User:Trappist the monk/HarvErrors.js

// tweaked copy of [[User:Ucucha/HarvErrors.js]]

mw.loader.addStyleTag (
	'.ttm_harv_err {color:DarkOrange}' +										// no target error message default color
	'.ttm_harv_err_dup {color:#C3F}' +											// duplicate target error message default color (sfn or harv template – source)
	'.ttm_err_harv_dup {color:#C3F}' +											// duplicate target error message default color (citation template – target)
	'.ttm_harv_warn {color:#ac6600}'											// warning message default color
	);

mw.hook( 'wikipage.content' ).add( function( $content )
	{
	var cites = $content.find('.citation');										// get all cites on the page
	var citeref_tgt_counts = {};												// associative array of citerefs as key, their number as value
	var citeref_source_ids = {};												// associative array of citerefs from sfn, harv templates
	var idi;																	// citeref id in and iterated loop
	
	for (var i=0; i < cites.length; i++)										// fill the array; to be used when identifying multiple targets
		{
		idi = cites[i].getAttribute('id');										// get citeref anchor from cites[]
		if ((idi in citeref_tgt_counts))										// is target in the associative array
			citeref_tgt_counts[idi]++;											// yep, bump the count
		else
			citeref_tgt_counts[idi] = 1;										// nope, add it with a count of one
		}

	var href, links = $content.find( 'a[href^="#CITEREF"]' );					// make a list of harv and sfn links
		
	links.each( function (i, elem)												// first check: do links in Harvard citations point to a valid citation?
		{
		href = elem.getAttribute( 'href' ).substring(1); //skip the #
		// IDs can contain characters like . that have meaning in selectors
		// use $.escapeSelector to make sure they are escaped
		if ( $content.find( '#' + $.escapeSelector(href) ).length < 1)
			$(elem.parentNode).append(" ", $("<span class=\"ttm_harv_err\">Harv error: link from " + href + " doesn't point to any citation.</span>"));
		else if (1 < citeref_tgt_counts[href])
			{
			$(elem.parentNode).append(" ", $("<span class=\"ttm_harv_err_dup\">Harv error: " + href + " has multiple targets (" + citeref_tgt_counts[href] + "×).</span>"));
			citeref_source_ids[href] = 1;										// add this citeref to the list; duplicates simply overwrite
			}
		});

	// second check: do CITEREF IDs have Harvard citations pointing to them?

	if (0 !== links.length)														// non-zero when links were found
		{
		var further_reading = $content.find('#Further_reading').parent().nextUntil('h2').find('.citation').get();	// get all cites inside a Further reading section
		var external_links = $content.find('#External_links').parent().nextUntil('h2').find('.citation').get();		// get all cites inside a External links section
		var publications = $content.find('#Publications').parent().nextUntil('h2').find('.citation').get();			// get all cites inside a Publications section

		var sections = [further_reading, external_links, publications];			// make an array of links in these sections
		
		var filtered_cites = cites.get().filter(function(x) { return further_reading.indexOf(x) === -1; });	// all cites not in 'Further reading' section
		filtered_cites = filtered_cites.filter(function(x) { return external_links.indexOf(x) === -1; });	// all cites not in 'External links' section
		filtered_cites = filtered_cites.filter(function(x) { return publications.indexOf(x) === -1; });		// all cites not in 'Publications' section
		
		var id;
		var query;

		for(i=0; i < filtered_cites.length; i++)
			{
			id = filtered_cites[i].getAttribute('id');
			// we only need to check citations with an id that is a CITEREF id
			if(!id || id.indexOf('CITEREF') !== 0)
				continue;
			// don't do cites that are inside a ref
			var parentid = filtered_cites[i].parentNode.parentNode.getAttribute('id');
			if(parentid && parentid.indexOf('cite_note') === 0)
				continue;
			// check for links to this citation
			query = 'a[href|="#' + $.escapeSelector(id) + '"]';
			if($content.find(query).length === 0)
				$(filtered_cites[i]).append(" ", $("<span class=\"ttm_harv_warn\">Harv warning: There is no link pointing to this citation. The anchor is named " + id + ".</span>"));
			}

		// third check: are Harvard citations pointing to citerefs in inappropriate article sections?
		for (var s=0; s < sections.length; s++)									// loop through the list of sections
			{
			for (i=0; i < sections[s].length; i++)								// look for linked cites in sections[s]
				{
				id = sections[s][i].getAttribute('id');
				query = 'a[href|="#' + $.escapeSelector(id) + '"]';
				if($content.find(query).length !== 0)
					$(sections[s][i]).append(" ", $("<span class=\"ttm_harv_err\">Harv error: linked from " + id + ".</span>"));
				}
			}

		// fourth check: do Harvard citations have multiple targets?
		for (i=0; i < cites.length; i++)
			{
			idi = cites[i].getAttribute('id');
			if (!(idi in citeref_source_ids))									// only check citations that have matching sfn or ref templates
				continue;
				
			if(!idi || idi.indexOf('CITEREF') !== 0)							// only check citations with an id that is a CITEREF id
				continue;

			if (1 < citeref_tgt_counts[idi])
				$(cites[i]).append(" ", $("<span class=\"ttm_err_harv_dup\">Harv error: duplicate target for " + idi + ".</span>"));
			}
		}
	});