From: David Read Date: Fri, 09 Nov 2012 16:15:35 +0000 Subject: Adjust popularity score to take account of number of days in the month. X-Git-Url: http://maxious.lambdacomplex.org/git/?p=ckanext-ga-report.git&a=commitdiff&h=20b6eca0a538a77122ce85cf588045784fa9b67e --- Adjust popularity score to take account of number of days in the month. --- --- a/README.rst +++ b/README.rst @@ -33,10 +33,9 @@ googleanalytics.id = UA-1010101-1 googleanalytics.account = Account name (e.g. data.gov.uk, see top level item at https://www.google.com/analytics) ga-report.period = monthly - ga-report.bounce_url = /data + ga-report.bounce_url = / - The ga-report.bounce_url specifies the path to use when calculating bounces. For DGU this is /data - but you may want to set this to /. + The ga-report.bounce_url specifies a particular path to record the bounce rate for. Typically it is / (the home page). 3. Set up this extension's database tables using a paster command. (Ensure your CKAN pyenv is still activated, run the command from ``src/ckanext-ga-report``, alter the ``--config`` option to point to your site config file):: --- a/ckanext/ga_report/command.py +++ b/ckanext/ga_report/command.py @@ -80,6 +80,11 @@ default=False, dest='delete_first', help='Delete data for the period first') + self.parser.add_option('-s', '--skip_url_stats', + action='store_true', + default=False, + dest='skip_url_stats', + help='Skip the download of URL data - just do site-wide stats') def command(self): self._load_config() @@ -95,7 +100,8 @@ return downloader = DownloadAnalytics(svc, profile_id=get_profile_id(svc), - delete_first=self.options.delete_first) + delete_first=self.options.delete_first, + skip_url_stats=self.options.skip_url_stats) time_period = self.args[1] if self.args and len(self.args) > 1 \ else 'latest' --- a/ckanext/ga_report/controller.py +++ b/ckanext/ga_report/controller.py @@ -71,13 +71,13 @@ entries = q.order_by('ga_stat.key').all() def clean_key(key, val): - if key in ['Average time on site', 'Pages per visit', 'New visits', 'Bounces']: + if key in ['Average time on site', 'Pages per visit', 'New visits', 'Bounce rate (home page)']: val = "%.2f" % round(float(val), 2) if key == 'Average time on site': mins, secs = divmod(float(val), 60) hours, mins = divmod(mins, 60) val = '%02d:%02d:%02d (%s seconds) ' % (hours, mins, secs, val) - if key in ['New visits','Bounces']: + if key in ['New visits','Bounce rate (home page)']: val = "%s%%" % val if key in ['Total page views', 'Total visits']: val = int(val) @@ -232,7 +232,7 @@ return render('ga_report/publisher/index.html') def _get_packages(self, publisher=None, count=-1): - '''Returns the datasets in order of visits''' + '''Returns the datasets in order of views''' if count == -1: count = sys.maxint @@ -244,11 +244,11 @@ if publisher: q = q.filter(GA_Url.department_id==publisher.name) q = q.filter(GA_Url.period_name==month) - q = q.order_by('ga_url.visitors::int desc') + q = q.order_by('ga_url.pageviews::int desc') top_packages = [] for entry,package in q.limit(count): if package: - top_packages.append((package, entry.pageviews, entry.visitors)) + top_packages.append((package, entry.pageviews, entry.visits)) else: log.warning('Could not find package associated package') @@ -306,13 +306,13 @@ month = c.month or 'All' connection = model.Session.connection() q = """ - select department_id, sum(pageviews::int) views, sum(visitors::int) visits + select department_id, sum(pageviews::int) views, sum(visits::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 + group by department_id order by views desc """ if limit: q = q + " limit %s;" % (limit) @@ -329,7 +329,7 @@ def _get_publishers(): ''' Returns a list of all publishers. Each item is a tuple: - (names, title) + (name, title) ''' publishers = [] for pub in model.Session.query(model.Group).\ --- a/ckanext/ga_report/download_analytics.py +++ b/ckanext/ga_report/download_analytics.py @@ -17,11 +17,13 @@ class DownloadAnalytics(object): '''Downloads and stores analytics info''' - def __init__(self, service=None, profile_id=None, delete_first=False): + def __init__(self, service=None, profile_id=None, delete_first=False, + skip_url_stats=False): self.period = config['ga-report.period'] self.service = service self.profile_id = profile_id self.delete_first = delete_first + self.skip_url_stats = skip_url_stats def specific_month(self, date): import calendar @@ -96,31 +98,32 @@ 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 this period "%s"', period_name) ga_model.delete(period_name) - # 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 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 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. + if not self.skip_url_stats: + # Clean out old url data before storing the new + 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 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 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 ) @@ -159,8 +162,8 @@ start_date = start_date.strftime('%Y-%m-%d') end_date = end_date.strftime('%Y-%m-%d') query = 'ga:pagePath=%s$' % path - metrics = 'ga:uniquePageviews, ga:visits' - sort = '-ga:uniquePageviews' + metrics = 'ga:pageviews, ga:visits' + sort = '-ga:pageviews' # Supported query params at # https://developers.google.com/analytics/devguides/reporting/core/v3/reference @@ -177,8 +180,12 @@ packages = [] for entry in results.get('rows'): (loc,pageviews,visits) = entry - url = _normalize_url('http:/' + loc) + url = _normalize_url('http:/' + loc) # strips off domain e.g. www.data.gov.uk or data.gov.uk + if not url.startswith('/dataset/') and not url.startswith('/publisher/'): + # filter out strays like: + # /data/user/login?came_from=http://data.gov.uk/dataset/os-code-point-open + # /403.html?page=/about&from=http://data.gov.uk/publisher/planning-inspectorate continue packages.append( (url, pageviews, visits,) ) # Temporary hack return dict(url=packages) @@ -212,8 +219,8 @@ results = self.service.data().ga().get( ids='ga:' + self.profile_id, start_date=start_date, - metrics='ga:uniquePageviews', - sort='-ga:uniquePageviews', + metrics='ga:pageviews', + sort='-ga:pageviews', max_results=10000, end_date=end_date).execute() result_data = results.get('rows') @@ -234,25 +241,27 @@ } ga_model.update_sitewide_stats(period_name, "Totals", data) - # Bounces from /data. This url is specified in configuration because - # for DGU we don't want /. - path = config.get('ga-report.bounce_url','/') - print path - results = self.service.data().ga().get( - ids='ga:' + self.profile_id, - filters='ga:pagePath=~%s$' % (path,), - start_date=start_date, - metrics='ga:bounces,ga:uniquePageviews', + # Bounces from / or another configurable page. + path = '/%s%s' % (config.get('googleanalytics.account'), + config.get('ga-report.bounce_url', '/')) + results = self.service.data().ga().get( + ids='ga:' + self.profile_id, + filters='ga:pagePath==%s' % (path,), + start_date=start_date, + metrics='ga:bounces,ga:pageviews', dimensions='ga:pagePath', max_results=10000, end_date=end_date).execute() result_data = results.get('rows') - for results in result_data: - if results[0] == path: - bounce, total = [float(x) for x in results[1:]] - pct = 100 * bounce/total - print "%d bounces from %d total == %s" % (bounce, total, pct) - ga_model.update_sitewide_stats(period_name, "Totals", {'Bounces': pct}) + if not result_data or len(result_data) != 1: + log.error('Could not pinpoint the bounces for path: %s. Got results: %r', + path, result_data) + return + results = result_data[0] + bounces, total = [float(x) for x in result_data[0][1:]] + pct = 100 * bounces/total + log.info('%d bounces from %d total == %s', bounces, total, pct) + ga_model.update_sitewide_stats(period_name, "Totals", {'Bounce rate (home page)': pct}) def _locale_stats(self, start_date, end_date, period_name): @@ -260,8 +269,8 @@ results = self.service.data().ga().get( ids='ga:' + self.profile_id, start_date=start_date, - metrics='ga:uniquePageviews', - sort='-ga:uniquePageviews', + metrics='ga:pageviews', + sort='-ga:pageviews', dimensions="ga:language,ga:country", max_results=10000, end_date=end_date).execute() @@ -284,8 +293,8 @@ results = self.service.data().ga().get( ids='ga:' + self.profile_id, start_date=start_date, - metrics='ga:uniquePageviews', - sort='-ga:uniquePageviews', + metrics='ga:pageviews', + sort='-ga:pageviews', dimensions="ga:socialNetwork,ga:referralPath", max_results=10000, end_date=end_date).execute() @@ -303,8 +312,8 @@ results = self.service.data().ga().get( ids='ga:' + self.profile_id, start_date=start_date, - metrics='ga:uniquePageviews', - sort='-ga:uniquePageviews', + metrics='ga:pageviews', + sort='-ga:pageviews', dimensions="ga:operatingSystem,ga:operatingSystemVersion", max_results=10000, end_date=end_date).execute() @@ -328,8 +337,8 @@ results = self.service.data().ga().get( ids='ga:' + self.profile_id, start_date=start_date, - metrics='ga:uniquePageviews', - sort='-ga:uniquePageviews', + metrics='ga:pageviews', + sort='-ga:pageviews', dimensions="ga:browser,ga:browserVersion", max_results=10000, end_date=end_date).execute() @@ -377,8 +386,8 @@ results = self.service.data().ga().get( ids='ga:' + self.profile_id, start_date=start_date, - metrics='ga:uniquePageviews', - sort='-ga:uniquePageviews', + metrics='ga:pageviews', + sort='-ga:pageviews', dimensions="ga:mobileDeviceBranding, ga:mobileDeviceInfo", max_results=10000, end_date=end_date).execute() --- a/ckanext/ga_report/ga_model.py +++ b/ckanext/ga_report/ga_model.py @@ -9,6 +9,8 @@ import ckan.model as model from ckan.lib.base import * + +log = __import__('logging').getLogger(__name__) def make_uuid(): return unicode(uuid.uuid4()) @@ -27,7 +29,7 @@ Column('period_name', types.UnicodeText), Column('period_complete_day', types.Integer), Column('pageviews', types.UnicodeText), - Column('visitors', types.UnicodeText), + Column('visits', types.UnicodeText), Column('url', types.UnicodeText), Column('department_id', types.UnicodeText), Column('package_id', types.UnicodeText), @@ -63,7 +65,7 @@ Column('period_name', types.UnicodeText), Column('publisher_name', types.UnicodeText), Column('views', types.UnicodeText), - Column('visitors', types.UnicodeText), + Column('visits', types.UnicodeText), Column('toplevel', types.Boolean, default=False), Column('subpublishercount', types.Integer, default=0), Column('parent', types.UnicodeText), @@ -111,9 +113,7 @@ >>> 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:]) + return '/' + '/'.join(url.split('/')[3:]) def _get_package_and_publisher(url): @@ -157,25 +157,6 @@ 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() @@ -189,7 +170,7 @@ stores them in GA_Url under the period and recalculates the totals for the 'All' period. ''' - for url, views, visitors in url_data: + for url, views, visits in url_data: package, publisher = _get_package_and_publisher(url) @@ -198,7 +179,7 @@ filter(GA_Url.url==url).first() if item: item.pageviews = item.pageviews + views - item.visitors = item.visitors + visitors + item.visits = item.visits + visits if not item.package_id: item.package_id = package if not item.department_id: @@ -210,7 +191,7 @@ 'period_complete_day': period_complete_day, 'url': url, 'pageviews': views, - 'visitors': visitors, + 'visits': visits, 'department_id': publisher, 'package_id': package } @@ -223,7 +204,7 @@ 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]) + old_visits = sum([int(o.visits) for o in old]) entries = model.Session.query(GA_Url).\ filter(GA_Url.period_name!='All').\ @@ -233,7 +214,7 @@ '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, + 'visits': sum([int(e.visits or 0) for e in entries]) + old_visits, 'department_id': publisher, 'package_id': package } @@ -283,7 +264,7 @@ filter(model.Group.type=='publisher').\ filter(model.Group.state=='active').all() for publisher in publishers: - views, visitors, subpub = update_publisher(period_name, publisher, publisher.name) + views, visits, subpub = update_publisher(period_name, publisher, publisher.name) parent, parents = '', publisher.get_groups('publisher') if parents: parent = parents[0].name @@ -292,7 +273,7 @@ filter(GA_Publisher.publisher_name==publisher.name).first() if item: item.views = views - item.visitors = visitors + item.visits = visits item.publisher_name = publisher.name item.toplevel = publisher in toplevel item.subpublishercount = subpub @@ -304,7 +285,7 @@ 'period_name': period_name, 'publisher_name': publisher.name, 'views': views, - 'visitors': visitors, + 'visits': visits, 'toplevel': publisher in toplevel, 'subpublishercount': subpub, 'parent': parent @@ -314,7 +295,7 @@ def update_publisher(period_name, pub, part=''): - views,visitors,subpub = 0, 0, 0 + views,visits,subpub = 0, 0, 0 for publisher in go_down_tree(pub): subpub = subpub + 1 items = model.Session.query(GA_Url).\ @@ -322,9 +303,9 @@ filter(GA_Url.department_id==publisher.name).all() for item in items: views = views + int(item.pageviews) - visitors = visitors + int(item.visitors) - - return views, visitors, (subpub-1) + visits = visits + int(item.visits) + + return views, visits, (subpub-1) def get_top_level(): @@ -364,3 +345,34 @@ q.delete() model.Session.commit() +def get_score_for_dataset(dataset_name): + ''' + Returns a "current popularity" score for a dataset, + based on how many views it has had recently. + ''' + import datetime + now = datetime.datetime.now() + last_month = now - datetime.timedelta(days=30) + period_names = ['%s-%02d' % (last_month.year, last_month.month), + '%s-%02d' % (now.year, now.month), + ] + + score = 0 + for period_name in period_names: + score /= 2 # previous periods are discounted by 50% + entry = model.Session.query(GA_Url)\ + .filter(GA_Url.period_name==period_name)\ + .filter(GA_Url.package_id==dataset_name).first() + # score + if entry: + views = float(entry.pageviews) + if entry.period_complete_day: + views_per_day = views / entry.period_complete_day + else: + views_per_day = views / 15 # guess + score += views_per_day + + score = int(score * 100) + log.debug('Popularity %s: %s', score, dataset_name) + return score + --- a/ckanext/ga_report/helpers.py +++ b/ckanext/ga_report/helpers.py @@ -106,7 +106,7 @@ if not p in datasets: datasets[p] = {'views':0, 'visits': 0} datasets[p]['views'] = datasets[p]['views'] + int(entry.pageviews) - datasets[p]['visits'] = datasets[p]['visits'] + int(entry.visitors) + datasets[p]['visits'] = datasets[p]['visits'] + int(entry.visits) results = [] for k, v in datasets.iteritems(): --- a/ckanext/ga_report/templates/ga_report/notes.html +++ b/ckanext/ga_report/templates/ga_report/notes.html @@ -6,9 +6,8 @@

  • Notes

    --- a/ckanext/ga_report/templates/ga_report/publisher/index.html +++ b/ckanext/ga_report/templates/ga_report/publisher/index.html @@ -41,14 +41,14 @@ - + - + --- a/ckanext/ga_report/templates/ga_report/publisher/read.html +++ b/ckanext/ga_report/templates/ga_report/publisher/read.html @@ -47,14 +47,14 @@
    PublisherDataset Visits Dataset Views
    ${h.link_to(publisher.title, h.url_for(controller='ckanext.ga_report.controller:GaDatasetReport', action='read_publisher', id=publisher.name))} ${visits} ${views}
    - + - + --- /dev/null +++ b/ckanext/ga_report/tests/test_model.py @@ -1,1 +1,18 @@ +from nose.tools import assert_equal +from ckanext.ga_report.ga_model import _normalize_url + +class TestNormalizeUrl: + def test_normal(self): + assert_equal(_normalize_url('http://data.gov.uk/dataset/weekly_fuel_prices'), + '/dataset/weekly_fuel_prices') + + def test_www_dot(self): + assert_equal(_normalize_url('http://www.data.gov.uk/dataset/weekly_fuel_prices'), + '/dataset/weekly_fuel_prices') + + def test_https(self): + assert_equal(_normalize_url('https://data.gov.uk/dataset/weekly_fuel_prices'), + '/dataset/weekly_fuel_prices') + +
    DatasetVisits Views
    ${h.link_to(package.title or package.name, h.url_for(controller='package', action='read', id=package.name))} ${visits} ${views}