Show top datasets cross-publisher. Drop-down for the publisher. Browser version numbers filtered on download, so you get this version in the CSV too - for privacy. single_popular_dataset now copes when not much data, and can return the figures so DGU can reskin it in its own repo. Notes about usage stats centralised to notes.html.
[ckanext-ga-report.git] / ckanext / ga_report / ga_model.py
blob:a/ckanext/ga_report/ga_model.py -> blob:b/ckanext/ga_report/ga_model.py
--- a/ckanext/ga_report/ga_model.py
+++ b/ckanext/ga_report/ga_model.py
@@ -1,10 +1,10 @@
 import re
 import uuid
 
-from sqlalchemy import Table, Column, MetaData
+from sqlalchemy import Table, Column, MetaData, ForeignKey
 from sqlalchemy import types
 from sqlalchemy.sql import select
-from sqlalchemy.orm import mapper
+from sqlalchemy.orm import mapper, relation
 from sqlalchemy import func
 
 import ckan.model as model
@@ -14,8 +14,6 @@
     return unicode(uuid.uuid4())
 
 metadata = MetaData()
-
-
 
 class GA_Url(object):
 
@@ -32,6 +30,7 @@
                       Column('visitors', types.UnicodeText),
                       Column('url', types.UnicodeText),
                       Column('department_id', types.UnicodeText),
+                      Column('package_id', types.UnicodeText),
                 )
 mapper(GA_Url, url_table)
 
@@ -112,12 +111,10 @@
     >>> normalize_url('http://data.gov.uk/dataset/weekly_fuel_prices')
     '/dataset/weekly_fuel_prices'
     '''
-    # Deliberately leaving a /
-    url = url.replace('http:/','')
-    return '/' + '/'.join(url.split('/')[2:])
-
-
-def _get_department_id_of_url(url):
+    return '/' + '/'.join(url.split('/')[3:])
+
+
+def _get_package_and_publisher(url):
     # e.g. /dataset/fuel_prices
     # e.g. /dataset/fuel_prices/resource/e63380d4
     dataset_match = re.match('/dataset/([^/]+)(/.*)?', url)
@@ -127,12 +124,13 @@
         if dataset:
             publisher_groups = dataset.get_groups('publisher')
             if publisher_groups:
-                return publisher_groups[0].name
+                return dataset_ref,publisher_groups[0].name
+        return dataset_ref, None
     else:
         publisher_match = re.match('/publisher/([^/]+)(/.*)?', url)
         if publisher_match:
-            return publisher_match.groups()[0]
-
+            return None, publisher_match.groups()[0]
+    return None, None
 
 def update_sitewide_stats(period_name, stat_name, data):
     for k,v in data.iteritems():
@@ -157,34 +155,91 @@
         model.Session.commit()
 
 
+def update_url_stat_totals(period_name):
+
+    """
+        items = model.Session.query(GA_Url).\
+            filter(GA_Url.period_name != "All").\
+            filter(GA_Url.url==url).all()
+        values = {'id': make_uuid(),
+                  'period_name': "All",
+                  'period_complete_day': "0",
+                  'url': url,
+                  'pageviews': sum([int(x.pageviews) for x in items]),
+                  'visitors': sum([int(x.visitors) for x in items]),
+                  'department_id': department_id,
+                  'package_id': package
+                 }
+        model.Session.add(GA_Url(**values))
+        model.Session.commit()
+    """
+
+def pre_update_url_stats(period_name):
+    model.Session.query(GA_Url).\
+            filter(GA_Url.period_name==period_name).delete()
+    model.Session.query(GA_Url).\
+            filter(GA_Url.period_name=='All').delete()
+
 
 def update_url_stats(period_name, period_complete_day, url_data):
+    '''
+    Given a list of urls and number of hits for each during a given period,
+    stores them in GA_Url under the period and recalculates the totals for
+    the 'All' period.
+    '''
     for url, views, visitors in url_data:
-        url = _normalize_url(url)
-        department_id = _get_department_id_of_url(url)
-
-        # see if the row for this url & month is in the table already
+        package, publisher = _get_package_and_publisher(url)
+
+
         item = model.Session.query(GA_Url).\
             filter(GA_Url.period_name==period_name).\
             filter(GA_Url.url==url).first()
         if item:
-            item.period_name = period_name
-            item.pageviews = views
-            item.visitors = visitors
-            item.department_id = department_id
+            item.pageviews = item.pageviews + views
+            item.visitors = item.visitors + visitors
+            if not item.package_id:
+                item.package_id = package
+            if not item.department_id:
+                item.department_id = publisher
             model.Session.add(item)
         else:
-            # create the row
             values = {'id': make_uuid(),
                       'period_name': period_name,
                       'period_complete_day': period_complete_day,
                       'url': url,
                       'pageviews': views,
                       'visitors': visitors,
-                      'department_id': department_id
+                      'department_id': publisher,
+                      'package_id': package
                      }
             model.Session.add(GA_Url(**values))
         model.Session.commit()
+
+        if package:
+            old_pageviews, old_visits = 0, 0
+            old = model.Session.query(GA_Url).\
+                filter(GA_Url.period_name=='All').\
+                filter(GA_Url.url==url).all()
+            old_pageviews = sum([int(o.pageviews) for o in old])
+            old_visits = sum([int(o.visitors) for o in old])
+
+            entries = model.Session.query(GA_Url).\
+                filter(GA_Url.period_name!='All').\
+                filter(GA_Url.url==url).all()
+            values = {'id': make_uuid(),
+                      'period_name': 'All',
+                      'period_complete_day': 0,
+                      'url': url,
+                      'pageviews': sum([int(e.pageviews) for e in entries]) + old_pageviews,
+                      'visitors': sum([int(e.visitors) for e in entries]) + old_visits,
+                      'department_id': publisher,
+                      'package_id': package
+                     }
+
+            model.Session.add(GA_Url(**values))
+            model.Session.commit()
+
+
 
 
 def update_social(period_name, data):