[FIX] website: seo, auto detection of keywords with accents in boundaries

Javascript regex \b is not unicode aware,
and words beginning or ending by accents won't match \b

We therefore use a custom regex to replace \b.
Basically, it's a regex matching all words separators
such as dot, comma, spaces, etc. and other unicode
separators as well.

opw-641005

addons/website/static/src/js/website.seo.js

@@ -6,6 +6,10 @@
 
     website.seo = {};
 
+    // This replaces \b, because accents(e.g. à, é) are not seen as word boundaries.
+    // Javascript \b is not unicode aware, and words beginning or ending by accents won't match \b
+    var WORD_SEPARATORS_REGEX = '([\\u2000-\\u206F\\u2E00-\\u2E7F\'!"#\\$%&\\(\\)\\*\\+,\\-\\.\\/:;<=>\\?¿¡@\\[\\]\\^_`\\{\\|\\}~\\s]+|^|$)';
+
     function analyzeKeyword(htmlPage, keyword) {
         return  htmlPage.isInTitle(keyword) ? {
                     title: 'label label-primary',
@@ -293,13 +297,13 @@
             return $('body').children().not('.js_seo_configuration').text();
         },
         isInBody: function (text) {
-            return new RegExp("\\b"+text+"\\b", "gi").test(this.bodyText());
+            return new RegExp(WORD_SEPARATORS_REGEX+text+WORD_SEPARATORS_REGEX, "gi").test(this.bodyText());
         },
         isInTitle: function (text) {
-            return new RegExp("\\b"+text+"\\b", "gi").test(this.title());
+            return new RegExp(WORD_SEPARATORS_REGEX+text+WORD_SEPARATORS_REGEX, "gi").test(this.title());
         },
         isInDescription: function (text) {
-            return new RegExp("\\b"+text+"\\b", "gi").test(this.description());
+            return new RegExp(WORD_SEPARATORS_REGEX+text+WORD_SEPARATORS_REGEX, "gi").test(this.description());
         },
     });