape

Monte

[boots]

I just picked up the book. Swimmingly good if you are into that sort of stuff!!