dokuwiki\Search\FulltextSearch::snippet()

Complexity

Conditions	14
Paths	7

Size

Total Lines

106

Duplication

Lines	0
Ratio	0 %

Importance

Changes

0

<?php
    public function snippet($id, $highlight)
    {
        $text = rawWiki($id);
        $text = str_replace("\xC2\xAD",'',$text); // remove soft-hyphens
        $evdata = array(
            'id'        => $id,
            'text'      => &$text,
            'highlight' => &$highlight,
            'snippet'   => '',
        );

        $evt = new Event('FULLTEXT_SNIPPET_CREATE', $evdata);
        if ($evt->advise_before()) {
            $match = array();
            $snippets = array();
            $utf8_offset = $offset = $end = 0;
            $len = Utf8\PhpString::strlen($text);

            // build a regexp from the phrases to highlight
            $re1 = '(' .
                join(
                    '|',
                    array_map(
                        [$this, 'snippetRePreprocess'],
                        array_map(
                            'preg_quote_cb',
                            array_filter((array) $highlight)
                        )
                    )
                ) .
                ')';
            $re2 = "$re1.{0,75}(?!\\1)$re1";
            $re3 = "$re1.{0,45}(?!\\1)$re1.{0,45}(?!\\1)(?!\\2)$re1";

            for ($cnt=4; $cnt--;) {
                if (0) {
                } elseif (preg_match('/'.$re3.'/iu', $text, $match, PREG_OFFSET_CAPTURE, $offset)) {
                } elseif (preg_match('/'.$re2.'/iu', $text, $match, PREG_OFFSET_CAPTURE, $offset)) {
                } elseif (preg_match('/'.$re1.'/iu', $text, $match, PREG_OFFSET_CAPTURE, $offset)) {
                } else {
                    break;
                }

                list($str, $idx) = $match[0];

                // convert $idx (a byte offset) into a utf8 character offset
                $utf8_idx = Utf8\PhpString::strlen(substr($text, 0, $idx));
                $utf8_len = Utf8\PhpString::strlen($str);

                // establish context, 100 bytes surrounding the match string
                // first look to see if we can go 100 either side,
                // then drop to 50 adding any excess if the other side can't go to 50,
                $pre = min($utf8_idx - $utf8_offset, 100);
                $post = min($len - $utf8_idx - $utf8_len, 100);

                if ($pre > 50 && $post > 50) {
                    $pre = $post = 50;
                } elseif ($pre > 50) {
                    $pre = min($pre, 100 - $post);
                } elseif ($post > 50) {
                    $post = min($post, 100 - $pre);
                } elseif ($offset == 0) {
                    // both are less than 50, means the context is the whole string
                    // make it so and break out of this loop - there is no need for the
                    // complex snippet calculations
                    $snippets = array($text);
                    break;
                }

                // establish context start and end points, try to append to previous
                // context if possible
                $start = $utf8_idx - $pre;
                $append = ($start < $end) ? $end : false;  // still the end of the previous context snippet
                $end = $utf8_idx + $utf8_len + $post;      // now set it to the end of this context

                if ($append) {

The expression $append of type integer|false is loosely compared to true; this is ambiguous if the integer can be zero. You might want to explicitly use !== null instead.

                    $snippets[count($snippets)-1] .= Utf8\PhpString::substr($text, $append, $end-$append);
                } else {
                    $snippets[] = Utf8\PhpString::substr($text, $start, $end-$start);
                }

                // set $offset for next match attempt
                // continue matching after the current match
                // if the current match is not the longest possible match starting at the current offset
                // this prevents further matching of this snippet but for possible matches of length
                // smaller than match length + context (at least 50 characters) this match is part of the context
                $utf8_offset = $utf8_idx + $utf8_len;
                $offset = $idx + strlen(Utf8\PhpString::substr($text, $utf8_idx, $utf8_len));
                $offset = Utf8\Clean::correctIdx($text, $offset);
            }

            $m = "\1";
            $snippets = preg_replace('/'.$re1.'/iu', $m.'$1'.$m, $snippets);
            $snippet = preg_replace(
                '/' . $m . '([^' . $m . ']*?)' . $m . '/iu',
                '<strong class="search_hit">$1</strong>',
                hsc(join('... ', $snippets))
            );

            $evdata['snippet'] = $snippet;
        }
        $evt->advise_after();
        unset($evt);

        return $evdata['snippet'];
    }