view js/taggingtext.js @ 88:61593b047289 extractapp

merged from extractapp_dev
author Zoe Hong <zhong@mpiwg-berlin.mpg.de>
date Wed, 03 Jun 2015 10:49:00 +0200
parents fb5049fc5dd7
children e681d693240e
line wrap: on
line source

/*
 * taggingtext.js
 * This file is part of Extraction-interface.
 *
 * Extraction-interface is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * Extraction-interface is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with Extraction-interface.  If not, see <http://www.gnu.org/licenses/>.
*/

/*! \file  
 * There are javscript functions for Extractapp. 
 * It is included by views/Extractapp/maintemplate.php and maintemplate_local.php.
*/ 

// === for taggingtext.php ===
var stringBeforeChange="";
var stringBeforeChangeStack = [];
var lastAddTag="";

var regex_element_index = 0;

window.onbeforeunload = function() {
    return "Are you sure?";
};


/*
$(document).on("mouseup", function (e) {
    e.stopPropagation();
    $('.questionMarkClass').remove();
    $('.tagItemDivClass').remove();
    alert("Remove");
});


$(document).on("click", '.questionMarkClass', function (e) {
    e.stopPropagation();
});
*/
function tagTime() {
    saveUndoText();
    var el = document.getElementById("editable-area");
    var str="" + el.innerHTML;

    var regexText=/(<?php echo $wordlistArray[0][2]; ?>)(○?)(一|二|三|四|五|六|七|八|九)?(○?)(十)?(○?)(一|二|三|四|五|六|七|八|九|元|十)(○?)(年)/g;
    var replaceText="<time>$1$2$3$4$5$6$7$8$9</time>";
    var matchedCount = str.match(regexText).length;
    str = str.replace(regexText, replaceText);
    
    var regexText=/(?!(>))(<?php echo $wordlistArray[0][2]; ?>)/g;
    var replaceText="<time>$1$2</time>";
    matchedCount += str.match(regexText).length;
    el.innerHTML = str.replace(regexText, replaceText);
    
    var regexText=/(一|二|三|四|五|六|七|八|九)?(○?)(十)?(○?)(一|二|三|四|五|六|七|八|九|元|十)(○?)(年)(?!(<))/g;
    var replaceText="<time>$1$2$3$4$5$6$7</time>";
    matchedCount += str.match(regexText).length;
    el.innerHTML = str.replace(regexText, replaceText);
    
    alert( "Tagged "+matchedCount+" entities!" );
}
function removeEmptyNodes( node ) {
    if (node.hasChildNodes()) {
        var children = node.childNodes;
        for (var i = 0; i < children.length; i++) {
            if (children[i].textContent == "" && children[i].innerHTML == "" && children[i].nodeName != "BR") {
                //console.log("removing child: "+children[i].nodeName+", its parent: "+node.nodeName);
                node.removeChild(children[i]);
            } else {
                //console.log("recursive to "+children[i].nodeName+"("+children[i].textContent+")")
                removeEmptyNodes(children[i]);  // recursive
            }

        };
    } else {
        //console.log(node.nodeName+" ("+node.textContent+") has no childNodes.");
    }

}


function removeTagTitle( range, container ) {
    saveUndoText();
    range.deleteContents();
    
    lastAddTag = $('#RemoveTitletagType').val();
    
    var stringtemp = container.innerHTML;
    
    if ( $('#RemoveTitletagName').val() == "" ) {
        var regexText="<"+lastAddTag+">〈(.*?)〉</"+lastAddTag+">";
        var replaceText="";
        stringtemp = stringtemp.replace(new RegExp(regexText, "g"), replaceText);
        
        var regexText="<"+lastAddTag+">(.*?)</"+lastAddTag+">";
        var replaceText="$1";
        stringtemp = stringtemp.replace(new RegExp(regexText, "g"), replaceText);
    } else {
        var regexText="<"+lastAddTag+">〈"+$('#RemoveTitletagName').val()+"〉</"+lastAddTag+">";
        var replaceText="";
        stringtemp = stringtemp.replace(new RegExp(regexText, "g"), replaceText);
        
        var regexText="<"+lastAddTag+">("+$('#RemoveTitletagName').val()+")</"+lastAddTag+">";
        var replaceText="$1";
        stringtemp = stringtemp.replace(new RegExp(regexText, "g"), replaceText);
    }
    
    var newdiv = document.createElement("aaaa");
    newdiv.innerHTML = stringtemp;
    range.insertNode(newdiv);
    
    var el = document.getElementById("editable-area");
    
    var regexText=/<aaaa>/gi;
    var replaceText='';
    el.innerHTML = el.innerHTML.replace(regexText, replaceText);
    
    var regexText=/<\/aaaa>/gi;
    var replaceText='';
    el.innerHTML = el.innerHTML.replace(regexText, replaceText);
    
    $('.tagItemDivClass').remove();
}

function addTagTitle( range, container ) {
    // TODO: bug: here generate self-closing tag when the end of text with tag on it

    saveUndoText();

    var el = document.getElementById("editable-area");
    
    lastAddTag = $('#TitletagType').val();
    var tag = "<"+$('#TitletagType').val()+">〈"+$('#TitletagName').val()+"〉</"+$('#TitletagType').val()+">";

    range.deleteContents(); // may causes self-closing tag

    var stringtemp = container.innerHTML;

    var regexText=/<br>/g;
    var replaceText="<br>"+tag;
    stringtemp = stringtemp.replace(regexText, replaceText);
  
    var newdiv = document.createElement("aaaa");
    newdiv.innerHTML = tag+stringtemp;

    range.insertNode(newdiv);
   

    var regexText=/<aaaa>/gi;
    var replaceText='';
    el.innerHTML = el.innerHTML.replace(regexText, replaceText);
    
    var regexText=/<\/aaaa>/gi;
    var replaceText='';
    el.innerHTML = el.innerHTML.replace(regexText, replaceText);
    
    $('.tagItemDivClass').remove();

    removeEmptyNodes(el);   // removing self-closing tags

}

function cleanUpTextArea() {
    var el = document.getElementById("editable-area");
    
    var regexText=/<div>/gi;
    var replaceText='<br>';
    el.innerHTML = el.innerHTML.replace(regexText, replaceText);
    
    var regexText=/<\/div>/gi;
    var replaceText='';
    el.innerHTML = el.innerHTML.replace(regexText, replaceText);
    
    
    var regexText=/<span style="(.*?)">/gi;
    var replaceText='';
    el.innerHTML = el.innerHTML.replace(regexText, replaceText);
    
    var regexText=/<\/span>/gi;
    var replaceText='';
    el.innerHTML = el.innerHTML.replace(regexText, replaceText);
}

function Undo() {
    if ( stringBeforeChangeStack.length > 0 ) {
        var el = document.getElementById("editable-area");
        el.innerHTML = stringBeforeChangeStack.pop();
    } else {
        var el = document.getElementById("buttonUndo");
        el.disabled = true;
    }
}

function saveUndoText() {
    var el = document.getElementById("editable-area");
    //stringBeforeChange = el.innerHTML;
    stringBeforeChangeStack.push(el.innerHTML); 
    var el = document.getElementById("buttonUndo");
    el.disabled = false;
}

function editText() {
    var el = document.getElementById("editable-area");
    if ( $("#editTextId").html() == "Edit the text" ) {
        // --- in editing mode ---
        saveUndoText();
        el.contentEditable = true;

        // disable/close functions
        $("button").attr("disabled", true);
        $("#editTextId").attr("disabled", false);
        $("#popups").css("display","none"); // close the popup windows
        $("input").attr('disabled', true); // disable input
        // clean editable-area
        $('.questionMarkClass').remove();
        $('.tagItemDivClass').remove();
        
        $("#editTextId").html("Edit completed!");
    } else {
        // --- not editing mode ---
        el.contentEditable = false;
        cleanUpTextArea();
        $("button").attr("disabled", false);
        $("#popups").css("display","block"); // open the popup windows
        $("input").attr('disabled', false); // enable input
        $("#editTextId").html("Edit the text"); 
    }
}


function removeTagNewDiv( eventObject, tagName, tagObject ) {
    // TODO: nesting tag representaion

    saveUndoText();
    var newdiv = document.createElement("div");
    $(newdiv).id = "questionMarkId";
    $(newdiv).attr("class", "questionMarkClass");

    // set z-index to 3 to bring popup tag windwo to front
    newdiv.style.cssText = 'top:'+eventObject.pageY+'; left:'+eventObject.pageX+';';
    
    newdiv.innerHTML = "Tag: "+tagName+"<br>Value: "+tagObject.text()+"<br>";
    
    var newbutton = $('<button class="btn btn-default">Remove this</button>').mouseup(function (e2) {
        var textKeep = $(this).parent().parent().html();
        var regexText=/<div(.*?)<\/div>/g;
        var replaceText="";
        textKeep = textKeep.replace(regexText, replaceText);
    
        $(this).parent().parent().replaceWith( textKeep );
    });
    newbutton.appendTo(newdiv);
    
    var newbutton = $('<button class="btn btn-default">Remove this(with newline)</button>').mouseup(function (e2) {
        var textKeep = $(this).parent().parent().html();
        var regexText=/<div(.*?)<\/div>/g;
        var replaceText="";
        textKeep = textKeep.replace(regexText, replaceText);
    
        var newLineBefore = $(this).parent().parent().prev();
        if ( newLineBefore.prop("tagName") == "BR" ) {
            $(this).parent().parent().prev().replaceWith( );
        }
        $(this).parent().parent().replaceWith( textKeep );
    });
    newbutton.appendTo(newdiv);
    
    var newbutton = $('<button class="btn btn-default">Remove all</button>').mouseup(function (e2) {
        var textKeep = $(this).parent().parent().html();
        var regexText=/<div(.*?)<\/div>/g;
        var replaceText="";
        textKeep = textKeep.replace(regexText, replaceText);
    
        $(this).parent().parent().replaceWith( textKeep );
        
        var el = document.getElementById("editable-area");
        var regexText=new RegExp("<"+tagName+">("+textKeep+")</"+tagName+">", "g");
        var replaceText="$1";
        var str="" + el.innerHTML;
        
        if ( str.match(regexText)==null ) {
            alert( "Removed 1 entity!" );
        } else {
            alert( "Removed "+(parseInt(str.match(regexText).length)+1)+" entities!" );
        }
        el.innerHTML = str.replace(regexText, replaceText);
    });
    newbutton.appendTo(newdiv);
    
    var newbutton = $('<button class="btn btn-default">Remove all(with newline)</button>').mouseup(function (e2) {
        var textKeep = $(this).parent().parent().html();
        var regexText=/<div(.*?)<\/div>/g;
        var replaceText="";
        textKeep = textKeep.replace(regexText, replaceText);
    
        $(this).parent().remove();
        
        var el = document.getElementById("editable-area");
        var regexText=new RegExp("<br><"+tagName+">("+textKeep+")</"+tagName+">", "g");
        var replaceText="$1";
        var str="" + el.innerHTML;
        
        alert( "Removed "+str.match(regexText).length+" entities!" );
        el.innerHTML = str.replace(regexText, replaceText);
    });
    newbutton.appendTo(newdiv);
    
    tagObject.append(newdiv);
}

function saveText(section_id) {
    var el = document.getElementById("editable-area");
    $.ajax({
        url : './'+section_id,
        async : false,
        type : 'POST',
        data : 'func=SaveFullText'+'&text='+encodeURIComponent(el.innerHTML)+'&filename='+section_id,
        // data : 'func=SaveFullText'+'&text='+el.innerHTML+'&filename='+section_id
        success: function (e) {
            alert("Saved!");
        },
        error: function (e) {
            alert("Haven't saved!");
        }
    }).done(function(result) {
    });
    
    
}



function preg_quote (str, delimiter) {
  // http://kevin.vanzonneveld.net
  // +   original by: booeyOH
  // +   improved by: Ates Goral (http://magnetiq.com)
  // +   improved by: Kevin van Zonneveld (http://kevin.vanzonneveld.net)
  // +   bugfixed by: Onno Marsman
  // +   improved by: Brett Zamir (http://brett-zamir.me)
  // *     example 1: preg_quote("$40");
  // *     returns 1: '\$40'
  // *     example 2: preg_quote("*RRRING* Hello?");
  // *     returns 2: '\*RRRING\* Hello\?'
  // *     example 3: preg_quote("\\.+*?[^]$(){}=!<>|:");
  // *     returns 3: '\\\.\+\*\?\[\^\]\$\(\)\{\}\=\!\<\>\|\:'
  return (str + '').replace(new RegExp('[.\\\\+*?\\[\\^\\]$(){}=!<>|:\\' + (delimiter || '') + '-]', 'g'), '\\$&');
}

function replaceRegex() {
    saveUndoText();
    
    var startPage = $('#regexPageStart2').val();
    var endPage = $('#regexPageEnd2').val();
    var el = document.getElementById("editable-area");
    var str="" + el.innerHTML;
    
    var regexText=document.getElementById("regexText").value;
    var replaceText=document.getElementById("replaceText").value;
    var str="" + el.innerHTML;
    
    if ( startPage == "" ) {
        alert( "Tagged "+str.match(new RegExp(regexText, "g")).length+" entities!" );
        el.innerHTML = str.replace(new RegExp(regexText, "g"), replaceText);
    } else {
        var regexText2="【<a([^<>]*?)>"+startPage+"</a>】(.*?)【<a([^<>]*?)>"+endPage+"</a>】";
        var partString = ""+str.match(new RegExp(regexText2, "g"));
        alert(partString);
        
        alert( "Tagged "+partString.match(new RegExp(regexText, "g")).length+" entities!" );
        var resultString = partString.replace(new RegExp(regexText, "g"), replaceText);
        
        str="" + el.innerHTML;
        el.innerHTML = str.replace(new RegExp(regexText, "g"), resultString);
    }
    
    //document.styleSheets[0].addRule("tag001", "color:green;")
}


//Tagging Items
function getSelected() {
    if(window.getSelection) {   // all browsers, except IE before version 9
        return window.getSelection();
    } else if (document.getSelection) { 
        return document.getSelection(); 
    } else {    // IE
        var selection = document.selection && document.selection.createRange();
        if(selection.text) {
            return selection.text;
        }

        return false;
    }
    return false;
}

function tagStringWithTag( stringValue, tag ) {
    saveUndoText();
    var el = document.getElementById("editable-area");
    var regexText="("+String(stringValue)+")";
    if ( tag=="person") {
        var replaceText="<br><"+tag+">$1</"+tag+">";
    } else {
        var replaceText="<"+tag+">$1</"+tag+">";
    }
    var str="" + el.innerHTML;

    $('.tagItemDivClass').remove();
    alert( "Tagged "+str.match(new RegExp(regexText, "g")).length+" entities!" );
    el.innerHTML = str.replace(new RegExp(regexText, "g"), replaceText);
}


function tagwithtitle( range, stringSelection ) {
    saveUndoText();
    range.deleteContents();
    //var newdiv = document.createElement("br");
    //range.insertNode(newdiv);
    range.insertNode(document.createTextNode("〈"+stringSelection+"〉"));
    var newdiv = document.createElement("br");
    range.insertNode(newdiv);
    
    $('.tagItemDivClass').remove();
}

function tagNameWithLastName() {
    saveUndoText();
    var el = document.getElementById("editable-area");
    var regexText=/(○|】|^)(王|李|張|趙|劉|陳|楊|吳|黃|黄|朱|孫|郭|胡|呂|高|宋|徐|程|林|鄭|范|何|韓|曹|馬|許|田|馮|杜|周|曾|汪|蘇|董|方|蔡|梁|石|謝|賈|薛|彭|崔|唐|潘|鄧|史|錢|侯|魏|羅|葉|沈|孟|姚|傅|丁|章|蕭|蔣|盧|陸|袁|晁|譚|邵|歐陽|孔|詹|俞|尹|廖|閻|洪|夏|雷|葛|文|柳|陶|毛|丘|龔|蒲|邢|郝|龐|安|裴|折|施|游|金|鄒|湯|虞|嚴|鍾)([^○(舉人|縣人|歲貢|間任)]{1,3}|○[^○])(?=(○|$))/g;
    var replaceText="$1<br><person>$2$3</person>$4";
    var str="" + el.innerHTML;
    
    alert( "Tagged "+str.match(regexText).length+" entities!" );
    el.innerHTML = str.replace(regexText, replaceText);
}

function tagNameWithLastName2() {
    saveUndoText();
    var el = document.getElementById("editable-area");
    var regexText="(○|】|^)("+$('#surname').val()+")([^○(舉人|縣人|歲貢|間任)]{1,3}|○[^○])(?=(○|$))";
    var replaceText="$1<br><person>$2$3</person>$4";
    var str="" + el.innerHTML;
    
    alert( "Tagged "+str.match(new RegExp(regexText, "g")).length+" entities!" );
    el.innerHTML = str.replace(new RegExp(regexText, "g"), replaceText);
}

function tagBiogAddr() {
    saveUndoText();
    var el = document.getElementById("editable-area");
    var str="" + el.innerHTML;

    var regexText=/(○)([^○]{1,6})(○?)(人)/g;
    var replaceText="$1<biog_addr>$2</biog_addr>$3$4";
    el.innerHTML = str.replace(regexText, replaceText);
    
    alert( "Tagged "+str.match(regexText).length+" entities!" );
}

function smartRegexNew() {

    var popup_status = $('#smartRegexPopUpDiv').css("display");
    if (popup_status == "block") {
        $('#smartRegexPopUpDiv').css("display", "none");
    } else {
        $('#smartRegexPopUpDiv').css("display", "block");
    }


    $('#smartRegexPopUpAdd').attr("disabled", false);
    $('#smartRegexPopUpEdit').attr("disabled", "disabled");
    $('#smartRegexPopUpDel').attr("disabled", "disabled");
    $('#smartRegexPopUpBack').attr("disabled", "disabled");
    $('#smartRegexPopUpFor').attr("disabled", "disabled");

}

function replaceSmartClose() {
    $('#smartRegexShowDiv > span').css("border","1px solid black");
    $('#smartRegexPopUpDiv').css("display", "none");
    $("#smartRegexPopUpSelectWord").val("NULL");
    $("#smartRegexPopUpText").val("");
    $("#smartRegexPopUpName").val("");
}

function replaceSmartEdit(){
    thisObject = $('#smartRegexPopUpDiv').attr("editID");

    $('#smartRegexShowDiv > #'+thisObject).attr("class", "span_"+$("#smartRegexPopUpSelectTag").val());
    $('#smartRegexShowDiv > #'+thisObject).attr("regexText", $("#smartRegexPopUpText").val());
    $('#smartRegexShowDiv > #'+thisObject).attr("regexReplace", $("#smartRegexPopUpSelectTag").val());
    $('#smartRegexShowDiv > #'+thisObject).text($("#smartRegexPopUpName").val());  // smartRgextPopUpName.val() is the name of this regex
    // id should be the same as it was
}   

function replaceSmartDel() {
    thisObject = $('#smartRegexPopUpDiv').attr("editID");

    $('#smartRegexShowDiv > #'+thisObject).remove();
}

function replaceSmartFor() {
    thisObject = $('#smartRegexPopUpDiv').attr("editID");
    //var wahaha = $('#'+thisObject).clone();
    //var hahawa = $('#'+thisObject).next();
    //$('#'+thisObject).remove();
    //wahaha.insertAfter(hahawa);
    $('#smartRegexShowDiv > #'+thisObject).insertAfter( $('#'+thisObject).next() );
}

function replaceSmartBack() {
    thisObject = $('#smartRegexPopUpDiv').attr("editID");
    //var wahaha = $('#'+thisObject).clone();
    //var hahawa = $('#'+thisObject).prev();
    //$('#'+thisObject).remove();
    //wahaha.insertBefore(hahawa);
    $('#smartRegexShowDiv > #'+thisObject).insertBefore( $('#'+thisObject).prev() );
}

function replaceSmartAdd() {
    var newdiv = document.createElement("span");
    //newdiv.innerHTML = " "+$("#smartRegexPopUpName").val()+" ";
    //newdiv.innerHTML = $("#smartRegexPopUpName").val();
    $(newdiv).text($("#smartRegexPopUpName").val());

    $(newdiv).css("border", "1px solid black");
    $(newdiv).css("width", "100px");
    
    $(newdiv).attr("class", "span_"+$("#smartRegexPopUpSelectTag").val());



    // id is the name from user's input, which dose not guarantee to be unique
    // id should be unique
    regex_element_index += 1;
    $(newdiv).attr("id", "regex_elem_"+regex_element_index);
    // $(newdiv).attr("id", "span_"+$("#smartRegexPopUpName").val());
    $(newdiv).attr("regexText", $("#smartRegexPopUpText").val());
    $(newdiv).attr("regexReplace", $("#smartRegexPopUpSelectTag").val());

    
    $('#smartRegexShowDiv').append(newdiv);
    
    replaceSmartClose();
}


$(document).on("click", '#smartRegexShowDiv > span', function (e) {
    
    

    $('#smartRegexPopUpDiv').css("display", "block");
    
    $('#smartRegexPopUpDiv').attr("editID", $(this).attr("id"));
    
    $('#smartRegexPopUpName').val($(this).text());
    // $('#smartRegexPopUpName').val($(this).html());

    $('#smartRegexPopUpText').val($(this).attr("regexText"));
    $('#smartRegexPopUpSelectTag').val($(this).attr("regexReplace"));
    
    $('#smartRegexPopUpAdd').attr("disabled", "disabled");
    $('#smartRegexPopUpEdit').attr("disabled", false);
    $('#smartRegexPopUpDel').attr("disabled", false);
    $('#smartRegexPopUpBack').attr("disabled", false);
    $('#smartRegexPopUpFor').attr("disabled", false);
});

function genRegexWindowOpen(){
    var btn_state = $('#regex_generator').css('display');
    if (btn_state == "block") {
        $("#regex_generator").css("display", "none");
        $("#gen_regex_window_open_id").text("Open Gen Regex");
    } else { 
        $('#regex_generator').css("display", "block");
        $("#gen_regex_window_open_id").text("Close Gen Regex");
    }
}
function genRegexWindowClose(){
    $('#regex_generator').css("display", "none");
}

function sharedStart_(array){
    var A= array.concat().sort(), 
    a1= A[0], a2= A[A.length-1], L= a1.length, i= 0;
    while(i<L && a1.charAt(i)=== a2.charAt(i)) i++;
    return a1.substring(0, i);
}

function longestCommonSubstring_(string1, string2){
    // init max value
    var longestCommonSubstring = 0;
    // init 2D array with 0
    var table = [],
        len1 = string1.length,
        len2 = string2.length,
        row, col;
    
    for(row = 0; row <= len1; row++){
        table[row] = [];
        for(col = 0; col <= len2; col++){
            table[row][col] = 0;
        }
    }
    // fill table
    var i, j;
    for(i = 0; i < len1; i++){
        for(j = 0; j < len2; j++){
            if(string1[i]==string2[j]){
                if(table[i][j] == 0){
                    table[i+1][j+1] = 1;
                } else {
                    table[i+1][j+1] = table[i][j] + 1;
                }
                if(table[i+1][j+1] > longestCommonSubstring){
                    longestCommonSubstring = table[i+1][j+1];
                }
            } else {
                table[i+1][j+1] = 0;
            }
        }
    }
    return longestCommonSubstring;
}

function longestCommonSubstring(s1, s2) {
   
    var start_idx = 0;
    var max_len = 0;
    for (var i = 0; i < s1.length; i++)
    {
        for (var j = 0; j < s2.length; j++)
        {
            var x = 0;
            while (s1.charAt(i + x) == s2.charAt(j + x))
            {
                x++;
                if ((i + x) >= s1.length || ((j + x) >= s2.length))
                    break;
            }
            if (x > max_len)
            {
                max_len = x;
                start_idx = i;
            }
         }
    }
    return s1.substring(start_idx, (start_idx + max_len));

}

function getRegex(_pattern) {
    console.log(_pattern[0]);
    console.log(_pattern[1]);
    var p0 = _pattern[0];
    var p1 = _pattern[1];

    // TODO: find common pattern
    var reg_str = "";
    // _p1 = 測試
    // _p2 = 測<tag_name>試</tag_name>一下
    var combined = [];
    if (p0.length > p1.length) {
        combined = p0;
    } else if(p0.length < p1.length) {
        combined = p1;
    } else {    // equal length
        // find matching string
        var cnt = p0.length;
        for (var i = 0;  i < cnt; i++) {
            if (p1[i].tag != null) {
                combined.push({tag:p1[i].tag, txt:"[^○如即而之有<>〈〉【】]{1,"+p1[i].txt.length+"}"});
            } else if (p0[i].tag != null) {
                combined.push({tag:p0[i].tag, txt:"[^○如即而之有<>〈〉【】]{1,"+p0[i].txt.length+"}"});
            } else {
                // find matching for text in each corresponding position
                var texts = [p0[i].txt, p1[i].txt];
                var common = longestCommonSubstring(p0[i].txt, p1[i].txt);
                
                /*
                var reg_for_common = "[";
                for (var i = 0; i < common.length; i++) {
                    common[i];
                    reg_for_common += common[i]+"|";
                };
                reg_for_common += "]";
                */
                combined.push({tag:null, txt:common});
            }
        };
    }

    for (var i = 0; i < combined.length; i++) {
       reg_str += combined[i].txt;
    };

    return reg_str;
}

var pattern_obj = [];  // record pattern array for regex generator. only contain pattern1 and pattern2

function genRegexBySelection(tag_item_div, _selection) {
    var add_gen_regex_button = document.createElement("button");
    $(add_gen_regex_button).id = "addToGenRegex";
    $(add_gen_regex_button).addClass("btn btn-md");
    $(add_gen_regex_button).click( function(){
        // popup for selected words regex gen
        console.log("Debug: ");
        console.log(_selection);

        if (_selection.type == "Range") {
            // select words, not just click on text
            var anchor_node = _selection.anchorNode;
            var focus_node = _selection.focusNode;
            var sibling_node = anchor_node.nextElementSibling;
            

            if (anchor_node && sibling_node && focus_node && container.innerHTML.indexOf( "br" ) == -1) {
                // Chrome can work on this.
                // Safari does not support some of the member in selection object
                // container.innerHTML.indexOf( "br" ) == -1: selection does not contain br.

                var seleted_div = document.createElement("div");
                var seleted_obj = [];   // array for selected text as well as its tag if it has any

                if (anchor_node == focus_node ) { 
                    // selected text in plain text
                    var text_all = anchor_node.textContent;
                    var text_ = text_all.substring(_selection.anchorOffset, _selection.focusOffset);
                    $(seleted_div).text(text_);
                    seleted_obj.push({tag:null, txt:text_});    // push object into array

                } else {
                    // selected text contain tags
                    var text_before = anchor_node.textContent.substring(_selection.anchorOffset, anchor_node.length);
                    var tag_name = sibling_node.nodeName.toLowerCase();
                    var tagged_text = sibling_node.textContent;

                    var text_after = _selection.focusNode.textContent.substring(0, _selection.focusOffset);

                    $(seleted_div).text(text_before+tagged_text+text_after);
                    seleted_obj.push({tag:null, txt:text_before});
                    seleted_obj.push({tag:tag_name, txt:tagged_text});
                    seleted_obj.push({tag:null, txt:text_after});


                    console.log(text_before);
                    console.log(tag_name);
                    console.log(tagged_text);
                    console.log(text_after);
                }
                
                
                var generated_regex = "";
                // show generate regex window
                $('#regex_generator').css("display", "block");
                $("#gen_regex_window_open_id").text("Close Gen Regex");

                //var seleted_text = String(_selection).replace(/^\s+|\s+$/g,'');
                var pattern1 = $('#regex_pattern1');
                var pattern2 = $('#regex_pattern2');
                if (pattern1.children().length == 0) {
                    pattern1.append(seleted_div);
                    pattern_obj.push(seleted_obj);
                    // pattern1.text(seleted_div.text());
                } else if (pattern2.children().length == 0) {
                    pattern2.append(seleted_div);
                    pattern_obj.push(seleted_obj);
                    //pattern2.text(seleted_div.text());
                    generated_regex = getRegex(pattern_obj);

                } else {
                    // pattern1 and pattern2 are already having text
                    pattern1.children().remove();
                    pattern1.append(pattern2.children());

                    pattern2.children().remove();
                    pattern2.append(seleted_div);

                    pattern_obj.shift();
                    pattern_obj.push(seleted_obj);

                    //pattern1.text(pattern2.text());
                    //pattern2.text(seleted_div);
                    generated_regex = getRegex(pattern_obj);

                }
                $('#generated_regex').text(generated_regex);
                // ---

                $('#regex_generator_error_msg').text("");
            } else {
                $('#regex_generator_error_msg').text("Note: Not a valid selection for regex generator.");
            }

            $('.tagItemDivClass').remove(); // close the tag window

        } else if (_selection.type == "Caret") {
            // TODO: click on tagged text case rather than select
            // If do this process, also need to consider between browers since not all of them support
            // and also need to modify pop_remove_tag_window
        }       
    });

    $(add_gen_regex_button).text("Add to Gen Regex");
    tag_item_div.appendChild(add_gen_regex_button);    
}


function smartRegexEmpty() {
    $('#smartRegexShowDiv').html("");
    regex_element_index = 0;
}

function replaceSmartRunWithBr() {

    var replaceSmartRegexString = "";
    var replaceSmartReplaceString = "";
    var count=1;
    
    saveUndoText();
    
    // skip everything inside "【】", including "【】". // (【([^【】])*】) //(【.*】)
    replaceSmartRegexString += "(【[^【】]+(?!.*})】)*";
    replaceSmartReplaceString += "$"+count;
    count ++;
    // ---

    $('#smartRegexShowDiv').children('span').each(function () {
        replaceSmartRegexString += "(" + $(this).attr("regexText") + ")";
        if ( $(this).attr("regexReplace") == "notag" || $(this).attr("regexReplace") == "NOTAG") {
            replaceSmartReplaceString += "$" + count;
        } else if ( $(this).attr("regexReplace") == "title" ) {
            replaceSmartReplaceString += "<br>〈" + "$" + count + "〉<br>";
        } else {
            // with <br> before
            replaceSmartReplaceString += "<br><" + $(this).attr("regexReplace") + ">" + "$" + count + "</"+ $(this).attr("regexReplace") +">";
        }
        count++;
    });
        
        
    var startPage = $('#regexPageStart').val();
    var endPage = $('#regexPageEnd').val();
    var el = document.getElementById("editable-area");
    var str="" + el.innerHTML;
    
    if ( startPage == "" ) {
        alert( "Tagged "+str.match(new RegExp(replaceSmartRegexString, "g")).length+" entities!" );
        el.innerHTML = str.replace(new RegExp(replaceSmartRegexString, "g"), replaceSmartReplaceString);
    } else {
        var regexText="【<a([^<>]*?)>"+startPage+"</a>】(.*?)【<a([^<>]*?)>"+endPage+"</a>】";
        var partString = ""+str.match(new RegExp(regexText, "g"));
        alert(partString);
        
        alert( "Tagged "+partString.match(new RegExp(replaceSmartRegexString, "g")).length+" entities!" );
        var resultString = partString.replace(new RegExp(replaceSmartRegexString, "g"), replaceSmartReplaceString);
        
        str="" + el.innerHTML;
        el.innerHTML = str.replace(new RegExp(regexText, "g"), resultString);
    }
}

function replaceSmartRun() {

    var replaceSmartRegexString = "";
    var replaceSmartReplaceString = "";
    var count=1;
    
    saveUndoText();
    
    // skip everything inside "【】", including "【】".
    replaceSmartRegexString += "(【[^【】]+(?!.*})】)*";
    replaceSmartReplaceString += "$"+count;
    count ++;
    // ---
 
    $('#smartRegexShowDiv').children('span').each(function () {
        replaceSmartRegexString += "(" + $(this).attr("regexText") + ")";
        if ( $(this).attr("regexReplace") == "notag" || $(this).attr("regexReplace") == "NOTAG" ) {
            replaceSmartReplaceString += "$" + count;
        } else if ( $(this).attr("regexReplace") == "title" ) {
            replaceSmartReplaceString += "<br>〈" + "$" + count + "〉<br>";
        } else {
            replaceSmartReplaceString += "<" + $(this).attr("regexReplace") + ">" + "$" + count + "</"+ $(this).attr("regexReplace") +">";
        }
        count++;
    });
        
        
    var startPage = $('#regexPageStart').val();
    var endPage = $('#regexPageEnd').val();
    var el = document.getElementById("editable-area");
    var str="" + el.innerHTML;
    
    if ( startPage == "" ) {
        alert( "Tagged "+str.match(new RegExp(replaceSmartRegexString, "g")).length+" entities!" );
        el.innerHTML = str.replace(new RegExp(replaceSmartRegexString, "g"), replaceSmartReplaceString);
    } else {
        var regexText="【<a([^<>]*?)>"+startPage+"</a>】(.*?)【<a([^<>]*?)>"+endPage+"</a>】";
        var partString = ""+str.match(new RegExp(regexText, "g"));
        alert(partString);
        
        alert( "Tagged "+partString.match(new RegExp(replaceSmartRegexString, "g")).length+" entities!" );
        var resultString = partString.replace(new RegExp(replaceSmartRegexString, "g"), replaceSmartReplaceString);
        
        str="" + el.innerHTML;
        el.innerHTML = str.replace(new RegExp(regexText, "g"), resultString);
    }
}

function replaceSmartRunSpace() {

    var replaceSmartRegexString = "";
    var replaceSmartReplaceString = "";
    var count=1;
    saveUndoText();
    
    $('#smartRegexShowDiv').children('span').each(function () {
        //alert($(this).attr("regexText"));
        replaceSmartRegexString += "(" + $(this).attr("regexText") + ")(○*)";
        if ( $(this).attr("regexReplace") == "notag" || $(this).attr("regexReplace") == "NOTAG") {
            replaceSmartReplaceString += "$" + count;
        } else if ( $(this).attr("regexReplace") == "title" ) {
            replaceSmartReplaceString += "<br>〈" + "$" + count + "〉<br>";
        } else {
            replaceSmartReplaceString += "<" + $(this).attr("regexReplace") + ">" + "$" + count + "</"+ $(this).attr("regexReplace") +">";
        }
        count++;
        replaceSmartReplaceString += "$" + count;
        count++;
    });
    
    var startPage = $('#regexPageStart').val();
    var endPage = $('#regexPageEnd').val();
    var el = document.getElementById("editable-area");
    var str="" + el.innerHTML;
    
    if ( startPage == "" ) {
        alert( "Tagged "+str.match(new RegExp(replaceSmartRegexString, "g")).length+" entities!" );
        el.innerHTML = str.replace(new RegExp(replaceSmartRegexString, "g"), replaceSmartReplaceString);
    } else {
        var regexText="【<a([^<>]*?)>"+startPage+"</a>】(.*?)【<a([^<>]*?)>"+endPage+"</a>】";
        var partString = ""+str.match(new RegExp(regexText, "g"));
        alert(partString);
        
        alert( "Tagged "+partString.match(new RegExp(replaceSmartRegexString, "g")).length+" entities!" );
        var resultString = partString.replace(new RegExp(replaceSmartRegexString, "g"), replaceSmartReplaceString);
        
        str="" + el.innerHTML;
        el.innerHTML = str.replace(new RegExp(regexText, "g"), resultString);
    }
    
    //alert( "Tagged "+str.match(new RegExp(replaceSmartRegexString, "g")).length+" entities!" );
    //el.innerHTML = str.replace(new RegExp(replaceSmartRegexString, "g"), replaceSmartReplaceString);
}

function smartRegexSave(topic_id) {
    console.log("topic_id"+topic_id);

    var today = new Date();
    var minute = today.getMinutes();
    var hour = today.getHours();
    var dd = today.getDate();
    var mm = today.getMonth()+1; //January is 0!
    var yyyy = today.getFullYear();

    if(dd<10) {
        dd='0'+dd
    } 
    if(mm<10) {
        mm='0'+mm
    } 
    if (hour<10) {
        hour='0'+hour;
    }
    if (minute<10) {
        minute='0'+minute;
    }
    today = hour+'_'+minute+'_'+dd+'_'+mm+'_'+yyyy;

    var name=prompt("Please enter this Regex name", RegexLoadedName+"_"+today);

    if (name!=null && name != ''){
        $.ajax({
            type : 'POST',
            url : './TaggingText',
            async : false,
            data : 'func=SmartRegexSave'+'&text='+encodeURIComponent($('#smartRegexShowDiv').html())+'&filename='+name+'&topic_id='+topic_id,
            error: function (result) {
                alert("Error");
            },
            success: function (result) {
                var obj = jQuery.parseJSON(result);

                if (obj == "ErrorDB") {
                    alert("Error when saving to database!!");

                } else if (obj == "ForceSave") {
                    var retVal = confirm("Danger! You will over write the previous regex file. Do you want to proceed?");
                    
                    if( retVal == true ){
                        // update regex file
                        $.ajax({
                            type : 'POST',
                            url : './TaggingText',
                            async : false,
                            data : 'func=SmartRegexSave'+'&text='+encodeURIComponent($('#smartRegexShowDiv').html())+'&filename='+name+'&topic_id='+topic_id+'&forcesave=1',
                            error: function (result) {
                                alert("Error");
                            },
                            success: function (result) {
                                alert("Update regex file.");
                            }
                        });
                    }else{
                        alert("You have not saved the regex file.");
                    }

                } else {
                    alert("Saved!");
                }
            }
        }).done(function(result) {
        });
    } else {
        alert("You haven't saved it.");
    }
}

function smartRegexLoad(topic_id) {
    $('#load_regex_div').html("");
    var popup_status = $('#load_regex_div').css("display");
    if (popup_status == "block") {
        $('#load_regex_div').css("display", "none");
    } else {
        $('#load_regex_div').css("display", "block");
    }

    var newselect = document.createElement("select");
    newselect.id = "loadRegexSelect";
            
    $.ajax({
        type: 'POST', 
        url: './TaggingText',
        dataType: 'json',
        data: "func=SmartRegexLoad&topic_id="+topic_id,
        //cache: false,
        success: function (data) {
            $.each(data, function(index, element) {
                // index is the filename (without '.txt')
                // element is the content in the file
                newselect.innerHTML += "<option value=\""+index+"\">"+index+"</option>\n";
                //alert(index);
                //alert(element);
                var newdiv = document.createElement("div");
                $(newdiv).css("display", "none");
                $(newdiv).html(element);
                
                $(newdiv).attr("id", "div_"+index);
                $('#load_regex_div').append(newdiv);
            });
            
            
        },
        error: function (data) {
            console.log("SmartRegexLoad fails");
        }

    });

    /*
    $.ajax({
        type: 'POST', 
        url: '../models/_extractapp_func.php',
        dataType: 'json',
        data: "func=SmartRegexLoad",
        //cache: false,
        success: function (data) {
            $.each(data, function(index, element) {
                newselect.innerHTML += "<option value=\""+index+"\">"+index+"</option>\n";
                //alert(index);
                //alert(element);
                var newdiv = document.createElement("div");
                $(newdiv).css("display", "none");
                $(newdiv).html(element);
                $(newdiv).attr("id", "div_"+index);
                $('#load_regex_div').append(newdiv);
            });
            
        },
        error: function (data) {
            console.log("SmartRegexLoad fails");
        }

    });
    */
    
    $('#load_regex_div').append(newselect);
    var newbutton = document.createElement("button");
    $(newbutton).html("Load");
    $(newbutton).addClass("btn btn-info");
    $(newbutton).attr("onclick", "loadRegexAdd()");
    $('#load_regex_div').append(newbutton);
    var newbutton = document.createElement("button");
    $(newbutton).html("Close");
    $(newbutton).addClass("btn btn-default");
    $(newbutton).attr("onclick", "$('#load_regex_div').css(\"display\", \"none\");");
    $('#load_regex_div').append(newbutton);
}

var RegexLoadedName = "";
function loadRegexAdd() {
    RegexLoadedName = $('#loadRegexSelect').val();
    var divName = "#div_"+RegexLoadedName;
    var regex_content = $(divName).html();

    $('#smartRegexShowDiv').html(regex_content);
    $('#load_regex_div').css("display", "none");

    // get the largest regex element index in the regex file
    // for all children (span) in #smartRegexShowDiv, find MAX(id)
    var regex_elem = $('#smartRegexShowDiv').children();   
    var max_id = 0;
    for (var i = 0; i < regex_elem.length; i++) {
        var r_id = parseInt(regex_elem[i].id.slice(11));    // cut the first 11 char out: "regex_elem_"
        if (r_id > max_id) {
            max_id = r_id;
        }

    }
    regex_element_index = max_id;


    // hover on #smartRegexShowDiv > span, change border width
    $("#smartRegexShowDiv > span").hover( function() {
        // hover in
        $(this).css("border","3px solid black");
        }, function() {
        // hover out
        // TODO: if this is clicked...
        $(this).css("border","1px solid black");
    });
    
    /*
    // TODO: click
    $("#smartRegexShowDiv > span").click( function() {
        $(this).css("border","3px double black");
        console.log("Debug: clicked");
    });


    $('#smartRegexShowDiv > span').on({
        mouseover: function(){
            $(this).css("border","3px solid black");
        },
        mouseleave: function(){
            $(this).css("border","1px solid black");
        },
        click: function(){
            $(this).off('mouseleave');
        }
    });
    */
}



// ===


// === for editwordlist.php ===
function addNewList() {
    var el = document.getElementById("listNameText");
    $.ajax({
        //url : '../../models/_extractapp_func.php',
        url : './EditWordlist',
        async : false,
        type : 'POST',
        data : 'func=AddNewList'+'&text='+el.value,
        success: function (e) {
            alert("Added!");
            document.location.reload(true);
        },
        error: function (e) {
            console.log("error when add new list");
            alert("Haven't added new list!!");
        }
    }).done(function(result) {
        
    });
}

function saveWordlist(id ) {
    var el = document.getElementById("editable-area");
    $.ajax({
        url : './EditWordlist',
        async : false,
        type : 'POST',
        data : 'func=SaveWordlist'+'&text='+el.innerHTML+'&filename='+id,
        success: function (e) {
            alert("Saved!");
        },
        error: function (e) {
            alert("Haven't saved!");
        }
    }).done(function(result) {
        
    });
    
}

function editWordlistText( id ) {
    var el = document.getElementById("editable-area");
    el.contentEditable = true;
}

function replaceRegex() {
    var el = document.getElementById("editable-area");
    var regexText=document.getElementById("regexText").value;
    var replaceText=document.getElementById("replaceText").value;
    var str="" + el.innerHTML;
    el.innerHTML = str.replace(new RegExp(regexText, "g"), replaceText);
}

function showListContent( id ) {
    var xhr = new XMLHttpRequest();
    xhr.onreadystatechange = process;
    xhr.open("POST", "../data/wordlist/"+id+".txt?t=" + Math.random(), true);
    // TODO: should show the latest wordlist file. eg. id_timestamp.txt, rather than open the original id.txt file 


    xhr.send();

    function process() {
        if (xhr.readyState == 4) {
            var el = document.getElementById("editable-area");
            var str = xhr.responseText
            var regexText=/\n/g;
            var replaceText="<br>\n";
            el.innerHTML = str.replace(regexText, replaceText);
            
            document.getElementById("button-area").innerHTML=
            "<form action=\"javascript:void(0);\"> \
                <fieldset><legend>Edit:</legend> \
                    <button id=\"buttonEditText\" onclick=\"editWordlistText("+id+")\" style=\"height: 30px; width: 220px\">Edit the text</button></br> \
                    <button id=\"buttonSaveText\" onclick=\"saveWordlist("+id+")\" style=\"height: 30px; width: 220px\">Save the text</button> \
                </fieldset> \
                <fieldset><legend>Replace By Regex:</legend> \
                    Regex: <input type=\"text\" size=\"30\" id=\"regexText\"></br> \
                    Replace: <input type=\"text\" size=\"30\" id=\"replaceText\"><br> \
                    <button onclick=\"replaceRegex()\">Replace!</button> \
                </fieldset> \
            </form>";
        }
    }
}

// ============

// === for edittaglist.php ===
function editTaglist(topic_id) {
    var form = document.createElement("form");
    form.setAttribute("method", "post");
    form.setAttribute("action", "./EditTaglist");
    form.setAttribute("target", "_blank");
    
    var hiddenField = document.createElement("input");      
    hiddenField.setAttribute("name", "topic_id");
    hiddenField.setAttribute("value", topic_id);
    form.appendChild(hiddenField);
    
    if(navigator.userAgent.toLowerCase().indexOf('firefox') > -1) {
        document.body.appendChild(form);
        form.submit();
    } else {
        form.submit(); // works under IE and Chrome, but not FF  
    }
}


// ============
//