Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
100.00% |
54 / 54 |
|
100.00% |
7 / 7 |
CRAP | |
100.00% |
1 / 1 |
| MediaModerationDatabaseLookup | |
100.00% |
54 / 54 |
|
100.00% |
7 / 7 |
15 | |
100.00% |
1 / 1 |
| __construct | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| fileExistsInScanTable | |
100.00% |
7 / 7 |
|
100.00% |
1 / 1 |
1 | |||
| getMatchStatusForSha1 | |
100.00% |
10 / 10 |
|
100.00% |
1 / 1 |
2 | |||
| getDb | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
2 | |||
| getDateFromTimestamp | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
2 | |||
| newSelectQueryBuilderForScan | |
100.00% |
25 / 25 |
|
100.00% |
1 / 1 |
6 | |||
| getSha1ValuesForScan | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
1 | |||
| 1 | <?php |
| 2 | |
| 3 | namespace MediaWiki\Extension\MediaModeration\Services; |
| 4 | |
| 5 | use ArchivedFile; |
| 6 | use File; |
| 7 | use IDBAccessObject; |
| 8 | use Wikimedia\Rdbms\IConnectionProvider; |
| 9 | use Wikimedia\Rdbms\IReadableDatabase; |
| 10 | use Wikimedia\Rdbms\SelectQueryBuilder; |
| 11 | use Wikimedia\Timestamp\ConvertibleTimestamp; |
| 12 | use Wikimedia\Timestamp\TimestampException; |
| 13 | |
| 14 | class MediaModerationDatabaseLookup { |
| 15 | |
| 16 | public const ANY_MATCH_STATUS = 'any'; |
| 17 | public const POSITIVE_MATCH_STATUS = '1'; |
| 18 | public const NEGATIVE_MATCH_STATUS = '0'; |
| 19 | public const NULL_MATCH_STATUS = null; |
| 20 | |
| 21 | private IConnectionProvider $connectionProvider; |
| 22 | |
| 23 | public function __construct( IConnectionProvider $connectionProvider ) { |
| 24 | $this->connectionProvider = $connectionProvider; |
| 25 | } |
| 26 | |
| 27 | /** |
| 28 | * Returns whether the given $file exists in the mediamoderation_scan table. |
| 29 | * |
| 30 | * @param File|ArchivedFile $file |
| 31 | * @param int $flags IDBAccessObject flags. Does not support READ_LOCKING or READ_EXCLUSIVE |
| 32 | * @return bool |
| 33 | */ |
| 34 | public function fileExistsInScanTable( $file, int $flags = IDBAccessObject::READ_NORMAL ): bool { |
| 35 | $db = $this->getDb( $flags ); |
| 36 | return (bool)$db->newSelectQueryBuilder() |
| 37 | ->select( 'COUNT(*)' ) |
| 38 | ->from( 'mediamoderation_scan' ) |
| 39 | ->where( [ 'mms_sha1' => $file->getSha1() ] ) |
| 40 | ->caller( __METHOD__ ) |
| 41 | ->fetchField(); |
| 42 | } |
| 43 | |
| 44 | /** |
| 45 | * Returns the match status for a given SHA-1. If the SHA-1 does not |
| 46 | * exist in the mediamoderation_scan table, this method will return null. |
| 47 | * |
| 48 | * @param string $sha1 |
| 49 | * @param int $flags IDBAccessObject flags. Does not support READ_LOCKING or READ_EXCLUSIVE |
| 50 | * @return bool|null The match status (null indicates the SHA-1 hasn't been scanned) |
| 51 | */ |
| 52 | public function getMatchStatusForSha1( string $sha1, int $flags = IDBAccessObject::READ_NORMAL ): ?bool { |
| 53 | $db = $this->getDb( $flags ); |
| 54 | $rawMatchStatus = $db->newSelectQueryBuilder() |
| 55 | ->select( 'mms_is_match' ) |
| 56 | ->from( 'mediamoderation_scan' ) |
| 57 | ->where( [ 'mms_sha1' => $sha1 ] ) |
| 58 | ->caller( __METHOD__ ) |
| 59 | ->fetchField(); |
| 60 | if ( is_string( $rawMatchStatus ) ) { |
| 61 | return boolval( $rawMatchStatus ); |
| 62 | } else { |
| 63 | return null; |
| 64 | } |
| 65 | } |
| 66 | |
| 67 | /** |
| 68 | * Gets the IReadableDatabase object for the virtual-mediamoderation DB domain |
| 69 | * for the given $flags. |
| 70 | * |
| 71 | * @param int $flags IDBAccessObject flags. |
| 72 | * @return IReadableDatabase |
| 73 | */ |
| 74 | public function getDb( int $flags ): IReadableDatabase { |
| 75 | if ( $flags & IDBAccessObject::READ_LATEST ) { |
| 76 | return $this->connectionProvider->getPrimaryDatabase( 'virtual-mediamoderation' ); |
| 77 | } else { |
| 78 | return $this->connectionProvider->getReplicaDatabase( 'virtual-mediamoderation' ); |
| 79 | } |
| 80 | } |
| 81 | |
| 82 | /** |
| 83 | * Converts a given timestamp to a string representing the date in the format YYYYMMDD. |
| 84 | * |
| 85 | * @param ConvertibleTimestamp|string|int $timestamp A ConvertibleTimestamp or timestamp recognised by |
| 86 | * ConvertibleTimestamp. |
| 87 | * @return string The timestamp as a date in the form YYYYMMDD |
| 88 | * @throws TimestampException If the $timestamp cannot be parsed |
| 89 | */ |
| 90 | public function getDateFromTimestamp( $timestamp ): string { |
| 91 | // Convert the $timestamp to a ConvertibleTimestamp instance |
| 92 | if ( !( $timestamp instanceof ConvertibleTimestamp ) ) { |
| 93 | $timestamp = new ConvertibleTimestamp( $timestamp ); |
| 94 | } |
| 95 | // Get the timestamp as in TS_MW form (YYYMMDDHHMMSS) |
| 96 | $timestampAsTSMW = $timestamp->getTimestamp( TS_MW ); |
| 97 | // Return the first 8 characters of the TS_MW timestamp, which |
| 98 | // means the YYYYMMDD part. |
| 99 | return substr( $timestampAsTSMW, 0, 8 ); |
| 100 | } |
| 101 | |
| 102 | /** |
| 103 | * Returns a SelectQueryBuilder that can be used to query SHA-1 values for a scan. |
| 104 | * |
| 105 | * The parameters to this method allow filtering for rows with a specific match status and/or rows that were |
| 106 | * last checked before or at a particular date. |
| 107 | * |
| 108 | * @param ConvertibleTimestamp|int|string|null $lastChecked See ::getSha1ValuesForScan |
| 109 | * @param string $direction See ::getSha1ValuesForScan |
| 110 | * @param array $excludedSha1Values See ::getSha1ValuesForScan |
| 111 | * @param string|null $matchStatus See ::getSha1ValuesForScan |
| 112 | * @return SelectQueryBuilder |
| 113 | * @throws TimestampException If the $lastChecked timestamp could not be parsed as a valid timestamp. |
| 114 | */ |
| 115 | protected function newSelectQueryBuilderForScan( |
| 116 | $lastChecked, string $direction, array $excludedSha1Values, ?string $matchStatus = self::ANY_MATCH_STATUS |
| 117 | ): SelectQueryBuilder { |
| 118 | // Get a replica DB connection. |
| 119 | $dbr = $this->getDb( IDBAccessObject::READ_NORMAL ); |
| 120 | // Create a SelectQueryBuilder that reads from the mediamoderation_scan table. |
| 121 | // The fields to read is set by the callers of this method. |
| 122 | $selectQueryBuilder = $dbr->newSelectQueryBuilder() |
| 123 | ->from( 'mediamoderation_scan' ); |
| 124 | if ( $lastChecked === null ) { |
| 125 | // If $lastChecked is null, then only get rows with the last checked value as null. |
| 126 | $selectQueryBuilder->where( [ 'mms_last_checked' => null ] ); |
| 127 | } else { |
| 128 | // If $lastChecked is not null, then treat it as a timestamp. |
| 129 | // Then using this timestamp as a date in the form YYYYMMDD, filter |
| 130 | // for rows with a smaller last checked date or which have never been |
| 131 | // checked (last checked as null). |
| 132 | $lastCheckedAsMWTimestamp = $this->getDateFromTimestamp( $lastChecked ); |
| 133 | $selectQueryBuilder->where( |
| 134 | $dbr->expr( |
| 135 | 'mms_last_checked', |
| 136 | '<=', |
| 137 | $lastCheckedAsMWTimestamp |
| 138 | )->or( 'mms_last_checked', '=', null ) |
| 139 | ); |
| 140 | } |
| 141 | if ( $dbr->getType() === 'postgres' ) { |
| 142 | // Postgres DBs treat NULLs by default as larger than non-NULL values. |
| 143 | // This is the opposite for Mariadb / SQLite. Postgres should have the same |
| 144 | // behaviour as Mariadb / SQLite. By using NULLS FIRST and NULLS LAST |
| 145 | // we can control where the NULL comes in the results list for postgres DBs |
| 146 | if ( $direction === SelectQueryBuilder::SORT_ASC ) { |
| 147 | $direction .= ' NULLS FIRST'; |
| 148 | } else { |
| 149 | $direction .= ' NULLS LAST'; |
| 150 | } |
| 151 | } |
| 152 | // Filter by match status if $matchStatus does not indicate to |
| 153 | // allow rows with any match status. |
| 154 | if ( $matchStatus !== self::ANY_MATCH_STATUS ) { |
| 155 | $selectQueryBuilder->where( [ 'mms_is_match' => $matchStatus ] ); |
| 156 | } |
| 157 | // Exclude the SHA-1 values specified by the caller, if any are provided. |
| 158 | if ( count( $excludedSha1Values ) ) { |
| 159 | $selectQueryBuilder->where( $dbr->expr( |
| 160 | 'mms_sha1', '!=', $excludedSha1Values |
| 161 | ) ); |
| 162 | } |
| 163 | // Return the constructed SelectQueryBuilder after adding the order by field. |
| 164 | return $selectQueryBuilder |
| 165 | ->orderBy( 'mms_last_checked', $direction ); |
| 166 | } |
| 167 | |
| 168 | /** |
| 169 | * Gets $limit rows from the mediamoderation_scan table that have mms_last_checked less than the timestamp |
| 170 | * in $lastChecked. The returned rows are ordered by last checked timestamp in $direction. |
| 171 | * |
| 172 | * @param int $limit The maximum number of scan rows to return |
| 173 | * @param ConvertibleTimestamp|int|string|null $lastChecked Filters for scan rows that have been last checked |
| 174 | * on or before this date. If null, only include files which have never been checked. If not null, treats this as |
| 175 | * a timestamp that can be parsed by ConvertibleTimestamp. |
| 176 | * @param string $direction Either SelectQueryBuilder::SORT_ASC or ::SORT_DESC. Used to control |
| 177 | * whether to start at the rows with the newest or oldest last checked timestamp. No-op if $lastChecked |
| 178 | * is null. |
| 179 | * @param array $excludedSha1Values SHA-1 values to exclude from the returned array. |
| 180 | * @param string|null $matchStatus Filter for rows that have this match status. Any constants of this |
| 181 | * service in the format ::*_MATCH_STATUS can be passed in this parameter. The default is to not filter |
| 182 | * by match status (using ::ANY_MATCH_STATUS). |
| 183 | * @return array The SHA-1 values from the selected rows |
| 184 | * @throws TimestampException If the $lastChecked timestamp could not be parsed as a valid timestamp. |
| 185 | */ |
| 186 | public function getSha1ValuesForScan( |
| 187 | int $limit, $lastChecked, string $direction, |
| 188 | array $excludedSha1Values, ?string $matchStatus |
| 189 | ): array { |
| 190 | // Return up to $limit SHA-1 values that match the given criteria. |
| 191 | return $this->newSelectQueryBuilderForScan( $lastChecked, $direction, $excludedSha1Values, $matchStatus ) |
| 192 | ->select( 'mms_sha1' ) |
| 193 | ->limit( $limit ) |
| 194 | ->fetchFieldValues(); |
| 195 | } |
| 196 | } |