whatizit

function HTMLescape(string)  {  
      var LT = new RegExp("<", "g");  
      var GT = new RegExp(">", "g");  
      var AMP = new RegExp("&", "g");  
      var TAB = new RegExp("\t", "g");  
      return  string.replace(AMP,"&amp;").replace(LT, "&lt;").replace(GT, "&gt;").replace(TAB, "...");
}  

CmdUtils.CreateCommand({  
    name: "whatizit",  
    takes: {"selected text": noun_arb_text},  
    execute: function(directObj) {},    
    preview: function(pblock, directObj) { 
      var t=directObj.text;
      if (t.match(/go/i)){t='go'} else
      if (t.match(/prot/i)){t='uniprot'} else {t='go'};
      var inputText=HTMLescape(CmdUtils.getSelection());

      var data="whatizitEBIMed\n"+
      "<document xmlns:xlink='http://www.w3.org/1999/xlink' xmlns:z='http://www.ebi.ac.uk/z' source='Whatizit'><text>"+inputText+"</text></document>"; 

      pblock.innerHTML = "whatizit is looking for <b>"+t+"</b> terms in <i>'"+inputText.substring(0,100)+"...'</i>";

      var soap = '<soap:Envelope xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:soap="http://schemas.xmlsoap.org/soap/envelope/" xmlns:wz="http://www.ebi.ac.uk/webservices/whatizit/ws"> ' +  
                  '    <soap:Body> ' +  
		  '        <wz:contact> '+      
		  '            <pipelineName>whatizitEBIMed</pipelineName> ' +
		  '            <text>' + inputText + '</text> ' +
		  '            <convertToHtml>false</convertToHtml> ' +  
		  '        </wz:contact> ' +  
		  '     </soap:Body> ' +  
		  '</soap:Envelope> '; 

      var url="http://www.ebi.ac.uk:80/webservices/whatizit/ws";

      jQuery.ajax({  
        type:     'POST',  
        url:      url,  
        dataType: "xml",  
        data:     soap,  
        beforeSend: function(req) {  
           req.setRequestHeader("Content-Type", "text/xml");
           req.setRequestHeader("Method", "POST http://www.ebi.ac.uk/webservices/whatizit/ws HTTP/1.1");
        },
        success: function(response) {
             pblock.innerHTML = "analyzing tagged document...";
             var serializer = new window.XMLSerializer();
             var parser = new window.DOMParser();
             xml=parser.parseFromString(jQuery(response).find("ns1\\:contactResponse return").text(),"text/xml");

             var tags={go: "z\\:go", uniprot: "z\\:uniprot"};
             var terms={};
             var count={};
             var inSentence={};
             terms[t]=[];
             count[t]=[];
             inSentence[t]=[];
             jQuery(xml).find("SENT").each(function(i){
               var sid=jQuery(this).attr("sid");
               var elements=jQuery(this).find(tags[t]);
               elements.each(function(i){
                 var ids=jQuery(this).attr("ids");
                 if (!terms[t][ids]){terms[t][ids]=jQuery(this).text()};
                 //problem: there migth be various instances of a entity
                 //we should keep order in sentence of instance
                 if (count[t][ids]){count[t][ids]++} else {count[t][ids]=1};
                 if (!inSentence[t][ids]){inSentence[t][ids]=[]};
                 if (inSentence[t][ids].indexOf(sid)<0){inSentence[t][ids].push(sid)};
               });
             });

             var sentences={};
             jQuery(xml).find("SENT").each(function(i){
               //will need to replace <z:*> tags by <span class="tag*" id="inlineTag*XXX">
               var sentence=jQuery(this).text();
               var sid=jQuery(this).attr("sid");
               sentences[sid]=sentence;
             });

             var doc=pblock.ownerDocument;
             doc.inSentence=inSentence;
             doc.sentences=sentences;
             doc.showSentences=function(id,tagType,term){
               var html="<ul>";
               list=inSentence[tagType][id];
               for (var i in list){
                 var re= new RegExp(term+"\\b","gi");
                 html+="<li>"+sentences[list[i]].replace(re,"<span class='hi'>"+term+"</span>")+"</li>"
               };
               html+="</ul>"
               return html;
             };

             //from http://news.hping.org/comp.lang.javascript.archive/0867.html
             function sortAssoc(aInput){
               var aTemp = [];
               for (var sKey in aInput)
                 aTemp.push([sKey, aInput[sKey]]);
                 aTemp.sort(function () {return arguments[0][1] > arguments[1][1]});
                 var aOutput = [];
                 for (var nIndex = aTemp.length-1; nIndex >=0; nIndex--)
                 aOutput[aTemp[nIndex][0]] = aTemp[nIndex][1];
               return aOutput;
             };
 
             var sorted=sortAssoc(count[t]);

             var style="<style type='text/css'>.cloud{padding:5px 5px; background: white; color: blue; font: arial} .reactive:hover{cursor: pointer; text-decoration:underline} .hi{text-decoration:underline; color: orange}</style>";
             var html="<div class='cloud'>";
             for (id in sorted){
                var term=terms[t][id];
                var n=sorted[id];
                var size=12+5*Math.round(Math.log(Math.min(n,10)));
                html+="<span style='font-size:"+size+"px' class='reactive' onmouseover='document.getElementById(\"sentences\").innerHTML=showSentences(\""+id+"\",\""+t+"\",\""+term+"\")'>"+term+"</span> ";
             };
             html+="</div>";
             html+="<div id='sentences'></div>";
             
             pblock.innerHTML = style+
                                //"<code>"+HTMLescape(serializer.serializeToString(xml))+"</code>"+"<br/><br/>"+
                                html;          
        },
        error: function(XMLHttpRequest, status, errorThrown){pblock.innerHTML = 'status: '+status;}
      })
   }  
})