Parse created_at consistently

data_quality/tasks/aggregate.py

@@ -53,6 +53,7 @@ def run(self, pipeline):
         if self.assess_timeliness:
             extractor = RelevancePeriodExtractor(self.config)
             extractor.run()
+            self.lookup = self.get_lookup()
 
         with compat.UnicodeAppender(self.result_file, quoting=csv.QUOTE_MINIMAL) as result_file:
             source = self.get_source(pipeline.data_source)
@@ -183,9 +184,8 @@ def get_publication_delay(self, source):
             raise ValueError(('Cannot assess timeliness: Bad value for "created_at"'
                               ' in source with id {0}.').format(source['id']))
         dates['period_end'] = dates['period_end'] or dates['period_start']
-        self.timeliness_period = int(self.timeliness_period) * 30
         timely_until = dates['period_end'] + \
-                       timedelta(days=self.timeliness_period)
+                       timedelta(days=(self.timeliness_period * 30))
         if dates['period_start'] <= dates['publication_date'] <= timely_until:
             delay = 0
         else:

data_quality/tasks/extract_relevance_period.py

@@ -54,13 +54,14 @@ def run(self):
 
         for source in sources:
             if source['period_id'] is None:
-                self.timeliness_strategy = ['created_at']
-                period_start, period_end = self.identify_period(source)
+                creation_date = utilities.date_from_string(source['created_at'])
+                dates = [creation_date, creation_date]
             else:
                 period_start, period_end = source['period_id']
-            periods = [period_start.date(), period_end.date()]
-            periods = [period.strftime('%d-%m-%Y') for period in periods]
-            source['period_id'] = '/'.join(periods)
+                dates = [period_start.date(), period_end.date()]
+            dates = [date.strftime('%d-%m-%Y') if isinstance(date, datetime.date)
+                     else '' for date in dates]
+            source['period_id'] = '/'.join(dates)
         self.update_sources_period(sources)
 
     def extract_period_from_sources(self):