Arlolra has uploaded a new change for review.
https://gerrit.wikimedia.org/r/91321
Change subject: Public API for page-less html2wt / wt2html
......................................................................
Public API for page-less html2wt / wt2html
Cleans up the existing POST api and adds two new paths, _wt2html and
_html2wt.
Bug: 55758
Change-Id: I7ce70b98be802ba0894826629b1a21a083a4fde3
---
M js/api/ParserService.js
1 file changed, 55 insertions(+), 37 deletions(-)
git pull ssh://gerrit.wikimedia.org:29418/mediawiki/extensions/Parsoid
refs/changes/21/91321/1
diff --git a/js/api/ParserService.js b/js/api/ParserService.js
index 6a1cfbe..281c974 100644
--- a/js/api/ParserService.js
+++ b/js/api/ParserService.js
@@ -440,28 +440,41 @@
getParserServiceEnv( res, parsoidConfig.defaultWiki, req.params[0], cb,
req );
} );
-app.post(/\/_html\/(.*)/, function ( req, res ) {
+app.post( /\/_html\/(.*)/, html2wt );
+app.post( /\/_html2wt\/(.*)/, html2wt );
+
+function html2wt( req, res ) {
var cb = function ( env ) {
- res.setHeader('Content-Type', 'text/html; charset=UTF-8');
- var doc = DU.parseHTML(req.body.content.replace(/\r/g, ''));
- res.write('<pre style="background-color: #efefef">');
+ var doc;
+ try {
+ doc = DU.parseHTML( req.body.content.replace( /\r/g, ''
) );
+ } catch ( e ) {
+ console.log( 'There was an error in the HTML5 parser.'
);
+ env.errCB( e );
+ return;
+ }
+
// Always use the non-selective serializer for this mode
- new WikitextSerializer({env: env}).serializeDOM(
- doc.body,
- function( c ) {
- res.write( htmlSpecialChars( c ) );
- },
- function() {
- res.write('</pre>');
- res.write( "<hr>Your HTML DOM:" );
- textarea( res, req.body.content.replace(/\r/g,
'') );
- res.end('');
- }
+ var out = [];
+ try {
+ new WikitextSerializer({ env: env }).serializeDOM(
+ doc.body,
+ function( chunk ) {
+ out.push( chunk );
+ },
+ function() {
+ res.setHeader( 'Content-Type',
'text/x-mediawiki; charset=UTF-8' );
+ res.setHeader( 'X-Parsoid-Performance',
env.getPerformanceHeader() );
+ res.end( out.join('') );
+ }
);
+ } catch ( e ) {
+ env.errCB( e );
+ }
};
getParserServiceEnv( res, parsoidConfig.defaultWiki, req.params[0], cb,
req );
-} );
+}
// Form-based wikitext -> HTML DOM interface for manual testing
app.get(/\/_wikitext\/(.*)/, function ( req, res ) {
@@ -475,33 +488,38 @@
getParserServiceEnv( res, parsoidConfig.defaultWiki, req.params[0], cb,
req );
} );
-app.post(/\/_wikitext\/(.*)/, function ( req, res ) {
+app.post( /\/_wikitext\/(.*)/, wt2html );
+app.post( /\/_wt2html\/(.*)/, wt2html );
+
+function wt2html( req, res ) {
var cb = function ( env ) {
- res.setHeader('Content-Type', 'text/html; charset=UTF-8');
- var parser = Util.getParserPipeline(env,
'text/x-mediawiki/full'),
- src = req.body.content.replace(/\r/g, '');
- parser.on('document', function ( document ) {
- if (req.body.format==='html') {
- res.write(DU.serializeNode(document));
- } else {
- res.write('<pre style="white-space: pre-wrap;
white-space: -moz-pre-wrap; white-space: -pre-wrap; white-space: -o-pre-wrap;
word-wrap: break-word;">');
-
res.write(htmlSpecialChars(DU.serializeNode(document.body)));
- res.write('</pre>');
- res.write('<hr/>');
- res.write(document.body.innerHTML);
- res.write('<hr style="clear:both;"/>Your
wikitext:');
- textarea( res, src );
- }
- res.end('');
- });
- if (env.conf.parsoid.allowCORS) {
+ var src = req.body.content.replace( /\r/g, '' );
+ console.log( 'starting parsing of ' + src );
+
+ if ( env.conf.parsoid.allowCORS ) {
// allow cross-domain requests (CORS) so that parsoid
service
// can be used by third-party sites
res.setHeader('Access-Control-Allow-Origin',
env.conf.parsoid.allowCORS);
}
+
+ var parser = Util.getParserPipeline( env,
'text/x-mediawiki/full' );
+ parser.on('document', function ( document ) {
+ var out;
+ if ( req.body.format === 'html' ) {
+ out = DU.serializeNode( document );
+ } else {
+ out = DU.serializeNode( document.body );
+ }
+
+ res.setHeader( 'Content-Type', 'text/html;
charset=UTF-8' );
+ res.setHeader( 'X-Parsoid-Performance',
env.getPerformanceHeader() );
+ res.end( out );
+
+ console.warn( "completed parsing of " + src + " in " +
env.performance.duration + " ms" );
+ });
+
try {
- console.log('starting parsing of ' + req.params[0]);
// FIXME: This does not handle includes or templates
correctly
env.setPageSrcInfo( src );
parser.processToplevelDoc( src );
@@ -513,7 +531,7 @@
};
getParserServiceEnv( res, parsoidConfig.defaultWiki, req.params[0], cb,
req );
-} );
+}
// Round-trip article testing
app.get( new RegExp('/_rt/(' + getInterwikiRE() + ')/(.*)'), function(req,
res) {
--
To view, visit https://gerrit.wikimedia.org/r/91321
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: newchange
Gerrit-Change-Id: I7ce70b98be802ba0894826629b1a21a083a4fde3
Gerrit-PatchSet: 1
Gerrit-Project: mediawiki/extensions/Parsoid
Gerrit-Branch: master
Gerrit-Owner: Arlolra <abrea...@wikimedia.org>
_______________________________________________
MediaWiki-commits mailing list
MediaWiki-commits@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/mediawiki-commits
