wikimedia/mediawiki-extensions-VisualEditor
1
0
Fork 0
mirror of https://gerrit.wikimedia.org/r/mediawiki/extensions/VisualEditor synced 2024-09-25 11:16:51 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge "Added html2wt command-line option to parse.js"

Browse source
This commit is contained in:
GWicke 2012-07-24 18:26:50 +00:00 committed by Gerrit Code Review
parent 01d6f17a3e ddc6899b1b
commit dfe082a258
1 changed files with 48 additions and 31 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

71
modules/parser/parse.js
View file

@ -11,7 +11,8 @@ var ParserPipelineFactory = require('./mediawiki.parser.js').ParserPipelineFacto
ConvertDOMToLM = require('./mediawiki.LinearModelConverter.js').ConvertDOMToLM,
DOMConverter = require('./mediawiki.DOMConverter.js').DOMConverter,
WikitextSerializer = require('./mediawiki.WikitextSerializer.js').WikitextSerializer,
optimist = require('optimist');
optimist = require('optimist'),
html5 = require('html5');
( function() {
var opts = optimist.usage( 'Usage: echo wikitext | $0', {
@ -30,6 +31,11 @@ var ParserPipelineFactory = require('./mediawiki.parser.js').ParserPipelineFacto
'boolean': true,
'default': false
},
'html2wt': {
description: 'Convert input HTML to Wikitext',
'boolean': true,
'default': false
},
'wikitext': {
description: 'Output WikiText instead of HTML',
'boolean': true,
@ -97,8 +103,6 @@ var ParserPipelineFactory = require('./mediawiki.parser.js').ParserPipelineFacto
maxDepth: argv.maxdepth,
pageName: argv.pagename
} );
var parserPipelineFactory = new ParserPipelineFactory( env );
var parser = parserPipelineFactory.makePipeline( 'text/x-mediawiki/full' );
process.stdin.resume();
process.stdin.setEncoding('utf8');
@ -108,34 +112,47 @@ var ParserPipelineFactory = require('./mediawiki.parser.js').ParserPipelineFacto
inputChunks.push( chunk );
} );
process.stdin.on( 'end', function() {
var input = inputChunks.join('');
parser.on('document', function ( document ) {
// Print out the html
if ( argv.linearmodel ) {
process.stdout.write(
JSON.stringify( ConvertDOMToLM( document.body ), null, 2 ) );
} else if ( argv.wikidom ) {
process.stdout.write(
JSON.stringify(
new DOMConverter().HTMLtoWiki( document.body ),
null,
2
));
} else if ( argv.wikitext ) {
new WikitextSerializer({env: env}).serializeDOM( document.body,
process.stdout.write.bind( process.stdout ) );
} else {
process.stdout.write( document.body.innerHTML );
}
if (argv.html2wt) {
var p = new html5.Parser();
p.parse('<html><body>' + input.replace(/\r/g, '') + '</body></html>');
var content = p.tree.document.childNodes[0].childNodes[1];
var stdout = process.stdout;
new WikitextSerializer({env: env}).serializeDOM(content, stdout.write.bind(stdout));
// add a trailing newline for shell user's benefit
process.stdout.write( "\n" );
stdout.write( "\n" );
process.exit(0);
});
// Kick off the pipeline by feeding the input into the parser pipeline
parser.process( input );
} else {
var parserPipelineFactory = new ParserPipelineFactory( env );
var parser = parserPipelineFactory.makePipeline( 'text/x-mediawiki/full' );
parser.on('document', function ( document ) {
// Print out the html
if ( argv.linearmodel ) {
process.stdout.write(
JSON.stringify( ConvertDOMToLM( document.body ), null, 2 ) );
} else if ( argv.wikidom ) {
process.stdout.write(
JSON.stringify(
new DOMConverter().HTMLtoWiki( document.body ),
null,
2
));
} else if ( argv.wikitext ) {
new WikitextSerializer({env: env}).serializeDOM( document.body,
process.stdout.write.bind( process.stdout ) );
} else {
process.stdout.write( document.body.innerHTML );
}
// add a trailing newline for shell user's benefit
process.stdout.write( "\n" );
process.exit(0);
});
// Kick off the pipeline by feeding the input into the parser pipeline
parser.process( input );
}
} );
} )();