diff --git a/newspaper/__init__.py b/newspaper/__init__.py
index eaef8855..b49c2486 100644
--- a/newspaper/__init__.py
+++ b/newspaper/__init__.py
@@ -12,6 +12,9 @@
                   popular_urls, NewsPool, Configuration as Config)
 from .source import Source
 from .version import __version__
+from .extractors import ContentExtractor
+from .parsers import Parser
+from .network import get_html
 
 news_pool = NewsPool()
 
diff --git a/newspaper/article.py b/newspaper/article.py
index 5fe75d88..a92cd8bd 100644
--- a/newspaper/article.py
+++ b/newspaper/article.py
@@ -33,14 +33,17 @@ class ArticleException(Exception):
 class Article(object):
     """Article objects abstract an online news article page
     """
-    def __init__(self, url, title='', source_url='', config=None, **kwargs):
+    def __init__(self, url, title='', source_url='', config=None, content_extractor=None, **kwargs):
         """The **kwargs argument may be filled with config values, which
         is added into the config object
         """
         self.config = config or Configuration()
         self.config = extend_config(self.config, kwargs)
 
-        self.extractor = ContentExtractor(self.config)
+        if content_extractor is None:
+            self.extractor = ContentExtractor(self.config)
+        else:
+            self.extractor = content_extractor
 
         if source_url == '':
             source_url = urls.get_scheme(url) + '://' + urls.get_domain(url)
@@ -212,6 +215,7 @@ def parse(self):
 
         text = ''
         self.top_node = self.extractor.calculate_best_node(self.doc)
+        #self.top_node = document_cleaner.clean(self.top_node)
         if self.top_node is not None:
             video_extractor = VideoExtractor(self.config, self.top_node)
             self.set_movies(video_extractor.get_videos())
diff --git a/newspaper/cleaners.py b/newspaper/cleaners.py
index 5de87b06..70f26dfc 100644
--- a/newspaper/cleaners.py
+++ b/newspaper/cleaners.py
@@ -64,7 +64,7 @@ def clean(self, doc_to_clean):
                                                self.facebook_braodcasting_re)
         doc_to_clean = self.remove_nodes_regex(doc_to_clean, self.twitter_re)
         doc_to_clean = self.clean_para_spans(doc_to_clean)
-        doc_to_clean = self.div_to_para(doc_to_clean, 'div')
+        #doc_to_clean = self.div_to_para(doc_to_clean, 'div')
         doc_to_clean = self.div_to_para(doc_to_clean, 'span')
         return doc_to_clean
 
diff --git a/newspaper/outputformatters.py b/newspaper/outputformatters.py
index c5726717..2e8ade4f 100644
--- a/newspaper/outputformatters.py
+++ b/newspaper/outputformatters.py
@@ -54,6 +54,7 @@ def get_formatted(self, top_node):
         self.remove_trailing_media_div()
         text = self.convert_to_text()
         # print(self.parser.nodeToString(self.get_top_node()))
+        #print(text)
         return (text, html)
 
     def convert_to_text(self):