Jump to content

User:N8wilson/AQFetcher.js

From Wikipedia, the free encyclopedia
This is an old revision of this page, as edited by N8wilson (talk | contribs) at 15:45, 31 October 2015 (Adding styling for some classifications). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.
Note: After saving, you have to bypass your browser's cache to see the changes. Google Chrome, Firefox, Microsoft Edge and Safari: Hold down the ⇧ Shift key and click the Reload toolbar button. For details and instructions about other browsers, see Wikipedia:Bypass your cache.
/* Using User:Anomie/linkclassifier.js as a template.  Original comment from that script included below:

-----------------------
If you want to use this script, simply add the following line to your [[Special:Mypage/monobook.js]]:

importScript('User:Anomie/linkclassifier.js'); // Linkback: [[User:Anomie/linkclassifier.js]]

* (Please keep the comment so I can see how many people use this). You will also want to
* add some CSS classes, such as those at [[User:Anomie/linkclassifier.css]].
*/

/* If you want this to run "on demand" instead of on every page, set "LinkClassifierOnDemand=true" and
 * use addPortletLink() or the like to add a button calling LinkClassifier.onDemand().
 */


var LinkClassifier={
    /* This object maps classes to the categories for which to apply them. Values may be an array of strings or a regex. */
    cats:{
        deletion:[
            'Category:All articles proposed for deletion',
            'Category:All books proposed for deletion',
            'Category:All disputed non-free Wikipedia files',
            'Category:All orphaned non-free use Wikipedia files',
            'Category:All possibly unfree Wikipedia files',
            'Category:All redirects for discussion',
            'Category:All replaceable non-free use Wikipedia files',
            'Category:All Wikipedia files with no copyright tag',
            'Category:All Wikipedia files with no non-free use rationale',
            'Category:All Wikipedia files with unknown copyright status',
            'Category:All Wikipedia files with unknown source',
            'Category:Articles for deletion',
            'Category:Articles on deletion review',
            'Category:Candidates for speedy deletion',
            'Category:Candidates for undeletion',
            'Category:Categories for conversion',
            'Category:Categories for deletion',
            'Category:Categories for listifying',
            'Category:Categories for merging',
            'Category:Categories for renaming',
            'Category:Categories for speedy renaming',
            'Category:Categories to be listified then deleted',
            'Category:Duplicate or hardcoded templates awaiting deletion',
            'Category:Items pending OTRS confirmation of permission for over 30 days',
            'Category:Miscellaneous pages for deletion',
            'Category:Stub categories for deletion',
            'Category:Stub template deletion candidates',
            'Category:Templates for deletion',
            'Category:Wikipedia deprecated and orphaned templates',
            'Category:Wikipedia files for deletion',
            'Category:Wikipedia files with unknown source for deletion in dispute',
            'Category:Wikipedia templates for deletion'
        ].sort(),
        disambiguation:[
            'Category:All disambiguation pages'
        ].sort(),
        'set-index':[
            'Category:All set index articles'
        ].sort(),
        'featured-content':[
            'Category:Featured articles',
            'Category:Featured lists',
            'Category:Featured pictures',
            'Category:Featured sounds',
            'Category:Featured videos',
            'Category:Featured portals'
        ].sort(),
        'good-content':[
            'Category:Good articles'
        ].sort(),
        'soft-redirect-cats':[
            'Category:Wikipedia soft redirected categories',
        ].sort(),
        'spoken-articles':[
            'Category:Spoken articles'
        ].sort(),
        stubcls:/^Category:.* stubs$/,
        'nonfree-media':[
            'Category:All non-free media'
        ].sort(),
        unprintworthy:[
            'Category:Unprintworthy redirects',
            'Category:Middle-earth redirects from redundant titles'
        ].sort(),
        'unprintworthy-shortcut':[
            'Category:Redirects from shortcuts',
        ].sort()
    },

    articleClasses: {
      'FA':[], 'GA':[], 'A':[], 'B':[], 'C':[], 'Start':[], 'Stub':[], 'FL':[], 'L':[], 'Unclassified':[]
    },


    /* This regex matches page titles to be marked as intentional links to disambiguation pages */
    intentionaldab: / \(disambiguation\)$/,
 
    callback:function(r, sts, xhr){
        if(!r.query) {
            if(typeof(window.console)=='undefined' || typeof(window.console.error)!='function')
                throw new Error('Bad response');
            window.console.error("Bad response", r);
            return;
        }
        if(r['query-continue']){
            var cc=this.rawdata;
            for(var k in r['query-continue']){
                for(var k2 in r['query-continue'][k]){
                    cc[k2]=r['query-continue'][k][k2];
                }
            }
            $.ajax({
                url:mw.util.wikiScript('api'),
                dataType:'json',
                type:'POST',
                data:cc,
                rawdata:cc,
                success:arguments.callee,
                error:function(xhr,textStatus,errorThrown){
                    throw new Error('AJAX error: '+textStatus+' '+errorThrown);
                }
            });
        }
        r=r.query;

        var a=document.getElementById('wikiPreview');
        if(!a) a=document.getElementById('bodyContent');
        if(!a) throw new Error('Huh? No body content?');
        a=a.getElementsByTagName('A');
        if(a.length==0) return;

        // Handy shortcut
        var LC = LinkClassifier;

        var prefix=(this.rawdata.redirects?'':'redir-');
        var cats={};
        var missing={};
        var classes={};
        if(r.pages) for(var i in r.pages){
            classes[r.pages[i].title] = [];
            missing[r.pages[i].title]=(typeof(r.pages[i].missing)!='undefined');
            if(typeof(r.pages[i].categories)!='undefined'){
                
                // If the corresponding Talk: page has appropriately-named categories we can
                // match each article page to it's classification

                unClassified = true;
                iCat = 0;
                while (unClassified && iCat < r.pages[i].categories.length) {
                   if (r.pages[i].categories[iCat].title!='undefined') {
                      if (r.pages[i].categories[iCat].title.indexOf('FA-Class') >= 0) {
                         LC.articleClasses['FA'].push(LC.getPageNameFromTalkPage(r.pages[i].title));
                         unClassified = false;
                      } else if (r.pages[i].categories[iCat].title.indexOf('A-Class') >= 0) {
                         LC.articleClasses['A'].push(LC.getPageNameFromTalkPage(r.pages[i].title));
                         unClassified = false;
                      } else if (r.pages[i].categories[iCat].title.indexOf('GA-Class') >= 0) {
                         LC.articleClasses['GA'].push(LC.getPageNameFromTalkPage(r.pages[i].title));
                         unClassified = false;
                      }  else if (r.pages[i].categories[iCat].title.indexOf('B-Class') >= 0) {
                         LC.articleClasses['B'].push(LC.getPageNameFromTalkPage(r.pages[i].title));
                         unClassified = false;
                      }  else if (r.pages[i].categories[iCat].title.indexOf('C-Class') >= 0) {
                         LC.articleClasses['C'].push(LC.getPageNameFromTalkPage(r.pages[i].title));
                         unClassified = false;
                      }  else if (r.pages[i].categories[iCat].title.indexOf('Start-Class') >= 0) {
                         LC.articleClasses['Start'].push(LC.getPageNameFromTalkPage(r.pages[i].title));
                         unClassified = false;
                      }  else if (r.pages[i].categories[iCat].title.indexOf('Stub-Class') >= 0) {
                         LC.articleClasses['Stub'].push(LC.getPageNameFromTalkPage(r.pages[i].title));
                         unClassified = false;
                      }  
                   }
                   iCat+=1;
                }
                if (unClassified) LC.articleClasses['Unclassified'].push(LC.getPageNameFromTalkPage(r.pages[i].title));
            }
        }

        Array.prototype.forEach.call(a, function(a){
            if(typeof(a.wikipage)=='undefined') return;
            if (LC.articleClasses['C'].indexOf(a.wikipage) >= 0 ||
                LC.articleClasses['Start'].indexOf(a.wikipage) >= 0 ||
                LC.articleClasses['Stub'].indexOf(a.wikipage) >= 0 ||
                LC.articleClasses['Unclassified'].indexOf(a.wikipage) >= 0) {
                $(a).addClass('classification-nonactive');
            }

            /*  UNUSED ...  TODO: Remove
            var m=a.href.match(/#.*/);
            if(m && m[0].substr(0,10)!=="#cite_note"){
                a.title=a.title.replace(/#.*/,'')+m[0].replace(/_/g,' ').replace(/\.([0-9A-F][0-9A-F])/gi, function(x,n){ return String.fromCharCode(parseInt(n,16)); });
            }
            if(LinkClassifier.intentionaldab.test(a.origwikipage)){
                $(a).addClass('intentional-disambiguation');
            }
            if(typeof(classes[a.wikipage])!='undefined'){
                for(var j=classes[a.wikipage].length-1; j>=0; j--)
                    $(a).addClass(classes[a.wikipage][j]);
            }
            if(a.wikipage!=a.origwikipage && typeof(classes[a.origwikipage])!='undefined'){
                for(var j=classes[a.origwikipage].length-1; j>=0; j--)
                    $(a).addClass(classes[a.origwikipage][j]);
            }
            var c1=[];
            if(typeof(cats[a.wikipage])!='undefined'){
                c1=c1.concat(cats[a.wikipage]);
            }
            if(a.wikipage!=a.origwikipage && typeof(cats[a.origwikipage])!='undefined'){
                c1=c1.concat(cats[a.origwikipage]);
            }
            if(c1.length>0){
                c1=c1.sort();
                for(var cls in LinkClassifier.cats){
                    var i1=c1.length-1;
                    var c2=LinkClassifier.cats[cls];
                    if(c2 instanceof RegExp){
                        while(i1>=0){
                            if(c2.test(c1[i1])){
                                $(a).addClass(cls);
                                break;
                            }
                            i1--;
                        }
                    } else {
                        var i2=c2.length-1;
                        while(i1>=0 && i2>=0){
                            if(c1[i1]==c2[i2]){
                                $(a).addClass(cls);
                                break;
                            }
                            (c1[i1]>c2[i2])?--i1:--i2;
                        }
                    }
                }
            }  // */
        });
    },

    /* Extract the name of the WP article from the URL */
    getPageName:function(url){
        var m=url.match(/\/wiki\/([^?#]+)/);
        if(!m) m=url.match(/\/w\/index.php\?(?:.*&)?title=([^&#]+)/);
        if(!m) return '';
        var t=decodeURIComponent(m[1]).replace(/_/g,' ');
        if(t.substr(0,6)=='Image:') t='File:'+t.substr(6);
        if(t.substr(0,11)=='Image talk:') t='File talk:'+t.substr(6);
        if(t.substr(0,8)=='Special:') t='';
        return t;
    },

    /* If given a pagename which is not in the 'Talk:' namespace, adds 'Talk:' and returns the name of the page in the talk 
     * namespace. Returns unaltered input if the provided name is already in the "Talk:" namespace */
    getTalkPageFromPageName:function(name){
       return (name.substr(0,5)=='Talk:') ? name : 'Talk:'+name;
    },

    /* If given a pagename in "Talk:" namespace, removes the namespace to return the page name
     * returns unaltered input if the provided name is not in the "Talk:" namespace */
    getPageNameFromTalkPage:function(name){
       return (name.substr(0,5)=='Talk:') ? name.substr(5,name.length) : name;
    },

    classifyChildren:function(node){
        mw.loader.using(['mediawiki.util','mediawiki.user'], function(){
            var a=node.getElementsByTagName('A');
            if(a.length==0) return;
            var self=LinkClassifier.getPageName(location.href);
            a=Array.prototype.map.call(a, function(a){
                a.wikipage='';
                if(/(^|\s)(external|extiw)(\s|$)/.test(a.className)) return '';
                if(!/(^|\s)(image)(\s|$)/.test(a.className)) a.className+=" nonimage";
                a.wikipage=LinkClassifier.getPageName(a.href);
                if(a.wikipage==self) a.wikipage='';
                a.origwikipage=a.wikipage;
                return LinkClassifier.getTalkPageFromPageName(a.wikipage); //return a.wikipage;
            }).sort().filter(function(e,i,a){
                return e!=='' && (i==0 || a[i-1]!==e);
            });
            // a is now a sorted array of links with wikipage attributes holding the title of talk pages


            function processLinks(limit){
                while(a.length>0){
                    var q={
                        format:'json',
                        action:'query',
                        titles:a.splice(0,limit).join('|'),
                        prop:'categories', // Only interested in Talk page categories
                        cllimit:'max',
                        rawcontinue:1
                    };
                    $.ajax({
                        url:mw.util.wikiScript('api'),
                        dataType:'json',
                        type:'POST',
                        data:q,
                        rawdata:q,
                        success:LinkClassifier.callback,
                        error:function(xhr,textStatus,errorThrown){
                            throw new Error('AJAX error: '+textStatus+' '+errorThrown);
                        }
                    });
                }
            }

            if(a.length<=100){
                // Not worth querying the API to see if the user has apihighlimits
                processLinks(50);
            } else {
                // Note mw.user.getRights queries the API
                mw.user.getRights(function(rights){
                    processLinks( (rights.indexOf('apihighlimits')>=0) ? 500 : 50 );
                });
            }  
        });
    },
 
    onLoad:function(){
        if(window.LinkClassifierOnDemand) return;
        if(window.AJAXPreview) window.AJAXPreview.AddOnLoadHook(LinkClassifier.classifyChildren);
        LinkClassifier.onDemand();
    },
 
    onDemand:function(){
        var node=document.getElementById('wikiPreview');
        if(!node) node=document.getElementById('bodyContent');
        if(node) LinkClassifier.classifyChildren(node);
    }
};

if(!window.LinkClassifierOnDemand) $(document).ready(LinkClassifier.onLoad);