User:Pathoschild/standardise.js
Jump to navigation
Jump to search
Note: After saving, changes may not occur immediately. Click here to learn how to bypass your browser's cache.
- Firefox / Safari: Hold Shift while clicking Reload, or press either Ctrl-F5 or Ctrl-R (Cmd-R on a Mac)
- Google Chrome: Press Ctrl-Shift-R (Cmd-Shift-R on a Mac)
- Internet Explorer: Hold Ctrl while clicking Refresh, or press Ctrl-F5
- Opera: Clear the cache in Tools → Preferences
For details and instructions about other browsers, see Wikipedia:Bypass your cache.
| Code that you insert on this page could contain malicious content capable of compromising your account. If you are unsure whether code you are adding to this page is safe, you can ask at the central discussion page, Scriptorium. The code will be executed when previewing this page under some skins, including Monobook. You can in the interim if you wish to refresh the content sooner under another skin. |
 | Documentation for this script can be added at User:Pathoschild/standardise. |
/* global $, mw, pathoschild */
// <syntaxhighlight lang="javascript">
/*************
*** Wikisource standardization extension (0.6)
*** for [[m:TemplateScript]]
*** by [[m:user:Pathoschild]]
*************/
/**
* TemplateScript adds configurable templates and scripts to the sidebar, and adds an example regex editor.
* @see https://meta.wikimedia.org/wiki/TemplateScript
* @update-token [[File:pathoschild/templatescript.js]]
*/
mw.loader.load('//tools-static.wmflabs.org/meta/scripts/pathoschild.templatescript.js');
function standardize(editor) {
editor = editor || pathoschild.TemplateScript.Context; // backwards compatibility for regex menu framework
// hide exceptions
editor.escape(/<(nowiki|poem|pre|source)[^>]*>[\s\S]*?<\/\1>/ig);
// main namespace
if(mw.config.get('wgNamespaceNumber') === 0) {
/*******************
*** header normalization
*******************/
/* prepare template for parsing if present */
if(editor.get().match(/{{\s*(?:msg:|template:)?\s*header/i)) {
editor
/* header syntax */
.replace(/{{\s*(?:msg:|template:)\s*header/i,'{{header')
/* fix delimiters */
// header parameters
.replace(/({{header2?[^\n]*)[\n\s]*\|[\s\n]*(previous|next|title|section|author|notes)\s*=\s*/ig,'$1~$2=',6)
// nested templates (up to 5)
.replace(/({{header[\s\S]*?{{[^\|}]*)\|/ig,'$1%%pipe%%',5)
.replace(/({{header[\s\S]*?){{([^}]+)}}/ig,'$1%%leftcurlies%%$2%%rightcurlies%%',5)
/* cleanup header */
// deprecated arrows and brackets
.replace(/((?:previous|next)[^~]*?)(?:&larr|→|←|→)/ig,'$1')
.replace(/(section\s*=\s*)\(([^~}]*)\)/,'$1$2')
// trailing whitespace
.replace(/{{(header[^}]+)\s*}}/i,'{{$1}}')
// non-semantic line-breaks
.replace(/({{header[^}]+(?:title|section|author|section)=[^~]*)<[^>]*br[^>]*>/ig,'$1',8);
}
/* prepare values */
var headerTemplate = String(editor.get().match(/{{header[\s\S]+?}}/i));
/* parse */
// title
var pattern = /^[\s\S]*~title=([^~}]*)[\s\S]*$/;
var headerTitle = headerTemplate.match(pattern)
? headerTemplate.replace(pattern,'$1')
: mw.config.get('wgTitle').match(/^[^\/]+/);
// author
var pattern = /^[\s\S]*~author=([^~}]*)[\s\S]*$/;
var headerAuthor = headerTemplate.match(pattern)
? headerTemplate.replace(pattern,'$1')
: '';
// section
var pattern = /^[\s\S]*~section=([^~}]*)[\s\S]*$/;
if(headerTemplate.match(pattern))
var headerSection = headerTemplate.replace(pattern,'$1');
else {
if(mw.config.get('wgTitle').match(/.+\//)) {
var newTitle = mw.config.get('wgTitle').replace(/^.*\/([^\/]+)/);
var headerSection = mw.config.get('wgTitle').replace(/^.*\/([^\/]+)/,'$1');
}
else
var headerSection = '';
}
// previous
var pattern = /^[\s\S]*~previous=([^~}]*)[\s\S]*$/;
if(headerTemplate.match(pattern))
var headerPrevious = headerTemplate.replace(pattern,'$1');
else {
if(mw.config.get('wgTitle').match(/\/Chapter \d+$/)) {
var newTitle = parseInt(mw.config.get('wgTitle').replace(/^.*\/Chapter (\d+)$/,'$1'))-1;
if(newTitle > 0)
var headerPrevious = '[[../Chapter '+newTitle+'|Chapter '+newTitle+']]';
else
var headerPrevious = '';
}
else
var headerPrevious = '';
}
// next
var pattern = /^[\s\S]*~next=([^~}]*)[\s\S]*$/;
if(headerTemplate.match(pattern))
var headerNext = headerTemplate.replace(pattern,'$1');
else {
if(mw.config.get('wgTitle').match(/\/Chapter \d+$/)) {
var newTitle = parseInt(mw.config.get('wgTitle').replace(/^.*\/Chapter (\d+)$/,'$1'))+1;
var headerNext = '[[../Chapter '+newTitle+'|Chapter '+newTitle+']]';
}
else
var headerNext = '';
}
// notes
var pattern = /^[\s\S]*~notes=([^~}]*)[\s\S]*$/;
if(headerTemplate.match(pattern)) {
var headerNotes = headerTemplate.replace(pattern,'$1');
headerNotes = headerNotes.replace(/^([\s\S]*?)[\n\s]*$/,'$1'); // trailing whitespace
}
else
var headerNotes = '';
/* replace template */
editor
.replace(/{{header[^}]*}}\n*/ig, '')
.set(
'{{header2'
+ '\n | title = ' + headerTitle
+ '\n | author = ' + headerAuthor
+ '\n | section = ' + headerSection
+ '\n | previous = ' + headerPrevious
+ '\n | next = ' + headerNext
+ '\n | notes = ' + headerNotes
+ '\n}}\n'
+ editor.get()
);
/*******************
*** fix false newlines
*******************/
editor
/* replace newlines with placeholders */
.replace(/([^\n])\n(\s*[^\n])/ig,'$1%%newline%%$2')
/* selectively restore legitimate newlines */
// paragraphs
.replace(/%%newline%%%%newline%%|%%newline%%\n|\n%%newline%%/ig,'\n\n')
// templates
.replace(/%%newline%%(\s*(?:}}|\|))/ig,'\n$1') // before
.replace(/}}\s*%%newline%%/ig,'}}\n') // after
// images, categories, interwiki links
.replace(/%%newline%%(\s*\[\[(?:Image|Category|[^:]+):[^\]]+\]\])/ig,'\n$1')
.replace(/(\s*\[\[(?:Image|Category|[^:]+):[^\]]+\]\])%%newline%%/ig,'$1\n')
// lists
.replace(/%%newline%%([*#:;])/ig,'\n$1') // lists
.replace(/([*#:;][^\n]*)%%newline%%/ig,'$1\n') // newlines closing list items
// tables
.replace(/%%newline%%{\|/ig,'\n{|')
.replace(/{\|%%newline%%/ig,'{|\n')
// rules
.replace(/%%newline%%(----+)/g,'\n$1')
.replace(/(----+)%%newline%%/g,'$1\n')
// tags
.replace(/(<[^>\n]+>)\s*%%newline%%/ig,'$1\n')
.replace(/%%newline%%(<[^>\n]+>)/ig,'\n$1')
/* remove remaining */
.replace(/-%%newline%%([^\s])/ig,'-$1') // hyphenated words
.replace(/\s*%%newline%%\s*/ig,' '); // all others
/*******************
*** Cleanup
*******************/
editor
/* restore delimiters */
.replace(/%%pipe%%/g,'|')
.replace(/%%leftcurlies%%/g,'{{')
.replace(/%%rightcurlies%%/g,'}}');
}
/*******************
*** Authorspace
*******************/
if(mw.config.get('wgNamespaceNumber') === 102) {
/*******************
*** {{author}} normalization
*******************/
editor
/* fix delimiters */
.replace(/[\n\s]*\|\s*((?:first|last)name|last_initial|(?:birth|death)year|description|image|(?:wikipedia|wikiquote|commons)_link|dates|name|defaultsort)\s*=\s*/ig,'~$1=') // author parameter delimiters
.replace(/({{author[\s\S]*?{{[^\|}]*)\|/ig,'$1%%pipe%%',5) // other template pipes
.replace(/({{author[\s\S]*?){{([^}]+)}}/ig,'$1%%leftcurlies%%$2%%rightcurlies%%',5) // other template delimiters
/* cleanup */
.replace(/{{(author[^}]+)\s*}}/i,'{{$1}}') // rm trailing whitespace
/* place standard template and move like parameters */
.replace(/{{author/i,'{{author\n |firstname =\n |lastname =\n |last_initial =\n |birthyear =\n |deathyear =\n |description =\n |image =\n |wikipedia_link =\n |wikiquote_link =\n |commons_link =\n}}\n{{author')
.replace(/(author[\s\S]*?\|((?:first|last)?name|last_initial|(?:birth|death)year|description|image|(?:wikipedia|wikiquote|commons)_link|dates|defaultsort)\s*)=([\s\S]*?)~?\2=([^~]*)/i,'$1=$4$3',10);
/* get dates if necessary */
if(editor.get().match(/(?:birth|death)year\s*=\s*\n/)) {
// cannibalise categories
editor
.replace(/(birthyear\s*)=(\s*\n[\s\S]*?)\n?\[\[\s*Category\s*:\s*(\d+(?:\s*BCE)?) births\s*[^\]]*\]\]/,'$1=$3$2')
.replace(/(deathyear\s*)=(\s*\n[\s\S]*?)\n?\[\[\s*Category\s*:\s*(\d+(?:\s*BCE)?) deaths\s*[^\]]*\]\]/,'$1=$3$2');
// if that failed, parse from old template
if(editor.get().match(/(?:birth|death)year\s*=\s*\n/) && editor.get().match(/~dates=[^~]/)) {
/* get dates */
// get raw parameter
var oldDates = editor.get().replace(/^[\s\S]*dates=[^\d~}]*([^~}]+)[\s\S]*$/,'$1'); // raw parameter
oldDates = oldDates.replace(/^(\d+)\s*BC?E/ig,'$1 BCE'); // fix eras
// extract dates
var birthYear = oldDates.replace(/^(\d+(?: BCE)?)[\s\S]*$/ig,'$1');
var deathYear = oldDates.replace(/^\d+[^\d]+?(\d+(?: BCE)?)$/ig,'$1');
/* fill in empty parameters */
if(editor.get().match(/birthyear\s*=\s*\n/))
editor.replace(/(birthyear\s*)=/,'$1='+birthYear);
if(editor.get().match(/deathyear\s*=\s*\n/) && deathYear > birthYear)
editor.replace(/(deathyear\s*)=/,'$1='+deathYear);
}
}
/* get names */
if(editor.replace(/(?:first|last)name\s*=\s*\n/)) {
// cannibalise name field
if(editor.replace(/(?:first|last)name\s*=\s*\n/)) {
// extract
var name = editor.get().replace(/^[\s\S]*~name=([^~}]*)[\s\S]*/,'$1');
var firstName = name.replace(/([\s\S]+)\s+[\s\S]*/,'$1');
var lastName = name.replace(/[\s\S]+\s+([\s\S]*)/,'$1');
// fill in empty parameters
if(editor.get().match(/firstname\s*=\s*\n/))
editor.replace(/(firstname\s*)=/,'$1='+firstName);
if(editor.get().match(/lastname\s*=\s*\n/))
editor.replace(/(lastname\s*)=/,'$1='+lastName);
}
}
/* cleanup */
editor
// remove old template
.replace(/({{author[\s\S]*?)[\n\s]*{{author[^}]*}}[\n\s]*/ig,'$1\n\n')
// restore delimiters
.replace(/%%pipe%%/g,'|')
.replace(/%%leftcurlies%%/g,'{{')
.replace(/%%rightcurlies%%/g,'}}')
// fix whitespace
.replace(/((?:(?:first|last)name|last_initial|(?:birth|death)year|description|image|(?:wikipedia|wikiquote|commons)_link)\s*)=\s*/ig,'$1= ')
.replace(/= \|/g,'= \n |')
.replace(/= }}/g,'= \n}}')
/* remove old categories */
.replace(/\[\[\s*Category\s*:\s*\d+[^\]]*?(?:births|deaths)[^\]]*\]\]\n?/ig,'') // authors by year
.replace(/\[\[\s*Category\s*:\s*(?:Ancient|Early modern|Medieval|Modern|Renaissance) authors[^\]]*\]\]\n?/ig,''); // authors by era
/*******************
*** Other tweaks
*******************/
/* update license templates */
editor.replace(/{{\s*(?:msg:|template:)?(?:author-)?(PD-[^\|\}]+)(?:\|[^}]*)?}}/ig,'{{$1}}');
/* normalize dates */
editor.replace(/^([#*:]+ \[\[[^\]]+\]\]),\s*(\d+)/mig,'$1 ($2)');
}
/*******************
*** miscellaneous cleanup
*******************/
editor
/* templates */
.replace(/{{\s*(?:msg:|template:)?([^}]+)}}/ig,'{{$1}}')
/* syntax */
// headers
.replace(/\n*^(=+)\s*(.*?)\s*\1\s*/mig,'\n\n$1$2$1\n') // whitespace
.replace(/=\n+=/ig,'=\n=') // fix consecutive headers
// categories
.replace(/\[\[\s*category\s*:\s*([^\|\]]+)(?:\s*(\|)([^\]]*))?\s*\]\]/ig,'[[' + 'category:$1$2$3]]')
//links
.replace(/\[\[\s*([^\|\]]+?)\s*(?:(\|)\s*([^\]]+?)\s*)?\]\]/ig,'[[$1$2$3]]') // redundant starting and ending whitespace
.replace(/\[\[([^\|\]]+?)\s*\|\s*\1\]\]/ig,'[[$1]]') // redundant link text
.replace(/\[\[([^\|\]]+?)_/ig,'[[$1 ',5) // underscores
// lists
.replace(/^([*#:]+)\s*/mig,'$1 ');
/*******************
*** sort elements
*******************/
/* store elements and remove from code */
// categories
var categories = editor.get().match(/\[\[category:[^\]]+\]\]/ig);
editor.replace(/\[\[category:[^\]]+\]\]\n?/ig,'');
// interlanguage links
var interwikilinks = editor.get().match(/\[\[[a-z]{2,3}(?:-[^:\|\]]+)?:[^\]]+\]\]/ig); // get codes
if(interwikilinks) {
for(var x in interwikilinks) { // filter out known non-interlanguage prefixes
if(interwikilinks[x].match(/\[\[(?:c2|cej|dcc|mw|rev|rfc|svn|wqy):/i)) {
interwikilinks.splice(x,0);
}
else
editor.replace(/\[\[[a-z]{2,3}(?:-[^:\|\]]+)?:[^\]]+\]\]\n?/ig,'');
}
}
// license templates
var licenses = editor.get().match(/{{(?:PD-|GFDL)[^}]*}}/ig);
editor.replace(/{{(?:PD-|GFDL)[^}]*}}\n?/ig,'');
/* sort and re-add */
// compare function for case-insensitivity
// courtesy <http://www.webreference.com/js/tips/000430.html>
function caseless(a, b) {
a = a.toLowerCase();
b = b.toLowerCase();
if (a < b) return -1;
if (a > b) return 1;
return 0;
}
// initial whitespace
editor.replace(/[\s\n]*$/,'\n\n');
// licenses
if(licenses && licenses.length > 0) {
// licenses
for(var x in licenses)
editor.set(editor.get() + licenses[x] + '\n');
// whitespace
editor.set(editor.get() + '\n');
}
// categories
if(categories && categories.length > 0) {
// sort and place
categories.sort(caseless);
for(var x in categories)
editor.set(editor.get() + categories[x] + '\n');
// whitespace
editor.set(editor.get() + '\n');
}
// interlanguage links
if(interwikilinks) {
// sort and place
interwikilinks.sort(caseless);
for(var x in interwikilinks)
editor.set(editor.get() + interwikilinks[x] + '\n');
}
/* restore exceptions */
editor.unescape();
/* edit summary */
editor.setEditSummary('[[WS:STYLE|standardization]], updates, and cleanup with [[m:TemplateScript|regex]]');
}
//</syntaxhighlight>