Fix the same bug again but for real this time

dataset.py

@@ -128,6 +128,8 @@ def get_crisis_dataset():
             date_path = topic_path / 'public' / 'content' / pub_date
             for article_filename in util.files(date_path, extension='.cont'):
                 article_file_path = date_path / article_filename
+                if '1093.htm.cont' not in str(article_file_path):
+                    continue
                 print(article_file_path)
                 article = {'pub_date': pub_date, 'sentences': []}
 

heideltime_util.py

@@ -49,9 +49,9 @@ def mentioned_dates_by_sentence(filename, pub_date):
                     match = date_format_regex.match(child.attrib['value'])
                     if match is not None:
                         dates_for_this_sentence.append(match[0])
-            dates_by_sentences.append(dates_for_this_sentence)
         except ET.ParseError:
-            continue
+            pass
+        dates_by_sentences.append(dates_for_this_sentence)
 
     # change directory back
     os.chdir(working_dir)