Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
47.62% |
40 / 84 |
|
28.57% |
2 / 7 |
CRAP | |
0.00% |
0 / 1 |
| DOMPostProcessor | |
47.62% |
40 / 84 |
|
28.57% |
2 / 7 |
149.87 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| getTimeProfile | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| registerProcessors | |
100.00% |
8 / 8 |
|
100.00% |
1 / 1 |
4 | |||
| setSourceOffsets | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| doPostProcess | |
46.77% |
29 / 62 |
|
0.00% |
0 / 1 |
66.86 | |||
| process | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
| processChunkily | |
0.00% |
0 / 5 |
|
0.00% |
0 / 1 |
6 | |||
| 1 | <?php |
| 2 | declare( strict_types = 1 ); |
| 3 | |
| 4 | namespace Wikimedia\Parsoid\Wt2Html; |
| 5 | |
| 6 | use Generator; |
| 7 | use Wikimedia\Parsoid\Config\Env; |
| 8 | use Wikimedia\Parsoid\Core\SelectiveUpdateData; |
| 9 | use Wikimedia\Parsoid\DOM\Node; |
| 10 | use Wikimedia\Parsoid\Ext\ParsoidExtensionAPI; |
| 11 | use Wikimedia\Parsoid\Tokens\SourceRange; |
| 12 | use Wikimedia\Parsoid\Utils\ContentUtils; |
| 13 | |
| 14 | /** |
| 15 | * Perform post-processing steps on an already-built HTML DOM. |
| 16 | */ |
| 17 | class DOMPostProcessor extends PipelineStage { |
| 18 | private array $options; |
| 19 | /** @var array[] */ |
| 20 | private array $processors = []; |
| 21 | private ParsoidExtensionAPI $extApi; // Provides post-processing support to extensions |
| 22 | private string $timeProfile = ''; |
| 23 | private ?SelectiveUpdateData $selparData = null; |
| 24 | |
| 25 | public function __construct( |
| 26 | Env $env, array $options = [], string $stageId = "", |
| 27 | ?PipelineStage $prevStage = null |
| 28 | ) { |
| 29 | parent::__construct( $env, $prevStage ); |
| 30 | |
| 31 | $this->options = $options; |
| 32 | $this->extApi = new ParsoidExtensionAPI( $env ); |
| 33 | } |
| 34 | |
| 35 | public function getTimeProfile(): string { |
| 36 | return $this->timeProfile; |
| 37 | } |
| 38 | |
| 39 | public function registerProcessors( array $processors ): void { |
| 40 | foreach ( $processors as $p ) { |
| 41 | if ( isset( $p['Processor'] ) ) { |
| 42 | // Internal processor w/ ::run() method, class name given |
| 43 | $p['proc'] = new $p['Processor']( $this ); |
| 44 | } else { |
| 45 | $t = new DOMPPTraverser( $this, $p['tplInfo'] ?? false ); |
| 46 | foreach ( $p['handlers'] as $h ) { |
| 47 | $t->addHandler( $h['nodeName'], $h['action'] ); |
| 48 | } |
| 49 | $p['proc'] = $t; |
| 50 | } |
| 51 | $this->processors[] = $p; |
| 52 | } |
| 53 | } |
| 54 | |
| 55 | /* --------------------------------------------------------------------------- |
| 56 | * FIXME: |
| 57 | * 1. PipelineFactory caches pipelines per env |
| 58 | * 2. PipelineFactory.parse uses a default cache key |
| 59 | * 3. ParserTests uses a shared/global env object for all tests. |
| 60 | * 4. ParserTests also uses PipelineFactory.parse (via env.getContentHandler()) |
| 61 | * => the pipeline constructed for the first test that runs wt2html |
| 62 | * is used for all subsequent wt2html tests |
| 63 | * 5. If we are selectively turning on/off options on a per-test basis |
| 64 | * in parser tests, those options won't work if those options are |
| 65 | * also used to configure pipeline construction (including which DOM passes |
| 66 | * are enabled). |
| 67 | * |
| 68 | * Ex: if (env.wrapSections) { addPP('wrapSections', wrapSections); } |
| 69 | * |
| 70 | * This won't do what you expect it to do. This is primarily a |
| 71 | * parser tests script issue -- but given the abstraction layers that |
| 72 | * are on top of the parser pipeline construction, fixing that is |
| 73 | * not straightforward right now. So, this note is a warning to future |
| 74 | * developers to pay attention to how they construct pipelines. |
| 75 | * --------------------------------------------------------------------------- */ |
| 76 | |
| 77 | /** |
| 78 | * @inheritDoc |
| 79 | */ |
| 80 | public function setSourceOffsets( SourceRange $so ): void { |
| 81 | $this->options['sourceOffsets'] = $so; |
| 82 | } |
| 83 | |
| 84 | public function doPostProcess( Node $node ): void { |
| 85 | $env = $this->env; |
| 86 | |
| 87 | $hasDumpFlags = $env->hasDumpFlags(); |
| 88 | |
| 89 | // FIXME: This works right now, but may not always be the right place to dump |
| 90 | // if custom DOM pipelines start getting more specialized and we enter this |
| 91 | // pipeline immediate after tree building. |
| 92 | if ( $hasDumpFlags && $env->hasDumpFlag( 'dom:post-builder' ) ) { |
| 93 | $opts = []; |
| 94 | $env->writeDump( ContentUtils::dumpDOM( $node, 'DOM: after tree builder', $opts ) ); |
| 95 | } |
| 96 | |
| 97 | $prefix = null; |
| 98 | $traceLevel = null; |
| 99 | $resourceCategory = null; |
| 100 | |
| 101 | $profile = null; |
| 102 | if ( $env->profiling() ) { |
| 103 | $profile = $env->getCurrentProfile(); |
| 104 | if ( $this->atTopLevel ) { |
| 105 | $this->timeProfile = str_repeat( "-", 85 ) . "\n"; |
| 106 | $prefix = 'TOP'; |
| 107 | // Turn off DOM pass timing tracing on non-top-level documents |
| 108 | $resourceCategory = 'DOMPasses:TOP'; |
| 109 | } else { |
| 110 | $prefix = '---'; |
| 111 | $resourceCategory = 'DOMPasses:NESTED'; |
| 112 | } |
| 113 | } |
| 114 | |
| 115 | foreach ( $this->processors as $pp ) { |
| 116 | // This is an optimization for the 'AddAnnotationIds' handler |
| 117 | // which is embedded in a DOMTraverser where we cannot check this flag. |
| 118 | if ( !empty( $pp['withAnnotations'] ) && !$this->env->hasAnnotations ) { |
| 119 | continue; |
| 120 | } |
| 121 | |
| 122 | $ppName = null; |
| 123 | $ppStart = null; |
| 124 | |
| 125 | // Trace |
| 126 | if ( $profile ) { |
| 127 | $ppName = $pp['name'] . str_repeat( |
| 128 | " ", |
| 129 | ( strlen( $pp['name'] ) < 30 ) ? 30 - strlen( $pp['name'] ) : 0 |
| 130 | ); |
| 131 | $ppStart = microtime( true ); |
| 132 | } |
| 133 | |
| 134 | $opts = null; |
| 135 | if ( $hasDumpFlags ) { |
| 136 | $opts = [ |
| 137 | 'env' => $env, |
| 138 | 'dumpFragmentMap' => $this->atTopLevel, |
| 139 | 'keepTmp' => true |
| 140 | ]; |
| 141 | |
| 142 | if ( $env->hasDumpFlag( 'dom:pre-' . $pp['shortcut'] ) |
| 143 | || $env->hasDumpFlag( 'dom:pre-*' ) |
| 144 | ) { |
| 145 | $env->writeDump( |
| 146 | ContentUtils::dumpDOM( $node, 'DOM: pre-' . $pp['shortcut'], $opts ) |
| 147 | ); |
| 148 | } |
| 149 | } |
| 150 | |
| 151 | // FIXME: env, extApi, frame, selparData, options, atTopLevel can all be |
| 152 | // put into a stdclass or a real class (DOMProcConfig?) and passed around. |
| 153 | $pp['proc']->run( |
| 154 | $this->env, |
| 155 | $node, |
| 156 | [ |
| 157 | 'extApi' => $this->extApi, |
| 158 | 'frame' => $this->frame, |
| 159 | 'selparData' => $this->selparData, |
| 160 | ] + $this->options, |
| 161 | $this->atTopLevel |
| 162 | ); |
| 163 | |
| 164 | if ( $hasDumpFlags && ( $env->hasDumpFlag( 'dom:post-' . $pp['shortcut'] ) |
| 165 | || $env->hasDumpFlag( 'dom:post-*' ) ) |
| 166 | ) { |
| 167 | $env->writeDump( |
| 168 | ContentUtils::dumpDOM( $node, 'DOM: post-' . $pp['shortcut'], $opts ) |
| 169 | ); |
| 170 | } |
| 171 | |
| 172 | if ( $profile ) { |
| 173 | $ppElapsed = 1000 * ( microtime( true ) - $ppStart ); |
| 174 | if ( $this->atTopLevel ) { |
| 175 | $this->timeProfile .= str_pad( $prefix . '; ' . $ppName, 65 ) . |
| 176 | ' time = ' . |
| 177 | str_pad( number_format( $ppElapsed, 2 ), 10, ' ', STR_PAD_LEFT ) . "\n"; |
| 178 | } |
| 179 | $profile->bumpTimeUse( $resourceCategory, $ppElapsed, 'DOM' ); |
| 180 | } |
| 181 | } |
| 182 | } |
| 183 | |
| 184 | /** |
| 185 | * @inheritDoc |
| 186 | */ |
| 187 | public function process( $node, array $opts ) { |
| 188 | if ( isset( $opts['selparData'] ) ) { |
| 189 | $this->selparData = $opts['selparData']; |
| 190 | } |
| 191 | '@phan-var Node $node'; // @var Node $node |
| 192 | $this->doPostProcess( $node ); |
| 193 | // @phan-suppress-next-line PhanTypeMismatchReturnSuperType |
| 194 | return $node; |
| 195 | } |
| 196 | |
| 197 | /** |
| 198 | * @inheritDoc |
| 199 | */ |
| 200 | public function processChunkily( $input, array $options ): Generator { |
| 201 | if ( $this->prevStage ) { |
| 202 | // The previous stage will yield a DOM. |
| 203 | // FIXME: Should we change the signature of that to return a DOM |
| 204 | // If we do so, a pipeline stage returns either a generator or |
| 205 | // concrete output (in this case, a DOM). |
| 206 | $node = $this->prevStage->processChunkily( $input, $options )->current(); |
| 207 | } else { |
| 208 | $node = $input; |
| 209 | } |
| 210 | $this->process( $node, $options ); |
| 211 | yield $node; |
| 212 | } |
| 213 | } |