Differential D18672 Diff 44838 src/infrastructure/query/policy/PhabricatorCursorPagedPolicyAwareQuery.php

Changeset View

Standalone View

src/infrastructure/query/policy/PhabricatorCursorPagedPolicyAwareQuery.php

Show First 20 Lines • Show All 1,694 Lines • ▼ Show 20 Lines	foreach ($this->ferretTokens as $fulltext_token) {
foreach ($ngrams as $ngram) {		foreach ($ngrams as $ngram) {
$flat[] = array(		$flat[] = array(
'table' => $ngram_table,		'table' => $ngram_table,
'ngram' => $ngram,		'ngram' => $ngram,
);		);
}		}
}		}

		// Remove common ngrams, like "the", which occur too frequently in
		// documents to be useful in constraining the query. The best ngrams
		// are obscure sequences which occur in very few documents.

		if ($flat) {
		$common_ngrams = queryfx_all(
		$conn,
		'SELECT ngram FROM %T WHERE ngram IN (%Ls)',
		$engine->getCommonNgramsTableName(),
		ipull($flat, 'ngram'));
		$common_ngrams = ipull($common_ngrams, 'ngram', 'ngram');

		foreach ($flat as $key => $spec) {
		$ngram = $spec['ngram'];
		if (isset($common_ngrams[$ngram])) {
		unset($flat[$key]);
		continue;
		}

		// NOTE: MySQL discards trailing whitespace in CHAR(X) columns.
		$trim_ngram = rtrim($ngram, ' ');
		if (isset($common_ngrams[$trim_ngram])) {
		unset($flat[$key]);
		continue;
		}
		}
		}

// MySQL only allows us to join a maximum of 61 tables per query. Each		// MySQL only allows us to join a maximum of 61 tables per query. Each
// ngram is going to cost us a join toward that limit, so if the user		// ngram is going to cost us a join toward that limit, so if the user
// specified a very long query string, just pick 16 of the ngrams		// specified a very long query string, just pick 16 of the ngrams
// at random.		// at random.
if (count($flat) > 16) {		if (count($flat) > 16) {
shuffle($flat);		shuffle($flat);
$flat = array_slice($flat, 0, 16);		$flat = array_slice($flat, 0, 16);
}		}
▲ Show 20 Lines • Show All 1,063 Lines • Show Last 20 Lines