Develop to master

ckanext/s3filestore/tests/test_uploader.py

@@ -190,23 +190,24 @@ def _test_dataset(self, private=False, title='Test Dataset', author='test'):
             author=author,
             owner_org=self.organisation['id'])
 
-    def _upload_test_resource(self, dataset=None):
+    def _upload_test_resource(self, dataset=None, filename='data.csv'):
         ''' Creates a test resource in the specified dataset
         by uploading a file.
         '''
         if not dataset:
             dataset = self._test_dataset()
-        file_path = os.path.join(os.path.dirname(__file__), 'data.csv')
+        file_path = os.path.join(os.path.dirname(__file__), filename)
         return helpers.call_action(
             'resource_create',
             package_id=dataset['id'],
             upload=FlaskFileStorage(io.open(file_path, 'rb')),
-            url='data.csv')
+            url=filename)
 
     def test_resource_upload(self):
         '''Test a basic resource file upload'''
         file_path = os.path.join(os.path.dirname(__file__), 'data.csv')
         resource = self._upload_test_resource()
+        assert_equal(resource['mimetype'], 'text/csv')
 
         key = _get_object_key(resource)
 
@@ -499,6 +500,11 @@ def test_ignoring_non_uploads(self):
                 mock_upload_to_key.assert_not_called()
                 mock_update_visibility.assert_called_once_with(resource['id'])
 
+    def test_detect_office_document_type(self):
+        dataset = self._test_dataset()
+        resource = self._upload_test_resource(dataset, 'example.docx')
+        assert_equal(resource['mimetype'], 'application/vnd.openxmlformats-officedocument.wordprocessingml.document')
+
     if toolkit.check_ckan_version(max_version='2.8.99'):
 
         def test_resource_upload_then_clear(self):

ckanext/s3filestore/uploader.py

@@ -542,8 +542,18 @@ def __init__(self, resource):
             self.mimetype = resource.get('mimetype')
             if not self.mimetype:
                 try:
-                    # 512 bytes should be enough for a mimetype check
-                    self.mimetype = resource['mimetype'] = mime.from_buffer(self.upload_file.read(512))
+                    try:
+                        # Get type from file extension if we recognise it
+                        self.mimetype = mimetypes.guess_type(self.filename, strict=False)[0]
+                    except Exception:
+                        pass
+                    if not self.mimetype:
+                        try:
+                            # 2048 bytes are needed to detect Office docs
+                            self.mimetype = mime.from_buffer(self.upload_file.read(2048))
+                        except Exception:
+                            pass
+                    resource['mimetype'] = self.mimetype
 
                     # additional check on text/plain mimetypes for
                     # more reliable result, if None continue with text/plain