// <nowiki>
/* THIS SCRIPT IS STILL IN BETA AND IS BUGGY ABOUT 25% OF THE TIME. Be sure to check the diff that pops up before submitting.
- Adds "Run DraftCleaner" link to the left sidebar
- Top uses:
- remove extra line breaks
- in the first sentence, bold the title
- convert curly quotes to regular quotes
- put <ref>s after periods
- clean external links out of the main article area (turn them into references)
- More detailed list of uses:
- converts [inline external links] to <ref>s
- reduces more than 2 enters in a row, to 2 enters
- removes spaces in front of <ref>s
- get rid of any level 2 heading that contains the article's title
- bolds the first occurrence of the article title
- removes bold from headings
- converts =TitleHeading= to ==H2Heading==
- replaces Covid-19 with COVID-19
- removes enter characters between <ref>s
- trims whitespace at beginning and end
- remove self wikilinks to the article title
- convert ==Reference== to ==References==
- swap ref period with period ref
- turn bare URLs into references
- fix errant spaces at beginning of lines, which makes a blockquote looking thing
- add references section if missing
- delete whitespace at the end of lines
- convert smart quotes to regular quotes
- convert double spaces to single spaces
- remove blank heading
- in refs, turn short links into long links, so you can see the domain
- change year range dash to ndash
- if in draftspace, and draft in categories, disable the categories
- delete <br>. in drafts, these are usually poorly placed
- fix empty references section
- right align images
- remove whitespace if that is the only character on a line
- correct capitalization of see also, references, further reading, external links
- if article has headings but no lead, remove first heading
- replace unicode bullets with asterisks
This page was assembled from 3 files using my publish.php script. I have an offline test suite with around 100 unit tests for the DraftCleaner and StringFilter classes.
*/
$(function() {
async function getWikicode(title) {
if ( ! mw.config.get('wgCurRevisionId') ) return ''; // if page is deleted, return blank
var wikicode = '';
title = encodeURIComponent(title);
await $.ajax({
url: 'https://en.wikipedia.org/w/api.php?action=parse&page='+title+'&prop=wikitext&formatversion=2&format=json',
success: function (result) {
wikicode = result['parse']['wikitext'];
},
dataType: "json",
});
return wikicode;
}
// borrowed from [[Wikipedia:User scripts/Guide#Edit a page and other common actions]]
function editPage(articleName, wikicode, summary) {
let debugInfo = $.ajax({
url: mw.util.wikiScript('api'),
type: 'POST',
dataType: 'json',
data: {
format: 'json',
action: 'edit',
title: articleName,
text: wikicode, // will replace entire page content
summary: summary,
token: mw.user.tokens.get('csrfToken')
},
async: false
});
}
function goToShowChangesScreen(titleWithNamespaceAndUnderscores, wikicode, editSummary) {
let titleEncoded = encodeURIComponent(titleWithNamespaceAndUnderscores);
let wgServer = mw.config.get('wgServer');
let wgScriptPath = mw.config.get('wgScriptPath');
let baseURL = wgServer + wgScriptPath + '/';
let editToken = mw.user.tokens.get( 'csrfToken' );
// https://stackoverflow.com/a/12464290/3480193
$(`<form action="${baseURL}index.php?title=${titleEncoded}&action=submit" method="POST"/>`)
.append($('<input type="hidden" name="wpTextbox1">').val(wikicode))
.append($('<input type="hidden" name="wpSummary">').val(editSummary))
.append($('<input type="hidden" name="mode">').val('preview'))
.append($('<input type="hidden" name="wpDiff">').val('Show changes'))
.append($('<input type="hidden" name="wpUltimateParam">').val('1'))
.appendTo($(document.body)) //it has to be added somewhere into the <body>
.submit();
}
/** returns the pagename, including the namespace name, but with spaces replaced by underscores */
function getArticleName() {
return mw.config.get('wgPageName');
}
function showMessage(messageText) {
$('#DraftCleaner').hide();
$('#DraftCleanerNoClick').empty();
$('#DraftCleanerNoClick').prepend(messageText);
$('#DraftCleanerNoClick').show();
}
function showClickableButton() {
$('#DraftCleanerNoClick').hide();
$('#DraftCleaner').show();
}
/** refresh AND clear cache */
function hardRefresh() {
// window.location.reload(true) is deprecated. use this instead
window.location.href = window.location.href;
}
// don't run when not viewing articles
let action = mw.config.get('wgAction');
if ( action != 'view' ) return;
// don't run when viewing diffs
let isDiff = mw.config.get('wgDiffNewId');
if ( isDiff ) return;
// Only run in mainspace, draftspace, and sandboxes
let titleWithNamespaceAndUnderscores = getArticleName();
let namespaceNumber = mw.config.get('wgNamespaceNumber');
let sandbox = titleWithNamespaceAndUnderscores.match(/sandbox/i);
if ( ! [0, 118].includes(namespaceNumber) && ! sandbox ) return;
// Add DraftCleaner to left sidebar
// Using two <li>s. One of the two is kept hidden at all times. This avoids having to delete #DraftCleanerLink, which would also delete the event listener.
$('#p-navigation .vector-menu-content-list').append(`
<li id="DraftCleaner">
<a id="DraftCleanerLink">Run DraftCleaner</a>
</li>
<li id="DraftCleanerNoClick" style="display:none">
</li>
`);
$('#DraftCleanerLink').on('click', async function() {
// prevent running the script while script is already in progress
showMessage('Editing. Please wait.');
// get page wikicode
let titleWithNamespaceAndSpaces = titleWithNamespaceAndUnderscores.replace(/_/g, ' ');
let originalWikicode = await getWikicode(titleWithNamespaceAndUnderscores);
let wikicode = originalWikicode;
let dc = new DraftCleaner();
wikicode = dc.cleanDraft(wikicode, namespaceNumber, titleWithNamespaceAndSpaces);
// if changes to be made
if ( wikicode != originalWikicode ) {
let summary = 'clean up ([[User:Novem Linguae/Scripts/DraftCleaner.js|DraftCleaner]])';
// editPage(titleWithNamespaceAndUnderscores, wikicode);
// hardRefresh();
await goToShowChangesScreen(titleWithNamespaceAndUnderscores, wikicode, summary);
// else display "no changes needed", then reset
} else {
showMessage('No changes needed.');
setTimeout(function (){
showClickableButton();
}, 2000);
}
});
});
// </nowiki>
// <nowiki>
class DraftCleaner {
cleanDraft(wikicode, namespaceNumber, titleWithNamespaceAndSpaces) {
wikicode = this.fixWikilinksContainingURL(wikicode);
wikicode = this.fixExternalLinksToWikipediaArticles(wikicode);
wikicode = this.deleteWeirdUnicodeCharacters(wikicode);
wikicode = this.trimEveryLine(wikicode);
wikicode = this.convertH1ToH2(wikicode);
wikicode = this.convertVeryLongHeadingToParagraph(wikicode);
wikicode = this.bareURLToRef(wikicode);
wikicode = this.refShortLinkToLongLink(wikicode);
wikicode = this.inlineExternalLinksToRefs(wikicode); // buggy
wikicode = this.deleteSpacesInFrontOfRefs(wikicode);
wikicode = this.deleteHeadingsWithTitle(wikicode, titleWithNamespaceAndSpaces);
wikicode = this.unlinkWikilinksToThisTitle(wikicode, titleWithNamespaceAndSpaces);
wikicode = this.capitalizeCOVID19(wikicode);
wikicode = this.removeBoldFromHeadings(wikicode);
wikicode = this.deleteNewLinesBetweenRefs(wikicode);
wikicode = this.convertReferenceToReferences(wikicode);
wikicode = this.swapRefPeriodWithPeriodRef(wikicode); // buggy
wikicode = this.addReferencesSectionIfMissing(wikicode); // buggy
wikicode = this.fixEmptyReferencesSection(wikicode);
wikicode = this.deleteWhitespaceAtEndOfLines(wikicode);
wikicode = this.convertSmartQuotesToRegularQuotes(wikicode);
wikicode = this.fixWordEmphasizedWithSingleQuotes(wikicode);
wikicode = this.convertDoubleSpacesToSingleSpaces(wikicode);
wikicode = this.deleteBlankHeadings(wikicode);
wikicode = this.changeYearRangeDashToNDash(wikicode);
wikicode = this.disableCategoriesInDraftspace(wikicode, namespaceNumber);
wikicode = this.deleteBRTagsOutsideInfoboxes(wikicode, namespaceNumber); // buggy
// wikicode = this.rightAlignImages(wikicode); // commenting out, too many false positives in featured articles
wikicode = this.correctCapitalizationOfEndMatterHeaders(wikicode);
wikicode = this.ifNoLeadSectionDeleteFirstHeading(wikicode);
wikicode = this.deleteCopyPastedEditAndEditSource(wikicode);
wikicode = this.replaceUnicodeBulletsWithAsterisks(wikicode);
wikicode = this.fixDoublePeriod(wikicode);
wikicode = this.deleteEmptySections(wikicode);
wikicode = this.fixHeadingsInAllCaps(wikicode);
wikicode = this.deleteDuplicateReferencesSection(wikicode);
wikicode = this.deleteSomeHTMLTags(wikicode);
wikicode = this.deleteBlankLinesBetweenBullets(wikicode);
wikicode = this.deleteBigTag(wikicode);
wikicode = this.removeUnderscoresFromWikilinks(wikicode);
// delete empty sections. for example, empty ==See Also== section
// all ==sections== should start with a capital letter
// after swap, if citation has no spaces on either side, and is not touching two other citations, add a space on the right
// strip [[File: from infobox's image field
// example 1: | image = [[File:SAMIR 1626.png|thumb|Samir Mohanty]]
// example 2: | image = [[File:SAMIR 1626.pngy]]
// trim whitespace inside refs, e.g. <ref> abc </ref>
// single quotes to double quotes, i.e. This 'thing' to This "thing"
// replace unreliable sources with {{cn}}.
// if adjacent to other sources, just delete
// if ref is used multiple times, account for that
// duplicate citation fixer
// move refs that are below {{Reflist}}, to above {{Reflist}}
// move refs out of headings
// delete AFC submission templates located mid-article, they end up self-hiding then appear as inexplicable whitespace. example: {{AfC submission|t||ts=20211212134609|u=Doezdemir|ns=118|demo=}}<!-- Important, do not remove this line before article has been created. -->
// delete <!-- Inline citations added to your article will automatically display here. See en.wikipedia.org/wiki/WP:REFB for instructions on how to add citations. -->
// remove <big></big> tags. test: == '''''<big>Saroj Kumar Basu</big>''''' ==, '''''<big>Saroj Kumar Basu</big>'''''
// in wikilinks, underscores to spaces
// fix redundant wikilinks, e.g. [[Spotify|Spotify]]
wikicode = this.boldArticleTitle(wikicode, titleWithNamespaceAndSpaces); // buggy
wikicode = this.trimEmptyLines(wikicode);
wikicode = this.deleteMoreThanTwoEntersInARow(wikicode);
return wikicode;
}
_escapeRegEx(string) {
return string.replace(/[.*+?^${}()|[\]\\]/g, '\\$&'); // $& means the whole matched string
}
// surround bare URL's with <ref></ref>
// Useful for seeing all URL's in the reflist section, and for CiteHighlighter ref quality highlighting
bareURLToRef(wikicode) {
return wikicode.replace(/^(http[^\n ]*) {0,}$/gm, "<ref>$1</ref>");
}
// in refs, turn [short links] into long links, so you can see the domain
// also fixes <ref> link </ref> issues with extra spaces in the ref
refShortLinkToLongLink(wikicode) {
return wikicode.replace(/(<ref[^>]*>) {0,}\[ {0,}([^\]]*) {0,}\] {0,}(<\/ref>)/gm, '$1$2$3');
}
// convert inline external links to references
inlineExternalLinksToRefs(wikicode) {
let externalLinksSectionInArticle = wikicode.match(/== ?External link/i);
let furtherReadingSectionInArticle = wikicode.match(/== ?Further reading/i);
// let infoboxInArticle = wikicode.match(/{{Infobox/i);
let sf = new StringFilter();
// skip external links and further reading sections
if ( externalLinksSectionInArticle || furtherReadingSectionInArticle ) {
let topHalf = wikicode.replace(/((== ?Further reading|== ?External link).*$)/is, '');
let bottomHalf = wikicode.match(/((== ?Further reading|== ?External link).*$)/is)[1];
let buffer = sf.surgicalReplaceOutsideTags(/(?<!>|> )\[(http[^ \]]+) ?(.*?)\](?!<\/ref>| <\/ref>)/gm, '$2<ref>$1</ref>', topHalf, ['<ref'], ['</ref>', '/>']);
wikicode = buffer + bottomHalf;
} else {
wikicode = sf.surgicalReplaceOutsideTags(/(?<!>|> )\[(http[^ \]]+) ?(.*?)\](?!<\/ref>| <\/ref>)/gm, '$2<ref>$1</ref>', wikicode, ['<ref'], ['</ref>', '/>']);
}
return wikicode;
}
// get rid of spaces in front of <refs>
deleteSpacesInFrontOfRefs(wikicode) {
return wikicode.replace(/(?<!\|) {1,}<ref/gm, "<ref");
}
// get rid of any level 2 heading that contains the article's title
// this takes care of 2 common cases: heading at the bottom next to the {{AFC Submission}} template, and heading at the top above the lead
deleteHeadingsWithTitle(wikicode, titleWithNamespaceAndSpaces) {
let headingNameToLookFor = titleWithNamespaceAndSpaces;
headingNameToLookFor = headingNameToLookFor.replace(/^Draft:/, '');
headingNameToLookFor = this._escapeRegEx(headingNameToLookFor);
let regEx = new RegExp("^== ?"+headingNameToLookFor+" ?==\n", "gmi");
wikicode = wikicode.replace(regEx, "");
// now look for titles that contain Draft: at the beginning, too
headingNameToLookFor = titleWithNamespaceAndSpaces;
headingNameToLookFor = this._escapeRegEx(headingNameToLookFor);
regEx = new RegExp("^== ?"+headingNameToLookFor+" ?==\n", "gmi");
wikicode = wikicode.replace(regEx, "");
return wikicode;
}
// remove wikilinks to article name
// Example: if title is Draft:Menna Shahin, change [[Menna Shahin]] to Menna Shahin
unlinkWikilinksToThisTitle(wikicode, titleWithNamespaceAndSpaces) {
let wikilinkToLookFor = titleWithNamespaceAndSpaces;
wikilinkToLookFor = wikilinkToLookFor.replace(/^Draft:/, '');
wikilinkToLookFor = this._escapeRegEx(wikilinkToLookFor);
let regEx = new RegExp("\\[\\[("+wikilinkToLookFor+")\\]\\]", "gm");
wikicode = wikicode.replace(regEx, "$1");
return wikicode;
}
// bold the first occurrence of the article title
// WARNING: this is buggy sometimes
boldArticleTitle(wikicode, titleWithNamespaceAndSpaces) {
let boldInArticle = wikicode.includes("'''");
if ( ! boldInArticle ) {
let titleToLookFor = titleWithNamespaceAndSpaces;
titleToLookFor = titleToLookFor.replace(/^Draft:/, '');
titleToLookFor = titleToLookFor.replace(/ \(.*?\)$/, '');
titleToLookFor = this._escapeRegEx(titleToLookFor);
// Be pretty strict, to avoid adding ''' to image paths and infoboxes, which messes up the image.
let regEx = new RegExp("^(The )?("+titleToLookFor+")([ <,])", "mi");
wikicode = wikicode.replace(regEx, "$1'''$2'''$3");
}
return wikicode;
}
// /covid-19/i -> COVID-19
// Careful of this string in URLs.
capitalizeCOVID19(wikicode) {
wikicode = wikicode.replace(/ covid-19/gmi, ' COVID-19');
wikicode = wikicode.replace(/\ncovid-19/gmi, "\nCOVID-19");
return wikicode;
}
// remove bold from headings
removeBoldFromHeadings(wikicode) {
return wikicode.replace(/^(=.*)'''(.*)'''(.*=)$/gm, '$1$2$3');
}
// remove enter characters between <ref>s
deleteNewLinesBetweenRefs(wikicode) {
return wikicode.replace(/<\/ref>\n{1,}<ref>/gm, '<\/ref><ref>');
}
// convert ==Reference== to ==References==
convertReferenceToReferences(wikicode) {
return wikicode.replace(/^== ?Reference ?==$/gmi, '== References ==');
}
// TOOL - swap ref period with period ref
swapRefPeriodWithPeriodRef(wikicode) {
wikicode = wikicode.replace(/((?:<ref[^>]*?>[^>]*?<\/ref>){1,})\. /gm, '.$1 ');
wikicode = wikicode.replace(/((?:<ref[^>]*?>[^>]*?<\/ref>){1,})\.\n/gm, ".$1\n");
return wikicode;
}
// fix errant spaces at beginning of lines, which makes a blockquote looking thing (AFCH does it)
trimEveryLine(wikicode) {
let output = '';
let lines = wikicode.split("\n");
let lineCount = lines.length;
let i = 0;
for ( let line of lines ) {
i++;
let trimmed = line.trim();
if ( trimmed.startsWith('|') || trimmed.startsWith('}') ) { // don't trim lines that start with | or }. It is common in FAs to indent these a bit.
output += line;
} else {
output += line.trim();
}
if ( i !== lineCount ) {
output += "\n";
}
}
return output;
}
// add references section if missing
addReferencesSectionIfMissing(wikicode) {
let hasRefSection = wikicode.match(/^== ?References ?==$/mi);
let hasReflist = wikicode.match(/(?:{{Reflist|<references)/mi);
if ( ! hasRefSection && ! hasReflist ) {
let hasBottomAFCTemplate = wikicode.match(/(\n{{AfC submission[^}]*}}\s*)$/);
if ( hasBottomAFCTemplate ) {
wikicode = wikicode.replace(/(\n{{AfC submission[^}]*}}\s*)$/, "\n\n== References ==\n{{Reflist}}$1");
} else {
wikicode = wikicode.replace(/$/, "\n\n== References ==\n{{Reflist}}");
}
}
return wikicode;
}
// fix empty references section
fixEmptyReferencesSection(wikicode) {
let hasRefSection = wikicode.match(/^== ?References ?==$/mi);
let hasReflist = wikicode.match(/(?:{{Reflist|<references)/mi);
if ( ! hasReflist && hasRefSection ) {
wikicode = wikicode.replace(/(?<=== ?References ?==)/gmi, "\n{{Reflist}}");
}
return wikicode;
}
// delete whitespace at the end of lines
// (?!\|)(?!\}\}) is to stop this from deleting spaces after = in infoboxes
deleteWhitespaceAtEndOfLines(wikicode) {
return wikicode.replace(/[ \t]+\n(?!\|)(?!\}\})/g, "\n");
}
// convert smart quotes to regular quotes
convertSmartQuotesToRegularQuotes(wikicode) {
wikicode = wikicode.replace(/”/g, '"');
wikicode = wikicode.replace(/“/g, '"');
wikicode = wikicode.replace(/‘/g, "'");
wikicode = wikicode.replace(/’/g, "'");
wikicode = wikicode.replace(/…/g, "...");
return wikicode;
}
// convert double spaces to single spaces
convertDoubleSpacesToSingleSpaces(wikicode) {
return wikicode.replace(/\. {2,}/g, '. ');
}
// remove blank heading
deleteBlankHeadings(wikicode) {
return wikicode.replace(/\n={2,} {0,}={2,}\n/g, "\n");
}
// Change year range dash to ndash. Skip text inside of [[File:
changeYearRangeDashToNDash(wikicode) {
let sf = new StringFilter();
// (1111-1111)
wikicode = sf.surgicalReplaceOutsideTags(/(\(\d{4}) ?- ?(\d{4}\))/gm, '$1–$2', wikicode, ['[[File:'], [']]']);
// 1839 - 1926)
wikicode = sf.surgicalReplaceOutsideTags(/( \d{4}) ?- ?(\d{4}\))/gm, '$1–$2', wikicode, ['[[File:'], [']]']);
return wikicode;
}
// if in draftspace, and draft has categories, disable the categories
disableCategoriesInDraftspace(wikicode, namespace) {
let draft = (namespace == 118);
if ( draft ) {
wikicode = wikicode.replace(/:?(\[\[)(Category:[^\]]*\]\])/gm, '$1:$2');
}
return wikicode;
}
// delete <br> in drafts, these are usually poorly placed
deleteBRTagsOutsideInfoboxes(wikicode) {
let output = '';
let lines = wikicode.split("\n");
let lineCount = lines.length;
let i = 0;
for ( let line of lines ) {
i++;
// Skip lines that start with { or |. This is the easiest way to detect infoboxes
if ( line.startsWith('{') || line.startsWith('|') ) {
output += line;
} else {
output += line.replace(/\<br ?\/?\>/gm, '');
}
if ( i !== lineCount ) {
output += "\n";
}
}
return output;
}
// right align images
rightAlignImages(wikicode) {
return wikicode.replace(/(\[\[File:[^\]]*\|)left(\|[^\]]*\]\])/gm, '$1right$2');
}
// correct capitalization of see also, references, further reading, external links
correctCapitalizationOfEndMatterHeaders(wikicode) {
wikicode = wikicode.replace(/^(== ?)References( ?==)$/gmi, "$1References$2");
wikicode = wikicode.replace(/^(== ?)External links( ?==)$/gmi, "$1External links$2");
wikicode = wikicode.replace(/^(== ?)Further reading( ?==)$/gmi, "$1Further reading$2");
wikicode = wikicode.replace(/^(== ?)See also( ?==)$/gmi, "$1See also$2");
return wikicode;
}
// if article has headings but no lead, remove first heading
ifNoLeadSectionDeleteFirstHeading(wikicode) {
let output = '';
let lines = wikicode.split("\n");
let lineCount = lines.length;
let i = 0;
let textCount = 0;
for ( let line of lines ) {
i++;
// scan for first heading. empty lines, lines with templates, or lines with images do not count.
if ( line.startsWith('{') || line.length === 0 || line.startsWith('[[File:') ) {
output += line;
} else if ( line.startsWith('==') && ! textCount ) {
continue; // delete this line by not putting it in the output string
} else {
textCount++;
output += line;
}
if ( i !== lineCount ) {
output += "\n";
}
}
return output;
}
// delete [edit], [edit source], and [editar] from headings
deleteCopyPastedEditAndEditSource(wikicode) {
wikicode = wikicode.replace(/\[edit\]( ?={2,})$/gm, '$1');
wikicode = wikicode.replace(/\[edit source\]( ?={2,})$/gm, '$1');
wikicode = wikicode.replace(/\[editar\]( ?={2,})$/gm, '$1');
return wikicode;
}
// at beginning of lines, replace unicode bullets with asterisks
replaceUnicodeBulletsWithAsterisks(wikicode) {
return wikicode.replace(/^\s{0,}[·•●]\s{0,}/gm, '* ');
}
// remove whitespace if that is the only character on a line
trimEmptyLines(wikicode) {
return wikicode.replace(/^\s*$/gm, '');
}
// no more than 2 enters in a row
// Note: AFCH does this too
deleteMoreThanTwoEntersInARow(wikicode) {
return wikicode.replace(/\n{3,}/gm, "\n\n");
}
// convert =TitleHeading= to ==H2Heading==
convertH1ToH2(wikicode) {
return wikicode.replace(/^= ?([^=]*) ?=$/gm, '== $1 ==');
}
convertVeryLongHeadingToParagraph(wikicode) {
let output = '';
let lines = wikicode.split("\n");
let lineCount = lines.length;
let i = 0;
for ( let line of lines ) {
i++;
if ( line.length > 150 && line.match(/^==.*==$/gm) && ! line.match(/<ref/) ) {
output += line.replace(/^={1,}\s*(.*?)\s*={1,}$/m, '$1');
} else {
output += line;
}
if ( i !== lineCount ) {
output += "\n";
}
}
return output;
}
fixWordEmphasizedWithSingleQuotes(wikicode) {
return wikicode.replace(/ '(\w+)' /g, ' "$1" ');
}
fixDoublePeriod(wikicode) {
return wikicode.replace(/(?<=[A-Za-z\]])\.\.(?=<ref| |\n)/g, '.');
}
fixWikilinksContainingURL(wikicode) {
// non-piped wikilink
wikicode = wikicode.replace(/\[\[https?:\/\/en\.wikipedia\.org\/wiki\/([^|]*)\]\]/g, '[[$1]]');
// piped wikilink
wikicode = wikicode.replace(/\[\[https?:\/\/en\.wikipedia\.org\/wiki\/([^|]*)\|([^\]]*)\]\]/g, '[[$1|$2]]');
// non-piped external link
wikicode = wikicode.replace(/\[\[(http[^|]*)\]\]/g, '[$1]');
// piped external link
wikicode = wikicode.replace(/\[\[(http[^|]*)\|([^\]]*)\]\]/g, '[$1 $2]');
return wikicode;
}
fixExternalLinksToWikipediaArticles(wikicode) {
// [https://en.wikipedia.org/wiki/Article] and [https://en.wikipedia.org/wiki/Article Article name]
return wikicode.replace(/(?<!\[)\[https?:\/\/en\.wikipedia\.org\/wiki\/([^ \]]*)( [^\]]*)?\]/gs, '[[$1]]');
}
deleteBlankLinesBetweenBullets(wikicode) {
let lines = wikicode.split('\n');
let buffer = [];
let length = lines.length;
for ( let i = 0; i < length; i++ ) {
let previous = lines[i-1];
let current = lines[i];
let next = lines[i+1];
if (
typeof previous !== 'undefined' &&
typeof next !== 'undefined' &&
previous.startsWith('*') &&
current === '' &&
next.startsWith('*')
) {
continue;
}
buffer.push(current);
}
return buffer.join('\n');
}
deleteWeirdUnicodeCharacters(wikicode) {
return wikicode.replace(/[]/g, '');
}
deleteSomeHTMLTags(wikicode) {
wikicode = wikicode.replace(/<\/?p>/g, '');
wikicode = wikicode.replace(/<\/?strong>/g, '');
wikicode = wikicode.replace(/<\/?em>/g, '');
wikicode = wikicode.replace(/<\/?nowiki>/g, '');
return wikicode;
}
fixHeadingsInAllCaps(wikicode) {
// create a concatenated string with the text from every heading
let matches = wikicode.matchAll(/== {0,}(.+) {0,}==/g);
let headingString = '';
for ( let match of matches ) {
headingString += match[1];
}
// if string only contains caps
if ( this._isUpperCase(headingString) ) {
// convert all headings to sentence case
let matches = wikicode.matchAll(/== {0,}(.+) {0,}==/g);
for ( let match of matches ) {
let matchRegex = this._escapeRegEx(match[1].trim());
matchRegex = new RegExp('== {0,}' + matchRegex + ' {0,}==', 'g');
let sentenceCase = this._toSentenceCase(match[1].trim());
wikicode = wikicode.replace(matchRegex, '== ' + sentenceCase + ' ==');
}
}
return wikicode;
}
deleteBigTag(wikicode) {
return wikicode.replace(/(?:<big>|<\/big>)/g, '');
}
deleteEmptySections(wikicode) {
// TODO:
return wikicode;
}
deleteDuplicateReferencesSection(wikicode) {
let matches = wikicode.match(/== ?References ?==/gi);
if ( matches.length > 1 ) {
// run regexes that are likely to delete the extra section
let attempt = wikicode.replace(
`== References ==
<!-- Inline citations added to your article will automatically display here. See en.wikipedia.org/wiki/WP:REFB for instructions on how to add citations. -->
{{reflist}}`
, '');
let matches2 = attempt.match(/== ?References ?==/gi);
if ( matches2.length === 1 ) {
wikicode = attempt.trim();
}
}
return wikicode;
}
removeUnderscoresFromWikilinks(wikicode) {
let sf = new StringFilter();
wikicode = sf.surgicalReplaceInsideTags(/_/g, ' ', wikicode, ['[['], [']]']);
return wikicode;
}
_isUpperCase(str) {
return str === str.toUpperCase();
}
_toSentenceCase(string) {
return string.charAt(0).toUpperCase() + string.slice(1).toLowerCase();
}
}
/** Lets you use regex to specify what parts of a very long string you want to specify as "off limits", then you can do additional regex's and search/replace to the remaining parts of the string. */
class StringFilter {
/** Does a replace, but specifies areas of the file that should NOT be replaced. Those areas are specified by providing an openingTag and a closingTag, and those areas are marked as off limits. */
surgicalReplaceOutsideTags(regex, replacement, haystack, openingTags, closingTags) {
let allTags = [...openingTags, ...closingTags];
let parts = this._splitStringUsingMultiplePatterns(haystack, allTags);
let resultArray = [];
for ( let part of parts ) {
let openingTagMatch = false;
for ( let tag of openingTags ) {
if ( part.startsWith(tag) ) {
openingTagMatch = true;
break;
}
}
if ( ! openingTagMatch ) {
part = part.replace(regex, replacement);
}
resultArray.push(part);
}
return resultArray.join('');
}
/** Does a replace, but specifies areas of the file that SHOULD be replaced, then skips the rest of the file. The area that should be replaced is specified by providing an openingTag and a closingTag. */
surgicalReplaceInsideTags(regex, replacement, haystack, openingTags, closingTags) {
let allTags = [...openingTags, ...closingTags];
let parts = this._splitStringUsingMultiplePatterns(haystack, allTags);
let resultArray = [];
for ( let part of parts ) {
for ( let tag of openingTags ) {
if ( part.startsWith(tag) ) {
part = part.replace(regex, replacement);
}
}
resultArray.push(part);
}
return resultArray.join('');
}
/**
Also keeps the pattern in the result, unlike string.prototype.split. Algorithm isn't perfect, will fail with this pattern: <ref>Test/>Test</ref>. But should be good enough for DraftCleaner stuff.
@param {Array} patterns
*/
_splitStringUsingMultiplePatterns(string, patterns) {
let length = string.length;
let result = [];
let positionOfLastMatch = 0;
for ( let i = 0; i < length; i++ ) {
let lookAhead = string.substring(i); // the rest of the string after current position
let patternMatch = false;
for ( let pattern of patterns ) {
if ( lookAhead.startsWith(pattern) ) {
patternMatch = true;
break;
}
}
if ( patternMatch ) {
let chunk = string.slice(positionOfLastMatch, i);
if ( ! chunk ) continue; // if blank (happens if i=0 matches), continue instead of putting an empty "" into the array
result.push(chunk);
positionOfLastMatch = i;
}
}
// Don't forget the last chunk.
result.push(string.substring(positionOfLastMatch));
return result;
}
}
// </nowiki>
