2 * @class Roo.htmleditor.FilterWord
3 * try and clean up all the mess that Word generates.
5 * This is the 'nice version' - see 'Heavy' that white lists a very short list of elements, and multi-filters
8 * Run a new Span Filter
9 * @param {Object} config Configuration options
12 Roo.htmleditor.FilterWord = function(cfg)
14 // no need to apply config.
15 this.replaceDocBullets(cfg.node);
17 // this is disabled as the removal is done by other filters;
18 // this.walk(cfg.node);
23 Roo.extend(Roo.htmleditor.FilterWord, Roo.htmleditor.Filter,
29 * Clean up MS wordisms...
31 replaceTag : function(node)
34 // no idea what this does - span with text, replaceds with just text.
36 node.nodeName == 'SPAN' &&
37 !node.hasAttributes() &&
38 node.childNodes.length == 1 &&
39 node.firstChild.nodeName == "#text"
41 var textNode = node.firstChild;
42 node.removeChild(textNode);
43 if (node.getAttribute('lang') != 'zh-CN') { // do not space pad on chinese characters..
44 node.parentNode.insertBefore(node.ownerDocument.createTextNode(" "), node);
46 node.parentNode.insertBefore(textNode, node);
47 if (node.getAttribute('lang') != 'zh-CN') { // do not space pad on chinese characters..
48 node.parentNode.insertBefore(node.ownerDocument.createTextNode(" ") , node);
51 node.parentNode.removeChild(node);
52 return false; // dont do chidren - we have remove our node - so no need to do chdhilren?
57 if (node.tagName.toLowerCase().match(/^(style|script|applet|embed|noframes|noscript)$/)) {
58 node.parentNode.removeChild(node);
59 return false; // dont do chidlren
61 //Roo.log(node.tagName);
62 // remove - but keep children..
63 if (node.tagName.toLowerCase().match(/^(meta|link|\\?xml:|st1:|o:|v:|font)/)) {
64 //Roo.log('-- removed');
65 while (node.childNodes.length) {
66 var cn = node.childNodes[0];
68 node.parentNode.insertBefore(cn, node);
69 // move node to parent - and clean it..
70 if (cn.nodeType == 1) {
75 node.parentNode.removeChild(node);
76 /// no need to iterate chidlren = it's got none..
77 //this.iterateChildren(node, this.cleanWord);
78 return false; // no need to iterate children.
81 if (node.className.length) {
83 var cn = node.className.split(/\W+/);
85 Roo.each(cn, function(cls) {
86 if (cls.match(/Mso[a-zA-Z]+/)) {
91 node.className = cna.length ? cna.join(' ') : '';
93 node.removeAttribute("class");
97 if (node.hasAttribute("lang")) {
98 node.removeAttribute("lang");
101 if (node.hasAttribute("style")) {
103 var styles = node.getAttribute("style").split(";");
105 Roo.each(styles, function(s) {
109 var kv = s.split(":");
110 if (kv[0].match(/^(mso-|line|font|background|margin|padding|color)/)) {
113 // what ever is left... we allow.
116 node.setAttribute("style", nstyle.length ? nstyle.join(';') : '');
117 if (!nstyle.length) {
118 node.removeAttribute('style');
121 return true; // do children
127 styleToObject: function(node)
129 var styles = (node.getAttribute("style") || '').split(";");
131 Roo.each(styles, function(s) {
135 var kv = s.split(":");
137 // what ever is left... we allow.
138 ret[kv[0].trim()] = kv[1];
144 replaceDocBullets : function(doc)
146 // this is a bit odd - but it appears some indents use ql-indent-1
147 Roo.log(doc.innerHTML);
149 var listpara = doc.getElementsByClassName('MsoListParagraphCxSpFirst');
150 for( var i = 0; i < listpara.length; i ++) {
151 listpara.item(i).className = "MsoListParagraph";
153 // this is a bit hacky - we had one word document where h2 had a miso-list attribute.
154 var htwo = doc.getElementsByTagName('h2');
155 for( var i = 0; i < htwo.length; i ++) {
156 if (htwo.item(i).hasAttribute('style') && htwo.item(i).getAttribute('style').match(/mso-list:/)) {
157 htwo.item(i).className = "MsoListParagraph";
160 listpara = doc.getElementsByClassName('MsoNormal');
161 while(listpara.length) {
162 if (listpara.item(i).hasAttribute('style') && listpara.item(i).getAttribute('style').match(/mso-list:/)) {
163 listpara.item(i).className = "MsoListParagraph";
165 listpara.item(i).className = "MsoNormalx";
168 listpara = doc.getElementsByClassName('ql-indent-1');
169 while(listpara.length) {
170 this.replaceDocBullet(listpara.item(0));
172 listpara = doc.getElementsByClassName('MsoListParagraph');
173 while(listpara.length) {
175 this.replaceDocBullet(listpara.item(0));
182 replaceDocBullet : function(p)
184 // gather all the siblings.
186 parent = p.parentNode,
187 doc = parent.ownerDocument,
192 if (ns.nodeType != 1) {
196 if (!ns.className.match(/(MsoListParagraph|ql-indent-1)/i)) {
199 if (ns.getAttribute('style').match(/mso-list/)) {
205 var spans = ns.getElementsByTagName('span');
209 var has_list = false;
210 for(var i = 0; i < spans.length; i++) {
211 if (spans[i].getAttribute('style').match(/mso-list/)) {
229 var ul = parent.ownerDocument.createElement('ul'); // what about number lists...
230 parent.insertBefore(ul, p);
236 items.forEach(function(n, ipos) {
237 //Roo.log("got innertHMLT=" + n.innerHTML);
239 var spans = n.getElementsByTagName('span');
241 //Roo.log("No spans found");
243 parent.removeChild(n);
246 return; // skip it...
252 for(var i = 0; i < spans.length; i++) {
254 style = this.styleToObject(spans[i]);
255 if (typeof(style['mso-list']) == 'undefined') {
259 spans[i].parentNode.removeChild(spans[i]); // remove the fake bullet.
262 //Roo.log("NOW GOT innertHMLT=" + n.innerHTML);
263 style = this.styleToObject(n); // mo-list is from the parent node.
264 if (typeof(style['mso-list']) == 'undefined') {
265 //Roo.log("parent is missing level");
268 parent.removeChild(n);
273 var nlvl = (style['mso-list'].split(' ')[1].replace(/level/,'') *1) - 1 ;
278 var nul = doc.createElement('ul'); // what about number lists...
280 last_li = doc.createElement('li');
281 stack[lvl].appendChild(last_li);
283 last_li.appendChild(nul);
289 var nli = stack[nlvl].appendChild(doc.createElement('li'));
291 nli.innerHTML = n.innerHTML;
292 //Roo.log("innerHTML = " + n.innerHTML);
293 parent.removeChild(n);