Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
85.90% |
67 / 78 |
|
44.44% |
4 / 9 |
CRAP | |
0.00% |
0 / 1 |
| BaseDump | |
87.01% |
67 / 77 |
|
44.44% |
4 / 9 |
45.86 | |
0.00% |
0 / 1 |
| __construct | |
66.67% |
4 / 6 |
|
0.00% |
0 / 1 |
2.15 | |||
| prefetch | |
100.00% |
26 / 26 |
|
100.00% |
1 / 1 |
14 | |||
| debug | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| nextPage | |
83.33% |
10 / 12 |
|
0.00% |
0 / 1 |
5.12 | |||
| nextRev | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
3 | |||
| nextText | |
66.67% |
2 / 3 |
|
0.00% |
0 / 1 |
2.15 | |||
| skipTo | |
81.82% |
9 / 11 |
|
0.00% |
0 / 1 |
7.29 | |||
| nodeContents | |
72.73% |
8 / 11 |
|
0.00% |
0 / 1 |
7.99 | |||
| close | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * Helper class for the --prefetch option of dumpTextPass.php |
| 4 | * |
| 5 | * Copyright © 2005 Brooke Vibber <bvibber@wikimedia.org> |
| 6 | * https://www.mediawiki.org/ |
| 7 | * |
| 8 | * @license GPL-2.0-or-later |
| 9 | * @file |
| 10 | * @ingroup Maintenance |
| 11 | */ |
| 12 | |
| 13 | namespace MediaWiki\Export; |
| 14 | |
| 15 | use MediaWiki\Revision\SlotRecord; |
| 16 | use XMLReader; |
| 17 | |
| 18 | /** |
| 19 | * Readahead helper for making large MediaWiki data dumps; |
| 20 | * reads in a previous XML dump to sequentially prefetch text |
| 21 | * records already normalized and decompressed. |
| 22 | * |
| 23 | * This can save load on the external database servers, hopefully. |
| 24 | * |
| 25 | * Assumes that dumps will be recorded in the canonical order: |
| 26 | * - ascending by page_id |
| 27 | * - ascending by rev_id within each page |
| 28 | * - text contents are immutable and should not change once |
| 29 | * recorded, so the previous dump is a reliable source |
| 30 | * |
| 31 | * @ingroup Maintenance |
| 32 | */ |
| 33 | class BaseDump { |
| 34 | /** @var XMLReader|null */ |
| 35 | protected $reader = null; |
| 36 | /** @var bool */ |
| 37 | protected $atEnd = false; |
| 38 | /** @var bool */ |
| 39 | protected $atPageEnd = false; |
| 40 | /** @var int */ |
| 41 | protected $lastPage = 0; |
| 42 | /** @var int */ |
| 43 | protected $lastRev = 0; |
| 44 | /** @var string[]|null */ |
| 45 | protected $infiles = null; |
| 46 | |
| 47 | /** |
| 48 | * @param string $infile |
| 49 | */ |
| 50 | public function __construct( $infile ) { |
| 51 | $this->infiles = explode( ';', $infile ); |
| 52 | $this->reader = new XMLReader(); |
| 53 | $infile = array_shift( $this->infiles ); |
| 54 | if ( !$this->reader->open( $infile, null, LIBXML_PARSEHUGE ) ) { |
| 55 | $this->debug( __METHOD__ . ' was unable to open xml' ); |
| 56 | $this->atEnd = true; |
| 57 | } |
| 58 | } |
| 59 | |
| 60 | /** |
| 61 | * Attempts to fetch the text of a particular page revision |
| 62 | * from the dump stream. May return null if the page is |
| 63 | * unavailable. |
| 64 | * |
| 65 | * @param int $page ID number of page to read |
| 66 | * @param int $rev ID number of revision to read |
| 67 | * @param string $slot Role name of the slot to read |
| 68 | * @return string|null |
| 69 | */ |
| 70 | public function prefetch( $page, $rev, $slot = SlotRecord::MAIN ) { |
| 71 | $page = intval( $page ); |
| 72 | $rev = intval( $rev ); |
| 73 | while ( $this->lastPage < $page && !$this->atEnd ) { |
| 74 | $this->debug( "BaseDump::prefetch at page $this->lastPage, looking for $page" ); |
| 75 | $this->nextPage(); |
| 76 | } |
| 77 | if ( $this->lastPage > $page || $this->atEnd ) { |
| 78 | $this->debug( "BaseDump::prefetch already past page $page or failed to open/read input file, " |
| 79 | . "looking for rev $rev [$this->lastPage, $this->lastRev]" ); |
| 80 | |
| 81 | return null; |
| 82 | } |
| 83 | while ( $this->lastRev < $rev && !$this->atEnd && !$this->atPageEnd ) { |
| 84 | $this->debug( "BaseDump::prefetch at page $this->lastPage, rev $this->lastRev, " |
| 85 | . "looking for $page, $rev" ); |
| 86 | $this->nextRev(); |
| 87 | } |
| 88 | if ( $this->lastRev == $rev && !$this->atEnd ) { |
| 89 | $this->debug( "BaseDump::prefetch hit on $page, $rev [$this->lastPage, $this->lastRev]" ); |
| 90 | |
| 91 | if ( $slot !== SlotRecord::MAIN ) { |
| 92 | $lastSlot = SlotRecord::MAIN; |
| 93 | while ( $lastSlot !== $slot ) { |
| 94 | if ( !$this->skipTo( 'content', 'revision' ) || |
| 95 | !$this->skipTo( 'role', 'revision' ) |
| 96 | ) { |
| 97 | return null; |
| 98 | } |
| 99 | $lastSlot = $this->nodeContents(); |
| 100 | } |
| 101 | } |
| 102 | |
| 103 | return $this->nextText(); |
| 104 | } else { |
| 105 | $this->debug( "BaseDump::prefetch already past rev $rev on page $page " |
| 106 | . "[$this->lastPage, $this->lastRev]" ); |
| 107 | |
| 108 | return null; |
| 109 | } |
| 110 | } |
| 111 | |
| 112 | /** |
| 113 | * @param string $str |
| 114 | */ |
| 115 | protected function debug( $str ) { |
| 116 | wfDebug( $str ); |
| 117 | // global $dumper; |
| 118 | // $dumper->progress( $str ); |
| 119 | } |
| 120 | |
| 121 | private function nextPage() { |
| 122 | if ( $this->skipTo( 'page', 'mediawiki' ) ) { |
| 123 | if ( $this->skipTo( 'id' ) ) { |
| 124 | $this->lastPage = intval( $this->nodeContents() ); |
| 125 | $this->lastRev = 0; |
| 126 | $this->atPageEnd = false; |
| 127 | } |
| 128 | } else { |
| 129 | $this->close(); |
| 130 | if ( count( $this->infiles ) ) { |
| 131 | $infile = array_shift( $this->infiles ); |
| 132 | if ( !$this->reader->open( $infile, null, LIBXML_PARSEHUGE ) ) { |
| 133 | $this->debug( __METHOD__ . ' was unable to open xml' ); |
| 134 | $this->atEnd = true; |
| 135 | } else { |
| 136 | $this->atEnd = false; |
| 137 | } |
| 138 | } |
| 139 | } |
| 140 | } |
| 141 | |
| 142 | private function nextRev() { |
| 143 | if ( $this->skipTo( 'revision' ) ) { |
| 144 | if ( $this->skipTo( 'id' ) ) { |
| 145 | $this->lastRev = intval( $this->nodeContents() ); |
| 146 | } |
| 147 | } else { |
| 148 | $this->atPageEnd = true; |
| 149 | } |
| 150 | } |
| 151 | |
| 152 | /** |
| 153 | * @return string|null |
| 154 | */ |
| 155 | private function nextText() { |
| 156 | if ( !$this->skipTo( 'text', 'revision' ) ) { |
| 157 | return null; |
| 158 | } |
| 159 | |
| 160 | return strval( $this->nodeContents() ); |
| 161 | } |
| 162 | |
| 163 | /** |
| 164 | * @param string $name |
| 165 | * @param string $parent |
| 166 | * @return bool|null |
| 167 | */ |
| 168 | private function skipTo( $name, $parent = 'page' ) { |
| 169 | if ( $this->atEnd ) { |
| 170 | return false; |
| 171 | } |
| 172 | while ( $this->reader->read() ) { |
| 173 | if ( $this->reader->nodeType == XMLReader::ELEMENT |
| 174 | && $this->reader->name == $name |
| 175 | ) { |
| 176 | return true; |
| 177 | } |
| 178 | if ( $this->reader->nodeType == XMLReader::END_ELEMENT |
| 179 | && $this->reader->name == $parent |
| 180 | ) { |
| 181 | $this->debug( "BaseDump::skipTo found </$parent> searching for <$name>" ); |
| 182 | |
| 183 | return false; |
| 184 | } |
| 185 | } |
| 186 | |
| 187 | return $this->close(); |
| 188 | } |
| 189 | |
| 190 | /** |
| 191 | * Shouldn't something like this be built-in to XMLReader? |
| 192 | * Fetches text contents of the current element, assuming |
| 193 | * no sub-elements or such scary things. |
| 194 | * |
| 195 | * @return string|null |
| 196 | */ |
| 197 | private function nodeContents() { |
| 198 | if ( $this->atEnd ) { |
| 199 | return null; |
| 200 | } |
| 201 | if ( $this->reader->isEmptyElement ) { |
| 202 | return ""; |
| 203 | } |
| 204 | $buffer = ""; |
| 205 | while ( $this->reader->read() ) { |
| 206 | switch ( $this->reader->nodeType ) { |
| 207 | case XMLReader::TEXT: |
| 208 | // case XMLReader::WHITESPACE: |
| 209 | case XMLReader::SIGNIFICANT_WHITESPACE: |
| 210 | $buffer .= $this->reader->value; |
| 211 | break; |
| 212 | case XMLReader::END_ELEMENT: |
| 213 | return $buffer; |
| 214 | } |
| 215 | } |
| 216 | |
| 217 | return $this->close(); |
| 218 | } |
| 219 | |
| 220 | /** |
| 221 | * @return null |
| 222 | */ |
| 223 | public function close() { |
| 224 | $this->reader->close(); |
| 225 | $this->atEnd = true; |
| 226 | |
| 227 | return null; |
| 228 | } |
| 229 | } |
| 230 | |
| 231 | /** @deprecated class alias since 1.46 */ |
| 232 | class_alias( BaseDump::class, 'BaseDump' ); |