fix(main): some bugs with new type of HTML-content

voxpelli · Sep 7, 2015 · 9d131e0 · 9d131e0
1 parent 8c33f35
commit 9d131e0
Show file tree

Hide file tree

Showing 2 changed files with 34 additions and 9 deletions.
diff --git a/index.js b/index.js
@@ -17,6 +17,22 @@ var kebabRegexp = /[^a-z0-9]+/g;
 var whitespaceRegexp = /\s+/g;
 var httpRegexp = /^http(s?):\/\//;
 
+var getMfValue = function (data) {
+  return _(data || [])
+    .map(function (item) {
+      if (!item) { return; }
+      if (item.value) { return item.value; }
+      if (item.html) {
+        return ent.decode(item.html.replace(htmlRegexp, ' '))
+          .replace(whitespaceRegexp, ' ')
+          .trim();
+      }
+      if (_.isString(item)) { return item; }
+    })
+    .filter()
+    .value();
+};
+
 var semiKebabCase = function (name) {
   // Convert camel case to spaces, then ensure everything is lower case and then finally – make kebab
   return _.deburr(name)
@@ -122,9 +138,8 @@ Formatter.prototype._formatSlug = function (data) {
   if (data.properties.name) {
     name = data.properties.name[0].trim();
   }
-  if (!name && data.properties.content && this.contentSlug) {
-    name = data.properties.content[0].trim();
-    name = ent.decode(name.replace(htmlRegexp, ''));
+  if (!name && !_.isEmpty(data.properties.content) && this.contentSlug) {
+    name = getMfValue(data.properties.content).join('\n');
   }
 
   if (name) {
@@ -197,10 +212,7 @@ Formatter.prototype.preFormat = function (data) {
 
   var strippedContent, estimatedLang;
   if (_.isEmpty(data.properties.lang) && this.deriveLanguages && !_.isEmpty(data.properties.content)) {
-    strippedContent = ((data.properties.content || [])[0] || '')
-      .replace(htmlRegexp, ' ')
-      .replace(whitespaceRegexp, ' ')
-      .trim();
+    strippedContent = getMfValue(data.properties.content).join('\n');
 
     if (strippedContent !== '') {
       estimatedLang = franc(strippedContent, this.deriveLanguages === true ? {} : { whitelist : this.deriveLanguages });

diff --git a/test/formatter.spec.js b/test/formatter.spec.js
@@ -281,7 +281,7 @@ describe('Formatter', function () {
 
     it('should ignore html-tags when basing slug on content', function () {
       delete baseMicroformatData.properties.name;
-      baseMicroformatData.properties.content = ['<h1>Foo</h1> Bar &amp; <strong>Abc</strong>'];
+      baseMicroformatData.properties.content = [{html: '<h1>Foo</h1> Bar &amp; <strong>Abc</strong>'}];
       formatter = new Formatter({ contentSlug: true });
       // Test twice so that we don't get a non-reusable regexp!
       formatter._formatSlug(baseMicroformatData).should.equal('foo-bar-abc');
@@ -511,7 +511,7 @@ describe('Formatter', function () {
         .that.deep.equals(['de']);
     });
 
-    it('should hanbdle undetectable language', function () {
+    it('should handle undetectable language', function () {
       baseMicroformatData.properties.content = ['Nope'];
 
       formatter = new Formatter({
@@ -545,6 +545,19 @@ describe('Formatter', function () {
         .that.deep.equals(['123']);
     });
 
+    it('should derive english language from HTML content', function () {
+      baseMicroformatData.properties.content = [{html: 'Another evening at the cottage and Bob was feeling brave. Make some food by himself? Surely, how hard could it be. But as the night went on and the dinner cooked Bob grew ever more desperate. Pepper, salt – all these crazy names of things he had never heard before. It would be a long night for poor Mr Bob.'}];
+
+      formatter = new Formatter({
+        deriveLanguages: ['eng', 'swe'],
+      });
+
+      return formatter.preFormat(baseMicroformatData)
+        .should.eventually
+        .have.deep.property('properties.lang')
+        .that.deep.equals(['en']);
+    });
+
   });
 
   describe('_formatFilesSlug', function () {