From: Ross Jones Date: Wed, 07 Nov 2012 14:35:17 +0000 Subject: Checking for duplicate "All" stats X-Git-Url: http://maxious.lambdacomplex.org/git/?p=ckanext-ga-report.git&a=commitdiff&h=f765d047ea03b3b1812edeb3efa84c89f57661d9 --- Checking for duplicate "All" stats --- --- a/ckanext/ga_report/controller.py +++ b/ckanext/ga_report/controller.py @@ -309,6 +309,8 @@ select department_id, sum(pageviews::int) views, sum(visitors::int) visits from ga_url where department_id <> '' + and package_id <> '' + and url like '/dataset/%%' and period_name=%s group by department_id order by visits desc """ --- a/ckanext/ga_report/download_analytics.py +++ b/ckanext/ga_report/download_analytics.py @@ -92,33 +92,40 @@ def download_and_store(self, periods): for period_name, period_complete_day, start_date, end_date in periods: + log.info('Period "%s" (%s - %s)', + self.get_full_period_name(period_name, period_complete_day), + start_date.strftime('%Y-%m-%d'), + end_date.strftime('%Y-%m-%d')) + if self.delete_first: - log.info('Deleting existing Analytics for period "%s"', + log.info('Deleting existing Analytics for this period "%s"', period_name) ga_model.delete(period_name) - log.info('Downloading Analytics for period "%s" (%s - %s)', - self.get_full_period_name(period_name, period_complete_day), - start_date.strftime('%Y %m %d'), - end_date.strftime('%Y %m %d')) # Clean up the entries before we run this ga_model.pre_update_url_stats(period_name) accountName = config.get('googleanalytics.account') + log.info('Downloading analytics for dataset views') data = self.download(start_date, end_date, '~/%s/dataset/[a-z0-9-_]+' % accountName) - log.info('Storing Dataset Analytics for period "%s"', - self.get_full_period_name(period_name, period_complete_day)) + + log.info('Storing dataset views (%i rows)', len(data.get('url'))) self.store(period_name, period_complete_day, data, ) + log.info('Downloading analytics for publisher views') data = self.download(start_date, end_date, '~/%s/publisher/[a-z0-9-_]+' % accountName) - log.info('Storing Publisher Analytics for period "%s"', - self.get_full_period_name(period_name, period_complete_day)) + + log.info('Storing publisher views (%i rows)', len(data.get('url'))) self.store(period_name, period_complete_day, data,) + log.info('Aggregating datasets by publisher') ga_model.update_publisher_stats(period_name) # about 30 seconds. + + log.info('Downloading and storing analytics for site-wide stats') self.sitewide_stats( period_name ) + log.info('Downloading and storing analytics for social networks') self.update_social_info(period_name, start_date, end_date) @@ -187,12 +194,10 @@ start_date = '%s-01' % period_name end_date = '%s-%s' % (period_name, last_day_of_month) - print 'Sitewide_stats for %s (%s -> %s)' % (period_name, start_date, end_date) - funcs = ['_totals_stats', '_social_stats', '_os_stats', '_locale_stats', '_browser_stats', '_mobile_stats'] for f in funcs: - print ' + Fetching %s stats' % f.split('_')[1] + log.info('Downloading analytics for %s' % f.split('_')[1]) getattr(self, f)(start_date, end_date, period_name) def _get_results(result_data, f): --- a/ckanext/ga_report/ga_model.py +++ b/ckanext/ga_report/ga_model.py @@ -116,7 +116,7 @@ return '/' + '/'.join(url.split('/')[2:]) -def _get_department_id_of_url(url): +def _get_package_and_publisher(url): # e.g. /dataset/fuel_prices # e.g. /dataset/fuel_prices/resource/e63380d4 dataset_match = re.match('/dataset/([^/]+)(/.*)?', url) @@ -126,12 +126,13 @@ 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(): @@ -185,25 +186,41 @@ def update_url_stats(period_name, period_complete_day, url_data): for url, views, visitors in url_data: - department_id = _get_department_id_of_url(url) - - package = None - if url.startswith('/dataset/'): - package = url[len('/dataset/'):] - - values = {'id': make_uuid(), - 'period_name': period_name, - 'period_complete_day': period_complete_day, - 'url': url, - 'pageviews': views, - 'visitors': visitors, - 'department_id': department_id, - 'package_id': package - } - model.Session.add(GA_Url(**values)) + 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.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: + values = {'id': make_uuid(), + 'period_name': period_name, + 'period_complete_day': period_complete_day, + 'url': url, + 'pageviews': views, + 'visitors': visitors, + '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() @@ -211,14 +228,14 @@ 'period_name': 'All', 'period_complete_day': 0, 'url': url, - 'pageviews': sum([int(e.pageviews) for e in entries]), - 'visitors': sum([int(e.visitors) for e in entries]), - 'department_id': department_id, + '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() - --- a/ckanext/ga_report/templates/ga_report/notes.html +++ b/ckanext/ga_report/templates/ga_report/notes.html @@ -6,11 +6,11 @@
  • Notes