Modified to collect visit instead of visitors.
--- a/.gitignore
+++ b/.gitignore
@@ -1,4 +1,6 @@
*.py[co]
+*.py~
+.gitignore
# Packages
*.egg
@@ -13,6 +15,10 @@
develop-eggs
.installed.cfg
+# Private info
+credentials.json
+token.dat
+
# Installer logs
pip-log.txt
--- a/README.md
+++ /dev/null
@@ -1,4 +1,1 @@
-ckanext-ga-report
-=================
-For creating detailed reports of CKAN analytics, sliced by group
--- /dev/null
+++ b/README.rst
@@ -1,1 +1,110 @@
+ckanext-ga-report
+=================
+**Status:** Development
+
+**CKAN Version:** 1.7.1+
+
+
+Overview
+--------
+
+For creating detailed reports of CKAN analytics, including totals per group.
+
+Whereas ckanext-googleanalytics focusses on providing page view stats a recent period and for all time (aimed at end users), ckanext-ga-report is more interested in building regular periodic reports (more for site managers to monitor).
+
+Contents of this extension:
+
+ * Use the CLI tool to download Google Analytics data for each time period into this extension's database tables
+
+ * Users can view the data as web page reports
+
+
+Installation
+------------
+
+1. Activate you CKAN python environment and install this extension's software::
+
+ $ pyenv/bin/activate
+ $ pip install -e git+https://github.com/datagovuk/ckanext-ga-report.git#egg=ckanext-ga-report
+
+2. Ensure you development.ini (or similar) contains the info about your Google Analytics account and configuration::
+
+ 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
+
+ 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 /.
+
+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)::
+
+ $ paster initdb --config=../ckan/development.ini
+
+4. Enable the extension in your CKAN config file by adding it to ``ckan.plugins``::
+
+ ckan.plugins = ga-report
+
+Problem shooting
+----------------
+
+* ``(ProgrammingError) relation "ga_url" does not exist``
+ This means that the ``paster initdb`` step has not been run successfully. Refer to the installation instructions for this extension.
+
+
+Authorization
+--------------
+
+Before you can access the data, you need to set up the OAUTH details which you can do by following the `instructions <https://developers.google.com/analytics/resources/tutorials/hello-analytics-api>`_ the outcome of which will be a file called credentials.json which should look like credentials.json.template with the relevant fields completed. These steps are below for convenience:
+
+1. Visit the `Google APIs Console <https://code.google.com/apis/console>`_
+
+2. Sign-in and create a project or use an existing project.
+
+3. In the `Services pane <https://code.google.com/apis/console#:services>`_ , activate Analytics API for your project. If prompted, read and accept the terms of service.
+
+4. Go to the `API Access pane <https://code.google.com/apis/console/#:access>`_
+
+5. Click Create an OAuth 2.0 client ID....
+
+6. Fill out the Branding Information fields and click Next.
+
+7. In Client ID Settings, set Application type to Installed application.
+
+8. Click Create client ID
+
+9. The details you need below are Client ID, Client secret, and Redirect URIs
+
+
+Once you have set up your credentials.json file you can generate an oauth token file by using the
+following command, which will store your oauth token in a file called token.dat once you have finished
+giving permission in the browser::
+
+ $ paster getauthtoken --config=../ckan/development.ini
+
+
+Tutorial
+--------
+
+Download some GA data and store it in CKAN's database. (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) and specifying the name of your auth file (token.dat by default) from the previous step::
+
+ $ paster loadanalytics token.dat latest --config=../ckan/development.ini
+
+The value after the token file is how much data you want to retrieve, this can be
+
+* **all** - data for all time (since 2010)
+
+* **latest** - (default) just the 'latest' data
+
+* **YYYY-MM-DD** - just data for all time periods going back to (and including) this date
+
+
+
+Software Licence
+================
+
+This software is developed by Cabinet Office. It is Crown Copyright and opened up under the Open Government Licence (OGL) (which is compatible with Creative Commons Attibution License).
+
+OGL terms: http://www.nationalarchives.gov.uk/doc/open-government-licence/
+
--- /dev/null
+++ b/ckanext/__init__.py
@@ -1,1 +1,8 @@
+# this is a namespace package
+try:
+ import pkg_resources
+ pkg_resources.declare_namespace(__name__)
+except ImportError:
+ import pkgutil
+ __path__ = pkgutil.extend_path(__path__, __name__)
--- /dev/null
+++ b/ckanext/ga_report/__init__.py
@@ -1,1 +1,8 @@
+# this is a namespace package
+try:
+ import pkg_resources
+ pkg_resources.declare_namespace(__name__)
+except ImportError:
+ import pkgutil
+ __path__ = pkgutil.extend_path(__path__, __name__)
--- /dev/null
+++ b/ckanext/ga_report/command.py
@@ -1,1 +1,110 @@
+import logging
+import datetime
+from ckan.lib.cli import CkanCommand
+# No other CKAN imports allowed until _load_config is run,
+# or logging is disabled
+
+
+class InitDB(CkanCommand):
+ """Initialise the extension's database tables
+ """
+ summary = __doc__.split('\n')[0]
+ usage = __doc__
+ max_args = 0
+ min_args = 0
+
+ def command(self):
+ self._load_config()
+
+ import ckan.model as model
+ model.Session.remove()
+ model.Session.configure(bind=model.meta.engine)
+ log = logging.getLogger('ckanext.ga-report')
+
+ import ga_model
+ ga_model.init_tables()
+ log.info("DB tables are setup")
+
+
+class GetAuthToken(CkanCommand):
+ """ Get's the Google auth token
+
+ Usage: paster getauthtoken <credentials_file>
+
+ Where <credentials_file> is the file name containing the details
+ for the service (obtained from https://code.google.com/apis/console).
+ By default this is set to credentials.json
+ """
+ summary = __doc__.split('\n')[0]
+ usage = __doc__
+ max_args = 0
+ min_args = 0
+
+ def command(self):
+ """
+ In this case we don't want a valid service, but rather just to
+ force the user through the auth flow. We allow this to complete to
+ act as a form of verification instead of just getting the token and
+ assuming it is correct.
+ """
+ from ga_auth import init_service
+ init_service('token.dat',
+ self.args[0] if self.args
+ else 'credentials.json')
+
+
+class LoadAnalytics(CkanCommand):
+ """Get data from Google Analytics API and save it
+ in the ga_model
+
+ Usage: paster loadanalytics <tokenfile> <time-period>
+
+ Where <tokenfile> is the name of the auth token file from
+ the getauthtoken step.
+
+ And where <time-period> is:
+ all - data for all time
+ latest - (default) just the 'latest' data
+ YYYY-MM - just data for the specific month
+ """
+ summary = __doc__.split('\n')[0]
+ usage = __doc__
+ max_args = 2
+ min_args = 1
+
+ def __init__(self, name):
+ super(LoadAnalytics, self).__init__(name)
+ self.parser.add_option('-d', '--delete-first',
+ action='store_true',
+ default=False,
+ dest='delete_first',
+ help='Delete data for the period first')
+
+ def command(self):
+ self._load_config()
+
+ from download_analytics import DownloadAnalytics
+ from ga_auth import (init_service, get_profile_id)
+
+ try:
+ svc = init_service(self.args[0], None)
+ except TypeError:
+ print ('Have you correctly run the getauthtoken task and '
+ 'specified the correct token file?')
+ return
+
+ downloader = DownloadAnalytics(svc, profile_id=get_profile_id(svc),
+ delete_first=self.options.delete_first)
+
+ time_period = self.args[1] if self.args and len(self.args) > 1 \
+ else 'latest'
+ if time_period == 'all':
+ downloader.all_()
+ elif time_period == 'latest':
+ downloader.latest()
+ else:
+ # The month to use
+ for_date = datetime.datetime.strptime(time_period, '%Y-%m')
+ downloader.specific_month(for_date)
+
--- /dev/null
+++ b/ckanext/ga_report/controller.py
@@ -1,1 +1,372 @@
-
+import re
+import csv
+import sys
+import logging
+import operator
+import collections
+from ckan.lib.base import (BaseController, c, g, render, request, response, abort)
+
+import sqlalchemy
+from sqlalchemy import func, cast, Integer
+import ckan.model as model
+from ga_model import GA_Url, GA_Stat, GA_ReferralStat
+
+log = logging.getLogger('ckanext.ga-report')
+
+
+def _get_month_name(strdate):
+ import calendar
+ from time import strptime
+ d = strptime(strdate, '%Y-%m')
+ return '%s %s' % (calendar.month_name[d.tm_mon], d.tm_year)
+
+
+def _month_details(cls):
+ months = []
+ vals = model.Session.query(cls.period_name).distinct().all()
+ for m in vals:
+ months.append( (m[0], _get_month_name(m[0])))
+ return sorted(months, key=operator.itemgetter(0), reverse=True)
+
+
+class GaReport(BaseController):
+
+ def csv(self, month):
+ import csv
+
+ q = model.Session.query(GA_Stat)
+ if month != 'all':
+ q = q.filter(GA_Stat.period_name==month)
+ entries = q.order_by('GA_Stat.period_name, GA_Stat.stat_name, GA_Stat.key').all()
+
+ response.headers['Content-Type'] = "text/csv; charset=utf-8"
+ response.headers['Content-Disposition'] = str('attachment; filename=stats_%s.csv' % (month,))
+
+ writer = csv.writer(response)
+ writer.writerow(["Period", "Statistic", "Key", "Value"])
+
+ for entry in entries:
+ writer.writerow([entry.period_name.encode('utf-8'),
+ entry.stat_name.encode('utf-8'),
+ entry.key.encode('utf-8'),
+ entry.value.encode('utf-8')])
+
+ def index(self):
+
+ # Get the month details by fetching distinct values and determining the
+ # month names from the values.
+ c.months = _month_details(GA_Stat)
+
+ # Work out which month to show, based on query params of the first item
+ c.month_desc = 'all months'
+ c.month = request.params.get('month', '')
+ if c.month:
+ c.month_desc = ''.join([m[1] for m in c.months if m[0]==c.month])
+
+ q = model.Session.query(GA_Stat).\
+ filter(GA_Stat.stat_name=='Totals')
+ if c.month:
+ q = q.filter(GA_Stat.period_name==c.month)
+ 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']:
+ 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']:
+ val = "%s%%" % val
+ if key in ['Total page views', 'Total visits']:
+ val = int(val)
+
+ return key, val
+
+ c.global_totals = []
+ if c.month:
+ for e in entries:
+ key, val = clean_key(e.key, e.value)
+ c.global_totals.append((key, val))
+ else:
+ d = collections.defaultdict(list)
+ for e in entries:
+ d[e.key].append(float(e.value))
+ for k, v in d.iteritems():
+ if k in ['Total page views', 'Total visits']:
+ v = sum(v)
+ else:
+ v = float(sum(v))/len(v)
+ key, val = clean_key(k,v)
+
+ c.global_totals.append((key, val))
+ c.global_totals = sorted(c.global_totals, key=operator.itemgetter(0))
+
+ keys = {
+ 'Browser versions': 'browser_versions',
+ 'Browsers': 'browsers',
+ 'Operating Systems versions': 'os_versions',
+ 'Operating Systems': 'os',
+ 'Social sources': 'social_networks',
+ 'Languages': 'languages',
+ 'Country': 'country'
+ }
+
+ def shorten_name(name, length=60):
+ return (name[:length] + '..') if len(name) > 60 else name
+
+ def fill_out_url(url):
+ import urlparse
+ return urlparse.urljoin(g.site_url, url)
+
+ c.social_referrer_totals, c.social_referrers = [], []
+ q = model.Session.query(GA_ReferralStat)
+ q = q.filter(GA_ReferralStat.period_name==c.month) if c.month else q
+ q = q.order_by('ga_referrer.count::int desc')
+ for entry in q.all():
+ c.social_referrers.append((shorten_name(entry.url), fill_out_url(entry.url),
+ entry.source,entry.count))
+
+ q = model.Session.query(GA_ReferralStat.url,
+ func.sum(GA_ReferralStat.count).label('count'))
+ q = q.filter(GA_ReferralStat.period_name==c.month) if c.month else q
+ q = q.order_by('count desc').group_by(GA_ReferralStat.url)
+ for entry in q.all():
+ c.social_referrer_totals.append((shorten_name(entry[0]), fill_out_url(entry[0]),'',
+ entry[1]))
+
+ for k, v in keys.iteritems():
+ q = model.Session.query(GA_Stat).\
+ filter(GA_Stat.stat_name==k)
+ if c.month:
+ entries = []
+ q = q.filter(GA_Stat.period_name==c.month).\
+ order_by('ga_stat.value::int desc')
+
+ d = collections.defaultdict(int)
+ for e in q.all():
+ d[e.key] += int(e.value)
+ entries = []
+ for key, val in d.iteritems():
+ entries.append((key,val,))
+ entries = sorted(entries, key=operator.itemgetter(1), reverse=True)
+
+ # Get the total for each set of values and then set the value as
+ # a percentage of the total
+ if k == 'Social sources':
+ total = sum([x for n,x in c.global_totals if n == 'Total visits'])
+ else:
+ total = sum([num for _,num in entries])
+ setattr(c, v, [(k,_percent(v,total)) for k,v in entries ])
+
+ return render('ga_report/site/index.html')
+
+
+class GaDatasetReport(BaseController):
+ """
+ Displays the pageview and visit count for datasets
+ with options to filter by publisher and time period.
+ """
+ def publisher_csv(self, month):
+ '''
+ Returns a CSV of each publisher with the total number of dataset
+ views & visits.
+ '''
+ c.month = month if not month == 'all' else ''
+ response.headers['Content-Type'] = "text/csv; charset=utf-8"
+ response.headers['Content-Disposition'] = str('attachment; filename=publishers_%s.csv' % (month,))
+
+ writer = csv.writer(response)
+ writer.writerow(["Publisher Title", "Publisher Name", "Views", "Visits", "Period Name"])
+
+ for publisher,view,visit in _get_top_publishers(None):
+ writer.writerow([publisher.title.encode('utf-8'),
+ publisher.name.encode('utf-8'),
+ view,
+ visit,
+ month])
+
+ def dataset_csv(self, id='all', month='all'):
+ '''
+ Returns a CSV with the number of views & visits for each dataset.
+
+ :param id: A Publisher ID or None if you want for all
+ :param month: The time period, or 'all'
+ '''
+ c.month = month if not month == 'all' else ''
+ if id != 'all':
+ c.publisher = model.Group.get(id)
+ if not c.publisher:
+ abort(404, 'A publisher with that name could not be found')
+
+ packages = self._get_packages(c.publisher)
+ response.headers['Content-Type'] = "text/csv; charset=utf-8"
+ response.headers['Content-Disposition'] = \
+ str('attachment; filename=datasets_%s_%s.csv' % (c.publisher_name, month,))
+
+ writer = csv.writer(response)
+ writer.writerow(["Dataset Title", "Dataset Name", "Views", "Visits", "Period Name"])
+
+ for package,view,visit in packages:
+ writer.writerow([package.title.encode('utf-8'),
+ package.name.encode('utf-8'),
+ view,
+ visit,
+ month])
+
+ def publishers(self):
+ '''A list of publishers and the number of views/visits for each'''
+
+ # Get the month details by fetching distinct values and determining the
+ # month names from the values.
+ c.months = _month_details(GA_Url)
+
+ # Work out which month to show, based on query params of the first item
+ c.month = request.params.get('month', '')
+ c.month_desc = 'all months'
+ if c.month:
+ c.month_desc = ''.join([m[1] for m in c.months if m[0]==c.month])
+
+ c.top_publishers = _get_top_publishers()
+
+ return render('ga_report/publisher/index.html')
+
+ def _get_packages(self, publisher=None, count=-1):
+ '''Returns the datasets in order of visits'''
+ if count == -1:
+ count = sys.maxint
+
+ q = model.Session.query(GA_Url)\
+ .filter(GA_Url.url.like('/dataset/%'))
+ if publisher:
+ q = q.filter(GA_Url.department_id==publisher.name)
+ if c.month:
+ q = q.filter(GA_Url.period_name==c.month)
+ q = q.order_by('ga_url.visitors::int desc')
+
+ if c.month:
+ top_packages = []
+ for entry in q.limit(count):
+ package_name = entry.url[len('/dataset/'):]
+ p = model.Package.get(package_name)
+ if p:
+ top_packages.append((p, entry.pageviews, entry.visitors))
+ else:
+ log.warning('Could not find package "%s"', package_name)
+ else:
+ ds = {}
+ for entry in q.limit(count):
+ package_name = entry.url[len('/dataset/'):]
+ p = model.Package.get(package_name)
+ if p:
+ if not p in ds:
+ ds[p] = {'views': 0, 'visits': 0}
+ ds[p]['views'] = ds[p]['views'] + int(entry.pageviews)
+ ds[p]['visits'] = ds[p]['visits'] + int(entry.visitors)
+ else:
+ log.warning('Could not find package "%s"', package_name)
+
+ results = []
+ for k, v in ds.iteritems():
+ results.append((k,v['views'],v['visits']))
+
+ top_packages = sorted(results, key=operator.itemgetter(1), reverse=True)
+ return top_packages
+
+ def read(self):
+ '''
+ Lists the most popular datasets across all publishers
+ '''
+ return self.read_publisher(None)
+
+ def read_publisher(self, id):
+ '''
+ Lists the most popular datasets for a publisher (or across all publishers)
+ '''
+ count = 20
+
+ c.publishers = _get_publishers()
+
+ id = request.params.get('publisher', id)
+ if id and id != 'all':
+ c.publisher = model.Group.get(id)
+ if not c.publisher:
+ abort(404, 'A publisher with that name could not be found')
+ c.publisher_name = c.publisher.name
+ c.top_packages = [] # package, dataset_views in c.top_packages
+
+ # Get the month details by fetching distinct values and determining the
+ # month names from the values.
+ c.months = _month_details(GA_Url)
+
+ # Work out which month to show, based on query params of the first item
+ c.month = request.params.get('month', '')
+ if not c.month:
+ c.month_desc = 'all months'
+ else:
+ c.month_desc = ''.join([m[1] for m in c.months if m[0]==c.month])
+
+ c.publisher_page_views = 0
+ q = model.Session.query(GA_Url).\
+ filter(GA_Url.url=='/publisher/%s' % c.publisher_name)
+ if c.month:
+ entry = q.filter(GA_Url.period_name==c.month).first()
+ c.publisher_page_views = entry.pageviews if entry else 0
+ else:
+ for e in q.all():
+ c.publisher_page_views = c.publisher_page_views + int(e.pageviews)
+
+ c.top_packages = self._get_packages(c.publisher, 20)
+
+ return render('ga_report/publisher/read.html')
+
+def _get_top_publishers(limit=20):
+ '''
+ Returns a list of the top 20 publishers by dataset visits.
+ (The number to show can be varied with 'limit')
+ '''
+ connection = model.Session.connection()
+ q = """
+ select department_id, sum(pageviews::int) views, sum(visitors::int) visits
+ from ga_url
+ where department_id <> ''"""
+ if c.month:
+ q = q + """
+ and period_name=%s
+ """
+ q = q + """
+ group by department_id order by visits desc
+ """
+ if limit:
+ q = q + " limit %s;" % (limit)
+
+ # Add this back (before and period_name =%s) if you want to ignore publisher
+ # homepage views
+ # and not url like '/publisher/%%'
+
+ top_publishers = []
+ res = connection.execute(q, c.month)
+
+ for row in res:
+ g = model.Group.get(row[0])
+ if g:
+ top_publishers.append((g, row[1], row[2]))
+ return top_publishers
+
+def _get_publishers():
+ '''
+ Returns a list of all publishers. Each item is a tuple:
+ (names, title)
+ '''
+ publishers = []
+ for pub in model.Session.query(model.Group).\
+ filter(model.Group.type=='publisher').\
+ filter(model.Group.state=='active').\
+ order_by(model.Group.name):
+ publishers.append((pub.name, pub.title))
+ return publishers
+
+def _percent(num, total):
+ p = 100 * float(num)/float(total)
+ return "%.2f%%" % round(p, 2)
+
--- /dev/null
+++ b/ckanext/ga_report/download_analytics.py
@@ -1,1 +1,396 @@
-
+import os
+import logging
+import datetime
+import collections
+from pylons import config
+
+import ga_model
+
+#from ga_client import GA
+
+log = logging.getLogger('ckanext.ga-report')
+
+FORMAT_MONTH = '%Y-%m'
+MIN_VIEWS = 50
+MIN_VISITS = 20
+
+class DownloadAnalytics(object):
+ '''Downloads and stores analytics info'''
+
+ def __init__(self, service=None, profile_id=None, delete_first=False):
+ self.period = config['ga-report.period']
+ self.service = service
+ self.profile_id = profile_id
+ self.delete_first = delete_first
+
+ def specific_month(self, date):
+ import calendar
+
+ first_of_this_month = datetime.datetime(date.year, date.month, 1)
+ _, last_day_of_month = calendar.monthrange(int(date.year), int(date.month))
+ last_of_this_month = datetime.datetime(date.year, date.month, last_day_of_month)
+ periods = ((date.strftime(FORMAT_MONTH),
+ last_day_of_month,
+ first_of_this_month, last_of_this_month),)
+ self.download_and_store(periods)
+
+
+ def latest(self):
+ if self.period == 'monthly':
+ # from first of this month to today
+ now = datetime.datetime.now()
+ first_of_this_month = datetime.datetime(now.year, now.month, 1)
+ periods = ((now.strftime(FORMAT_MONTH),
+ now.day,
+ first_of_this_month, now),)
+ else:
+ raise NotImplementedError
+ self.download_and_store(periods)
+
+
+ def for_date(self, for_date):
+ assert isinstance(since_date, datetime.datetime)
+ periods = [] # (period_name, period_complete_day, start_date, end_date)
+ if self.period == 'monthly':
+ first_of_the_months_until_now = []
+ year = for_date.year
+ month = for_date.month
+ now = datetime.datetime.now()
+ first_of_this_month = datetime.datetime(now.year, now.month, 1)
+ while True:
+ first_of_the_month = datetime.datetime(year, month, 1)
+ if first_of_the_month == first_of_this_month:
+ periods.append((now.strftime(FORMAT_MONTH),
+ now.day,
+ first_of_this_month, now))
+ break
+ elif first_of_the_month < first_of_this_month:
+ in_the_next_month = first_of_the_month + datetime.timedelta(40)
+ last_of_the_month = datetime.datetime(in_the_next_month.year,
+ in_the_next_month.month, 1)\
+ - datetime.timedelta(1)
+ periods.append((now.strftime(FORMAT_MONTH), 0,
+ first_of_the_month, last_of_the_month))
+ else:
+ # first_of_the_month has got to the future somehow
+ break
+ month += 1
+ if month > 12:
+ year += 1
+ month = 1
+ else:
+ raise NotImplementedError
+ self.download_and_store(periods)
+
+ @staticmethod
+ def get_full_period_name(period_name, period_complete_day):
+ if period_complete_day:
+ return period_name + ' (up to %ith)' % period_complete_day
+ else:
+ return period_name
+
+
+ def download_and_store(self, periods):
+ for period_name, period_complete_day, start_date, end_date in periods:
+ if self.delete_first:
+ log.info('Deleting existing Analytics for 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'))
+
+ data = self.download(start_date, end_date, '~/dataset/[a-z0-9-_]+')
+ log.info('Storing Dataset Analytics for period "%s"',
+ self.get_full_period_name(period_name, period_complete_day))
+ self.store(period_name, period_complete_day, data, )
+
+ data = self.download(start_date, end_date, '~/publisher/[a-z0-9-_]+')
+ log.info('Storing Publisher Analytics for period "%s"',
+ self.get_full_period_name(period_name, period_complete_day))
+ self.store(period_name, period_complete_day, data,)
+
+ ga_model.update_publisher_stats(period_name) # about 30 seconds.
+ self.sitewide_stats( period_name )
+
+ self.update_social_info(period_name, start_date, end_date)
+
+ def update_social_info(self, period_name, start_date, end_date):
+ start_date = start_date.strftime('%Y-%m-%d')
+ end_date = end_date.strftime('%Y-%m-%d')
+ query = 'ga:hasSocialSourceReferral=~Yes$'
+ metrics = 'ga:entrances'
+ sort = '-ga:entrances'
+
+ # Supported query params at
+ # https://developers.google.com/analytics/devguides/reporting/core/v3/reference
+ results = self.service.data().ga().get(
+ ids='ga:' + self.profile_id,
+ filters=query,
+ start_date=start_date,
+ metrics=metrics,
+ sort=sort,
+ dimensions="ga:landingPagePath,ga:socialNetwork",
+ max_results=10000,
+ end_date=end_date).execute()
+ data = collections.defaultdict(list)
+ rows = results.get('rows',[])
+ for row in rows:
+ from ga_model import _normalize_url
+ data[_normalize_url(row[0])].append( (row[1], int(row[2]),) )
+ ga_model.update_social(period_name, data)
+
+
+ def download(self, start_date, end_date, path='~/dataset/[a-z0-9-_]+'):
+ '''Get data from GA for a given time period'''
+ start_date = start_date.strftime('%Y-%m-%d')
+ end_date = end_date.strftime('%Y-%m-%d'