Wikipedia

User:Polygnotus/DuplicateReferences.js: Difference between revisions

Browse history interactively
← Previous edit
Content deleted Content added
No edit summary
No edit summary
 
(42 intermediate revisions by the same user not shown)
Line 1:
//Testpage: https://en.wikipedia.org/wiki/User:Polygnotus/DuplicateReferencesTest
 
// <nowiki>
mw.loader.using(['mediawiki.util'], function () {
$(document).ready(function () {
const DEBUG = false;
function debug(...args) {
if (DEBUG) {
console.log(...args);
}
}
 
debug("Scriptconst started")DEBUG = false;
 
function debug(...args) {
if ((mw.config.get('wgNamespaceNumber') !== 0 && mw.config.get('wgPageName') !== 'User:Polygnotus/dupreftest') || mw.config.get('wgAction') !== 'view') {
if (DEBUG) {
debug("Not the correct page or action, script terminated");
return console.log('[DuplicateReferences]', ...args);
}
}
 
if (
mw.config.get('wgAction') !== 'view' ||
mw.config.get('wgDiffNewId') ||
mw.config.get('wgDiffOldId') ||
(mw.config.get('wgNamespaceNumber') !== 0 && mw.config.get('wgPageName') !== 'User:Polygnotus/DuplicateReferencesTest')
) {
debug("Not the correct page or action, script terminated");
return;
}
 
debug("Page title:", document.title);
Line 23 ⟶ 29:
let nextElement = element.nextElementSibling;
while (nextElement) {
if (nextElement.tagName.toLowerCase() === 'div' &&
(nextElement.classList.contains('reflist') || nextElement.classList.contains('mw-references-wrap'))) {
return nextElement;
Line 70 ⟶ 76:
document.head.appendChild(style);
 
function addDuplicateCitationsTemplate(linkElement) {
debug("Adding duplicate citations template");
showLoading(linkElement);
var api = new mw.Api();
var pageTitle = mw.config.get('wgPageName');
 
let duplicateInfo = getDuplicateInfo();
 
// Get current date
const currentDate = new Date();
Line 93 ⟶ 100:
rvslots: 'main',
formatversion: 2
}).then(function (data) {
var page = data.query.pages[0];
var content = page.revisions[0].slots.main.content;
 
// Define the templates to check for
const templatesToCheck = [
'{{short description',
'{{DISPLAYTITLE',
'{{Lowercase title',
'{{Italic title',
'{{about',
'{{redirect'
];
 
// Find the position to insert the new template
let insertPosition = 0;
let lines = content.split('\n');
for (let i = 0; i < lines.length; i++) {
let line = lines[i].trim().toLowerCase();
if (templatesToCheck.some(template => line.startsWith(template.toLowerCase()))) {
insertPosition = i + 1;
} else if (line && !line.startsWith('{{') && !line.startsWith('__')) {
break;
}
}
 
// Create the reason string
let reason = 'DuplicateReferences script detected:\n\n';
if (duplicateInfo.length > 0) {
duplicateInfo.forEach((info, index) => {
reason += `* ${info.url} (refs: ${info.refs.map(r => r.number).join(', ')})\n\n`;
if (index < duplicateInfo.length - 1) {
reason += '; ';
}
});
}
 
// InsertCreate the new template with the reasonto parameterinsert
lines.splice(insertPosition,const 0,templateToInsert = `{{DuplicateDuplicated citations|reason=${reason}${dateParam}}}\n`);
 
var newContent = lines.join('\n');
// Use Morebits to handle the template insertion
const wikitextPage = new Morebits.wikitext.page(content);
// Define templates that should come before the duplicated citations template
const precedingTemplates = [
'short description',
'displaytitle',
'lowercase title',
'italic title',
'about',
'redirect',
'distinguish',
'for',
'Featured list',
'Featured article',
'Good article',
'Other uses',
'Redirect2',
'Use mdy dates',
'Use dmy dates',
'Use American English',
'Use British English'
];
 
// Insert the template after the specified templates
// The third parameter is flags (default 'i' for case-insensitive)
// The fourth parameter can include pre-template content like HTML comments
wikitextPage.insertAfterTemplates(templateToInsert, precedingTemplates, 'i', ['<!--[\\s\\S]*?-->']);
var newContent = wikitextPage.getText();
 
let summary = `+{{DuplicateTagged [[WP:DUPREF|duplicate citations]] using [[User:Polygnotus/DuplicateReferences|reason=${reason}${dateParam}}}DuplicateReferences]]`;
 
return api.postWithToken('csrf', {
Line 142 ⟶ 154:
summary: summary
});
}).then(function () {
mw.notifyshowSuccess('Successfully added the Duplicate citations template!'linkElement);
___location.reloadsetTimeout(function (); {
}) ___location.catchreload(function(error) {;
}, 100); // Reload after 0.1 second
}).catch(function (error) {
console.error('Error:', error);
showError(linkElement);
mw.notify('Failed to add the template. See console for details.', {type: 'error'});
});
}
 
function showLoading(element) {
element.innerHTML = '<sup><small>[ Working... ]</small></sup>';
}
 
function showSuccess(element) {
element.innerHTML = '<sup><small>[ Done ]</small></sup>';
}
 
function showError(element) {
element.innerHTML = '<sup><small>[ Error ]</small></sup>';
}
 
function getVisibleText(element) {
// Recursively get the visible text content of an element
let text = '';
for (let node of element.childNodes) {
if (node.nodeType === Node.TEXT_NODE) {
text += node.textContent.trim() + ' ';
} else if (node.nodeType === Node.ELEMENT_NODE) {
// Skip hidden elements
const style = window.getComputedStyle(node);
if (style.display !== 'none' && style.visibility !== 'hidden') {
text += getVisibleText(node) + ' ';
}
}
}
return text.trim();
}
 
function calculateLevenshteinDistance(a, b) {
debug("Comparing:");
debug("Text 1:", a);
debug("Text 2:", b);
 
if (a.length === 0) return b.length;
if (b.length === 0) return a.length;
 
const matrix = [];
 
// Increment along the first column of each row
for (let i = 0; i <= b.length; i++) {
matrix[i] = [i];
}
 
// Increment each column in the first row
for (let j = 0; j <= a.length; j++) {
matrix[0][j] = j;
}
 
// Fill in the rest of the matrix
for (let i = 1; i <= b.length; i++) {
for (let j = 1; j <= a.length; j++) {
if (b.charAt(i - 1) === a.charAt(j - 1)) {
matrix[i][j] = matrix[i - 1][j - 1];
} else {
matrix[i][j] = Math.min(
matrix[i - 1][j - 1] + 1, // substitution
Math.min(
matrix[i][j - 1] + 1, // insertion
matrix[i - 1][j] + 1 // deletion
)
);
}
}
}
 
debug("Levenshtein distance:", matrix[b.length][a.length]);
return matrix[b.length][a.length];
}
 
function calculateSimilarityPercentage(distance, maxLength) {
const similarity = ((maxLength - distance) / maxLength) * 100;
debug("Similarity percentage:", similarity.toFixed(2) + "%");
return Math.round(similarity) + '%';
}
 
function getDuplicateInfo() {
debug("Getting duplicate info");
 
const referenceItems = referencesList.children;
debug("Number of reference items:", referenceItems.length);
const urlMap = new Map();
const duplicates = [];
const urlMap = new Map();
const referenceItems = Array.from(referencesList.children);
 
letdebug("Number refNumberof =reference 0items:", referenceItems.length);
 
for (let item of referenceItems) {
referenceItems.forEach((item, index) => {
if (item.tagName.toLowerCase() === 'li') {
refNumber++;
const refId = item.id;
const refNumber = index + 1;
debug(`Processing reference item ${refNumber} (${refId})`);
const span = item.querySelector('span.reference-text');
if (!span) {
debug(` No reference-text span found in item ${refNumber}`);
continue;
}
const links = span.querySelectorAll('a');
debug(` Number of links in this span: ${links.length}`);
 
// Get the visible text of the entire reference item
const refText = getVisibleText(item);
debug(` Reference text: ${refText}`);
 
// Find the first valid link in the reference
const links = item.querySelectorAll('a');
let validLink = null;
for (let link of links) {
const url = link.href;
 
// Skip this reference if the URL doesn't contain 'http'
if (!url.includes('http')) {
debug(` Skipping reference ${refNumber} - URL does not contain 'http'`);
return; // This 'return' is equivalent to 'continue' in a regular for loop
}
const linkText = link.textContent.trim();
 
if (
// (!url.includes("wikipedia.org/wiki/") || url.includes("Special:BookSources")) &&
linkText !== "Archived" &&
(!url.includes("wikipedia.org/wiki/") || url.includes("Special:BookSources")) &&
!url.includes("_(identifier)") && // Templates like ISBN and ISSN and OCLC and S2CID contain (identifier)
!url.startsWith("https://search.worldcat.org/") && // |issn= parameter in cite news
!url.startsWith("https://www.bbc.co.uk/news/live/") && // live articles get frequent updates
!url.startsWith("https://www.aljazeera.com/news/liveblog/") &&
!url.startsWith("https://www.nbcnews.com/news/world/live-blog/") &&
!url.startsWith("https://www.theguardian.com/world/live/") &&
!url.startsWith("https://www.nytimes.com/live/") &&
!url.startsWith("https://edition.cnn.com/world/live-news/") &&
!url.startsWith("https://www.timesofisrael.com/liveblog") &&
!url.startsWith("https://www.france24.com/en/live-news/") &&
!url.startsWith("https://books.google.com/") && //may be 2 different pages of the same book
!url.startsWith("https://archive.org/details/isbn_")
) {
validLink = link;
Line 193 ⟶ 300:
const url = validLink.href;
if (urlMap.has(url)) {
urlMap.get(url).push({id: refId, number: refNumber, text: refText});
debug(` Duplicate found for URL: ${url}`);
} else {
urlMap.set(url, [{id: refId, number: refNumber, text: refText}]);
debug(` New URL added to map: ${url}`);
}
Line 203 ⟶ 310:
}
}
});
 
urlMap.forEach((refs, url) => {
if (refs.length > 1) {
duplicates.push({// url,Calculate refsLevenshtein });distance for each pair of refs
for (let i = 0; i < refs.length - 1; i++) {
for (let j = i + 1; j < refs.length; j++) {
debug(`Comparing references ${refs[i].number} and ${refs[j].number}:`);
const distance = calculateLevenshteinDistance(refs[i].text, refs[j].text);
const maxLength = Math.max(refs[i].text.length, refs[j].text.length);
const similarity = calculateSimilarityPercentage(distance, maxLength);
refs[i].similarity = refs[i].similarity || {};
refs[i].similarity[refs[j].id] = similarity;
}
}
duplicates.push({url, refs});
}
});
Line 227 ⟶ 345:
const headerCell = document.createElement('td');
headerCell.innerHTML = '<strong>Duplicate References</strong>';
 
const toggleSpan = document.createElement('span');
toggleSpan.className = 'mw-collapsible-toggle';
toggleSpan.innerHTML = '[<a href="#" class="mw-collapsible-text">hide</a>]';
headerCell.appendChild(toggleSpan);
 
 
// Check if the {{Duplicated citations}} template is already present
const duplicatedCitationsTemplate = document.querySelector('table.box-Duplicated_citations');
 
// Only add the link if the template is not present
if (!duplicatedCitationsTemplate) {
 
// Add the "add {{duplicated citations}}" link to the header
const addTemplateLink = document.createElement('a');
addTemplateLink.textContent = ' add {{duplicated citations}} ';
addTemplateLink.href = '#';
addTemplateLink.addEventListener('click', function (e) {
e.preventDefault();
addDuplicateCitationsTemplate(this);
});
//headerCell.appendChild(document.createTextNode(' ['));
headerCell.appendChild(addTemplateLink);
//headerCell.appendChild(document.createTextNode(']'));
}
headerRow.appendChild(headerCell);
tbody.appendChild(headerRow);
 
duplicateInfoconst pageTitle = mw.config.forEachget('wgPageName').replace({ url/_/g, refs' }') => {;
 
duplicateInfo.forEach(({url, refs}) => {
const row = document.createElement('tr');
const cell = document.createElement('td');
 
// Create report icon
const reportIcon = document.createElement('a');
reportIcon.href = `https://en.wikipedia.org/wiki/User_talk:Polygnotus?action=edit&section=new&preloadtitle=Reporting%20%5B%5BUser%3APolygnotus%2FDuplicateReferences%7CDuplicateReferences%5D%5D%20false-positive&preload=User:Polygnotus/$1&preloadparams%5b%5d=${encodeURIComponent(`[[${pageTitle}]] ${url}`)}%20~~~~`;
reportIcon.innerHTML = '<img src="https://upload.wikimedia.org/wikipedia/commons/thumb/e/ef/Cross_CSS_Red.svg/15px-Cross_CSS_Red.svg.png" width="15" height="15" alt="Report false positive" title="Report false positive" />';
reportIcon.style.marginRight = '5px';
cell.appendChild(reportIcon);
 
let urlLink = document.createElement('a');
Line 249 ⟶ 395:
cell.appendChild(document.createTextNode(' in refs: '));
 
const originalRef = refs[0];
refs.forEach((ref, index) => {
let link = document.createElement('a');
Line 255 ⟶ 402:
cell.appendChild(link);
 
// Add similarity information
if (index > 0) {
const similarity = calculateSimilarityPercentage(
calculateLevenshteinDistance(originalRef.text, ref.text),
Math.max(originalRef.text.length, ref.text.length)
);
let similarityInfo = document.createElement('span');
similarityInfo.textContent = ` (${similarity})`;
cell.appendChild(similarityInfo);
}
link.addEventListener('mouseover', () => {
refs.forEach(r => {
Line 304 ⟶ 461:
debug("Checking for duplicate reference links");
const duplicateInfo = getDuplicateInfo();
 
if (duplicateInfo.length > 0) {
debug("Duplicates found, creating collapsible table");
 
if (document.querySelector('table.box-Duplicated_citations') === null) {
const editSections = containerDiv.querySelectorAll('span.mw-editsection');
editSections.forEach(editSection => {
let spanBefore = document.createElement('span');
spanBefore.className = 'mw-editsection-bracket';
spanBefore.textContent = '[';
let addTemplateLink = document.createElement('a');
addTemplateLink.textContent = ' add {{duplicated citations}} ';
addTemplateLink.href = '#';
addTemplateLink.addEventListener('click', function(e) {
e.preventDefault();
addDuplicateCitationsTemplate();
});
let spanAfter = document.createElement('span');
spanAfter.className = 'mw-editsection-bracket';
spanAfter.textContent = ']';
editSection.appendChild(spanBefore);
editSection.appendChild(addTemplateLink);
editSection.appendChild(spanAfter);
});
}
const table = createCollapsibleTable(duplicateInfo);
containerDiv.after(table);
Line 357 ⟶ 488:
setTableState(initialState);
 
toggleLink.addEventListener('click', function (e) {
e.preventDefault();
const isCurrentlyCollapsed = tableBody.is(':hidden');
Line 366 ⟶ 497:
}
}
 
checkDuplicateReferenceLinks();
debug("Script execution completed");
Retrieved from "https://en.wikipedia.org/wiki/User:Polygnotus/DuplicateReferences.js"