Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
55.47% |
71 / 128 |
|
25.00% |
1 / 4 |
CRAP | |
0.00% |
0 / 1 |
| MigrateFileRepoLayout | |
55.47% |
71 / 128 |
|
25.00% |
1 / 4 |
122.43 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
1 | |||
| execute | |
51.75% |
59 / 114 |
|
0.00% |
0 / 1 |
108.87 | |||
| getRepo | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| runBatch | |
85.71% |
6 / 7 |
|
0.00% |
0 / 1 |
3.03 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * Copy all files in FileRepo to an originals container using SHA1 paths. |
| 4 | * |
| 5 | * This program is free software; you can redistribute it and/or modify |
| 6 | * it under the terms of the GNU General Public License as published by |
| 7 | * the Free Software Foundation; either version 2 of the License, or |
| 8 | * (at your option) any later version. |
| 9 | * |
| 10 | * This program is distributed in the hope that it will be useful, |
| 11 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 13 | * GNU General Public License for more details. |
| 14 | * |
| 15 | * You should have received a copy of the GNU General Public License along |
| 16 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 17 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
| 18 | * http://www.gnu.org/copyleft/gpl.html |
| 19 | * |
| 20 | * @file |
| 21 | * @ingroup Maintenance |
| 22 | */ |
| 23 | |
| 24 | use Wikimedia\FileBackend\FileBackend; |
| 25 | |
| 26 | // @codeCoverageIgnoreStart |
| 27 | require_once __DIR__ . '/Maintenance.php'; |
| 28 | // @codeCoverageIgnoreEnd |
| 29 | |
| 30 | /** |
| 31 | * Copy all files in FileRepo to an originals container using SHA1 paths. |
| 32 | * |
| 33 | * This script should be run while the repo is still set to the old layout. |
| 34 | * |
| 35 | * @ingroup Maintenance |
| 36 | */ |
| 37 | class MigrateFileRepoLayout extends Maintenance { |
| 38 | public function __construct() { |
| 39 | parent::__construct(); |
| 40 | $this->addDescription( 'Copy files in repo to a different layout.' ); |
| 41 | $this->addOption( 'oldlayout', "Old layout; one of 'name' or 'sha1'", true, true ); |
| 42 | $this->addOption( 'newlayout', "New layout; one of 'name' or 'sha1'", true, true ); |
| 43 | $this->addOption( 'since', "Copy only files from after this timestamp", false, true ); |
| 44 | $this->setBatchSize( 50 ); |
| 45 | } |
| 46 | |
| 47 | public function execute() { |
| 48 | $oldLayout = $this->getOption( 'oldlayout' ); |
| 49 | if ( !in_array( $oldLayout, [ 'name', 'sha1' ] ) ) { |
| 50 | $this->fatalError( "Invalid old layout." ); |
| 51 | } |
| 52 | $newLayout = $this->getOption( 'newlayout' ); |
| 53 | if ( !in_array( $newLayout, [ 'name', 'sha1' ] ) ) { |
| 54 | $this->fatalError( "Invalid new layout." ); |
| 55 | } |
| 56 | $since = $this->getOption( 'since' ); |
| 57 | |
| 58 | $repo = $this->getRepo(); |
| 59 | |
| 60 | $be = $repo->getBackend(); |
| 61 | if ( $be instanceof FileBackendDBRepoWrapper ) { |
| 62 | // avoid path translations for this script |
| 63 | $be = $be->getInternalBackend(); |
| 64 | } |
| 65 | |
| 66 | $dbw = $repo->getPrimaryDB(); |
| 67 | |
| 68 | $origBase = $be->getContainerStoragePath( "{$repo->getName()}-original" ); |
| 69 | $startTime = wfTimestampNow(); |
| 70 | |
| 71 | // Do current and archived versions... |
| 72 | $conds = []; |
| 73 | if ( $since ) { |
| 74 | $conds[] = $dbw->expr( 'img_timestamp', '>=', $dbw->timestamp( $since ) ); |
| 75 | } |
| 76 | |
| 77 | $batchSize = $this->getBatchSize(); |
| 78 | $batch = []; |
| 79 | $lastName = ''; |
| 80 | do { |
| 81 | $res = $dbw->newSelectQueryBuilder() |
| 82 | ->select( [ 'img_name', 'img_sha1' ] ) |
| 83 | ->from( 'image' ) |
| 84 | ->where( $dbw->expr( 'img_name', '>', $lastName ) ) |
| 85 | ->andWhere( $conds ) |
| 86 | ->orderBy( 'img_name' ) |
| 87 | ->limit( $batchSize ) |
| 88 | ->caller( __METHOD__ )->fetchResultSet(); |
| 89 | |
| 90 | foreach ( $res as $row ) { |
| 91 | $lastName = $row->img_name; |
| 92 | /** @var LocalFile $file */ |
| 93 | $file = $repo->newFile( $row->img_name ); |
| 94 | // Check in case SHA1 rows are not populated for some files |
| 95 | $sha1 = strlen( $row->img_sha1 ) ? $row->img_sha1 : $file->getSha1(); |
| 96 | |
| 97 | if ( !strlen( $sha1 ) ) { |
| 98 | $this->error( "Image SHA-1 not known for {$row->img_name}." ); |
| 99 | } else { |
| 100 | if ( $oldLayout === 'sha1' ) { |
| 101 | $spath = "{$origBase}/{$sha1[0]}/{$sha1[1]}/{$sha1[2]}/{$sha1}"; |
| 102 | } else { |
| 103 | $spath = $file->getPath(); |
| 104 | } |
| 105 | |
| 106 | if ( $newLayout === 'sha1' ) { |
| 107 | $dpath = "{$origBase}/{$sha1[0]}/{$sha1[1]}/{$sha1[2]}/{$sha1}"; |
| 108 | } else { |
| 109 | $dpath = $file->getPath(); |
| 110 | } |
| 111 | |
| 112 | $status = $be->prepare( [ |
| 113 | 'dir' => dirname( $dpath ), 'bypassReadOnly' => true ] ); |
| 114 | if ( !$status->isOK() ) { |
| 115 | $this->error( $status ); |
| 116 | } |
| 117 | |
| 118 | $batch[] = [ 'op' => 'copy', 'overwrite' => true, |
| 119 | 'src' => $spath, 'dst' => $dpath, 'img' => $row->img_name ]; |
| 120 | } |
| 121 | |
| 122 | foreach ( $file->getHistory() as $ofile ) { |
| 123 | $sha1 = $ofile->getSha1(); |
| 124 | if ( !strlen( $sha1 ) ) { |
| 125 | $this->error( "Image SHA-1 not set for {$ofile->getArchiveName()}." ); |
| 126 | continue; |
| 127 | } |
| 128 | |
| 129 | if ( $oldLayout === 'sha1' ) { |
| 130 | $spath = "{$origBase}/{$sha1[0]}/{$sha1[1]}/{$sha1[2]}/{$sha1}"; |
| 131 | } elseif ( $ofile->isDeleted( File::DELETED_FILE ) ) { |
| 132 | $spath = $be->getContainerStoragePath( "{$repo->getName()}-deleted" ) . |
| 133 | '/' . $repo->getDeletedHashPath( $sha1 ) . |
| 134 | $sha1 . '.' . $ofile->getExtension(); |
| 135 | } else { |
| 136 | $spath = $ofile->getPath(); |
| 137 | } |
| 138 | |
| 139 | if ( $newLayout === 'sha1' ) { |
| 140 | $dpath = "{$origBase}/{$sha1[0]}/{$sha1[1]}/{$sha1[2]}/{$sha1}"; |
| 141 | } else { |
| 142 | $dpath = $ofile->getPath(); |
| 143 | } |
| 144 | |
| 145 | $status = $be->prepare( [ |
| 146 | 'dir' => dirname( $dpath ), 'bypassReadOnly' => true ] ); |
| 147 | if ( !$status->isOK() ) { |
| 148 | $this->error( $status ); |
| 149 | } |
| 150 | $batch[] = [ 'op' => 'copy', 'overwrite' => true, |
| 151 | 'src' => $spath, 'dst' => $dpath, 'img' => $ofile->getArchiveName() ]; |
| 152 | } |
| 153 | |
| 154 | if ( count( $batch ) >= $batchSize ) { |
| 155 | $this->runBatch( $batch, $be ); |
| 156 | $batch = []; |
| 157 | } |
| 158 | } |
| 159 | } while ( $res->numRows() ); |
| 160 | |
| 161 | if ( count( $batch ) ) { |
| 162 | $this->runBatch( $batch, $be ); |
| 163 | } |
| 164 | |
| 165 | // Do deleted versions... |
| 166 | $conds = []; |
| 167 | if ( $since ) { |
| 168 | $conds[] = $dbw->expr( 'fa_deleted_timestamp', '>=', $dbw->timestamp( $since ) ); |
| 169 | } |
| 170 | |
| 171 | $batch = []; |
| 172 | $lastId = 0; |
| 173 | do { |
| 174 | $res = $dbw->newSelectQueryBuilder() |
| 175 | ->select( [ 'fa_storage_key', 'fa_id', 'fa_name' ] ) |
| 176 | ->from( 'filearchive' ) |
| 177 | ->where( $dbw->expr( 'fa_id', '>', $lastId ) ) |
| 178 | ->andWhere( $conds ) |
| 179 | ->orderBy( 'fa_id' ) |
| 180 | ->limit( $batchSize ) |
| 181 | ->caller( __METHOD__ )->fetchResultSet(); |
| 182 | |
| 183 | foreach ( $res as $row ) { |
| 184 | $lastId = $row->fa_id; |
| 185 | $sha1Key = $row->fa_storage_key; |
| 186 | if ( !strlen( $sha1Key ) ) { |
| 187 | $this->error( "Image SHA-1 not set for file #{$row->fa_id} (deleted)." ); |
| 188 | continue; |
| 189 | } |
| 190 | $sha1 = substr( $sha1Key, 0, strpos( $sha1Key, '.' ) ); |
| 191 | |
| 192 | if ( $oldLayout === 'sha1' ) { |
| 193 | $spath = "{$origBase}/{$sha1[0]}/{$sha1[1]}/{$sha1[2]}/{$sha1}"; |
| 194 | } else { |
| 195 | $spath = $be->getContainerStoragePath( "{$repo->getName()}-deleted" ) . |
| 196 | '/' . $repo->getDeletedHashPath( $sha1Key ) . $sha1Key; |
| 197 | } |
| 198 | |
| 199 | if ( $newLayout === 'sha1' ) { |
| 200 | $dpath = "{$origBase}/{$sha1[0]}/{$sha1[1]}/{$sha1[2]}/{$sha1}"; |
| 201 | } else { |
| 202 | $dpath = $be->getContainerStoragePath( "{$repo->getName()}-deleted" ) . |
| 203 | '/' . $repo->getDeletedHashPath( $sha1Key ) . $sha1Key; |
| 204 | } |
| 205 | |
| 206 | $status = $be->prepare( [ |
| 207 | 'dir' => dirname( $dpath ), 'bypassReadOnly' => true ] ); |
| 208 | if ( !$status->isOK() ) { |
| 209 | $this->error( $status ); |
| 210 | } |
| 211 | |
| 212 | $batch[] = [ 'op' => 'copy', 'src' => $spath, 'dst' => $dpath, |
| 213 | 'overwriteSame' => true, 'img' => "(ID {$row->fa_id}) {$row->fa_name}" ]; |
| 214 | |
| 215 | if ( count( $batch ) >= $batchSize ) { |
| 216 | $this->runBatch( $batch, $be ); |
| 217 | $batch = []; |
| 218 | } |
| 219 | } |
| 220 | } while ( $res->numRows() ); |
| 221 | |
| 222 | if ( count( $batch ) ) { |
| 223 | $this->runBatch( $batch, $be ); |
| 224 | } |
| 225 | |
| 226 | $this->output( "Done (started $startTime)\n" ); |
| 227 | } |
| 228 | |
| 229 | protected function getRepo() { |
| 230 | return $this->getServiceContainer()->getRepoGroup()->getLocalRepo(); |
| 231 | } |
| 232 | |
| 233 | /** |
| 234 | * @param array[] $ops |
| 235 | * @param FileBackend $be |
| 236 | */ |
| 237 | protected function runBatch( array $ops, FileBackend $be ) { |
| 238 | $this->output( "Migrating file batch:\n" ); |
| 239 | foreach ( $ops as $op ) { |
| 240 | $this->output( "\"{$op['img']}\" (dest: {$op['dst']})\n" ); |
| 241 | } |
| 242 | |
| 243 | $status = $be->doOperations( $ops, [ 'bypassReadOnly' => true ] ); |
| 244 | if ( !$status->isOK() ) { |
| 245 | $this->error( $status ); |
| 246 | } |
| 247 | |
| 248 | $this->output( "Batch done\n\n" ); |
| 249 | } |
| 250 | } |
| 251 | |
| 252 | // @codeCoverageIgnoreStart |
| 253 | $maintClass = MigrateFileRepoLayout::class; |
| 254 | require_once RUN_MAINTENANCE_IF_MAIN; |
| 255 | // @codeCoverageIgnoreEnd |