gitphp 0.2.9.1 :: ckanext-ga-report.git/commitdiff

A fix for the copy and paste error that accounts for #211

3 files changed: (show all)
ckanext/ga_report/ga_model.py
ckanext/ga_report/helpers.py
ckanext/ga_report/templates/ga_report/notes.html

file:a/ckanext/ga_report/ga_model.py -> file:b/ckanext/ga_report/ga_model.py

import re	import re
import uuid	import uuid

from sqlalchemy import Table, Column, MetaData, ForeignKey	from sqlalchemy import Table, Column, MetaData, ForeignKey
from sqlalchemy import types	from sqlalchemy import types
from sqlalchemy.sql import select	from sqlalchemy.sql import select
from sqlalchemy.orm import mapper, relation	from sqlalchemy.orm import mapper, relation
from sqlalchemy import func	from sqlalchemy import func

import ckan.model as model	import ckan.model as model
from ckan.lib.base import *	from ckan.lib.base import *

log = __import__('logging').getLogger(__name__)	log = __import__('logging').getLogger(__name__)

def make_uuid():	def make_uuid():
return unicode(uuid.uuid4())	return unicode(uuid.uuid4())

metadata = MetaData()	metadata = MetaData()

class GA_Url(object):	class GA_Url(object):

def __init__(self, **kwargs):	def __init__(self, **kwargs):
for k,v in kwargs.items():	for k,v in kwargs.items():
setattr(self, k, v)	setattr(self, k, v)

url_table = Table('ga_url', metadata,	url_table = Table('ga_url', metadata,
Column('id', types.UnicodeText, primary_key=True,	Column('id', types.UnicodeText, primary_key=True,
default=make_uuid),	default=make_uuid),
Column('period_name', types.UnicodeText),	Column('period_name', types.UnicodeText),
Column('period_complete_day', types.Integer),	Column('period_complete_day', types.Integer),
Column('pageviews', types.UnicodeText),	Column('pageviews', types.UnicodeText),
Column('visits', types.UnicodeText),	Column('visits', types.UnicodeText),
Column('url', types.UnicodeText),	Column('url', types.UnicodeText),
Column('department_id', types.UnicodeText),	Column('department_id', types.UnicodeText),
Column('package_id', types.UnicodeText),	Column('package_id', types.UnicodeText),
)	)
mapper(GA_Url, url_table)	mapper(GA_Url, url_table)


class GA_Stat(object):	class GA_Stat(object):

def __init__(self, **kwargs):	def __init__(self, **kwargs):
for k,v in kwargs.items():	for k,v in kwargs.items():
setattr(self, k, v)	setattr(self, k, v)

stat_table = Table('ga_stat', metadata,	stat_table = Table('ga_stat', metadata,
Column('id', types.UnicodeText, primary_key=True,	Column('id', types.UnicodeText, primary_key=True,
default=make_uuid),	default=make_uuid),
Column('period_name', types.UnicodeText),	Column('period_name', types.UnicodeText),
Column('period_complete_day', types.UnicodeText),	Column('period_complete_day', types.UnicodeText),
Column('stat_name', types.UnicodeText),	Column('stat_name', types.UnicodeText),
Column('key', types.UnicodeText),	Column('key', types.UnicodeText),
Column('value', types.UnicodeText), )	Column('value', types.UnicodeText), )
mapper(GA_Stat, stat_table)	mapper(GA_Stat, stat_table)


class GA_Publisher(object):	class GA_Publisher(object):

def __init__(self, **kwargs):	def __init__(self, **kwargs):
for k,v in kwargs.items():	for k,v in kwargs.items():
setattr(self, k, v)	setattr(self, k, v)

pub_table = Table('ga_publisher', metadata,	pub_table = Table('ga_publisher', metadata,
Column('id', types.UnicodeText, primary_key=True,	Column('id', types.UnicodeText, primary_key=True,
default=make_uuid),	default=make_uuid),
Column('period_name', types.UnicodeText),	Column('period_name', types.UnicodeText),
Column('publisher_name', types.UnicodeText),	Column('publisher_name', types.UnicodeText),
Column('views', types.UnicodeText),	Column('views', types.UnicodeText),
Column('visits', types.UnicodeText),	Column('visits', types.UnicodeText),
Column('toplevel', types.Boolean, default=False),	Column('toplevel', types.Boolean, default=False),
Column('subpublishercount', types.Integer, default=0),	Column('subpublishercount', types.Integer, default=0),
Column('parent', types.UnicodeText),	Column('parent', types.UnicodeText),
)	)
mapper(GA_Publisher, pub_table)	mapper(GA_Publisher, pub_table)


class GA_ReferralStat(object):	class GA_ReferralStat(object):

def __init__(self, **kwargs):	def __init__(self, **kwargs):
for k,v in kwargs.items():	for k,v in kwargs.items():
setattr(self, k, v)	setattr(self, k, v)

referrer_table = Table('ga_referrer', metadata,	referrer_table = Table('ga_referrer', metadata,
Column('id', types.UnicodeText, primary_key=True,	Column('id', types.UnicodeText, primary_key=True,
default=make_uuid),	default=make_uuid),
Column('period_name', types.UnicodeText),	Column('period_name', types.UnicodeText),
Column('source', types.UnicodeText),	Column('source', types.UnicodeText),
Column('url', types.UnicodeText),	Column('url', types.UnicodeText),
Column('count', types.Integer),	Column('count', types.Integer),
)	)
mapper(GA_ReferralStat, referrer_table)	mapper(GA_ReferralStat, referrer_table)



def init_tables():	def init_tables():
metadata.create_all(model.meta.engine)	metadata.create_all(model.meta.engine)


cached_tables = {}	cached_tables = {}


def get_table(name):	def get_table(name):
if name not in cached_tables:	if name not in cached_tables:
meta = MetaData()	meta = MetaData()
meta.reflect(bind=model.meta.engine)	meta.reflect(bind=model.meta.engine)
table = meta.tables[name]	table = meta.tables[name]
cached_tables[name] = table	cached_tables[name] = table
return cached_tables[name]	return cached_tables[name]


def _normalize_url(url):	def _normalize_url(url):
'''Strip off the hostname etc. Do this before storing it.	'''Strip off the hostname etc. Do this before storing it.

>>> normalize_url('http://data.gov.uk/dataset/weekly_fuel_prices')	>>> normalize_url('http://data.gov.uk/dataset/weekly_fuel_prices')
'/dataset/weekly_fuel_prices'	'/dataset/weekly_fuel_prices'
'''	'''
return '/' + '/'.join(url.split('/')[3:])	return '/' + '/'.join(url.split('/')[3:])


def _get_package_and_publisher(url):	def _get_package_and_publisher(url):
# e.g. /dataset/fuel_prices	# e.g. /dataset/fuel_prices
# e.g. /dataset/fuel_prices/resource/e63380d4	# e.g. /dataset/fuel_prices/resource/e63380d4
dataset_match = re.match('/dataset/([^/]+)(/.*)?', url)	dataset_match = re.match('/dataset/([^/]+)(/.*)?', url)
if dataset_match:	if dataset_match:
dataset_ref = dataset_match.groups()[0]	dataset_ref = dataset_match.groups()[0]
dataset = model.Package.get(dataset_ref)	dataset = model.Package.get(dataset_ref)
if dataset:	if dataset:
publisher_groups = dataset.get_groups('publisher')	publisher_groups = dataset.get_groups('publisher')
if publisher_groups:	if publisher_groups:
return dataset_ref,publisher_groups[0].name	return dataset_ref,publisher_groups[0].name
return dataset_ref, None	return dataset_ref, None
else:	else:
publisher_match = re.match('/publisher/([^/]+)(/.*)?', url)	publisher_match = re.match('/publisher/([^/]+)(/.*)?', url)
if publisher_match:	if publisher_match:
return None, publisher_match.groups()[0]	return None, publisher_match.groups()[0]
return None, None	return None, None

def update_sitewide_stats(period_name, stat_name, data, period_complete_day):	def update_sitewide_stats(period_name, stat_name, data, period_complete_day):
for k,v in data.iteritems():	for k,v in data.iteritems():
item = model.Session.query(GA_Stat).\	item = model.Session.query(GA_Stat).\
filter(GA_Stat.period_name==period_name).\	filter(GA_Stat.period_name==period_name).\
filter(GA_Stat.key==k).\	filter(GA_Stat.key==k).\
filter(GA_Stat.stat_name==stat_name).first()	filter(GA_Stat.stat_name==stat_name).first()
if item:	if item:
item.period_name = period_name	item.period_name = period_name
item.key = k	item.key = k
item.value = v	item.value = v
item.period_complete_day = period_complete_day	item.period_complete_day = period_complete_day
model.Session.add(item)	model.Session.add(item)
else:	else:
# create the row	# create the row
values = {'id': make_uuid(),	values = {'id': make_uuid(),
'period_name': period_name,	'period_name': period_name,
'period_complete_day': period_complete_day,	'period_complete_day': period_complete_day,
'key': k,	'key': k,
'value': v,	'value': v,
'stat_name': stat_name	'stat_name': stat_name
}	}
model.Session.add(GA_Stat(**values))	model.Session.add(GA_Stat(**values))
model.Session.commit()	model.Session.commit()


def pre_update_url_stats(period_name):	def pre_update_url_stats(period_name):
q = model.Session.query(GA_Url).\	q = model.Session.query(GA_Url).\
filter(GA_Url.period_name==period_name)	filter(GA_Url.period_name==period_name)
log.debug("Deleting %d '%s' records" % (q.count(), period_name))	log.debug("Deleting %d '%s' records" % (q.count(), period_name))
q.delete()	q.delete()

q = model.Session.query(GA_Url).\	q = model.Session.query(GA_Url).\
filter(GA_Url.period_name == 'All')	filter(GA_Url.period_name == 'All')
log.debug("Deleting %d 'All' records..." % q.count())	log.debug("Deleting %d 'All' records..." % q.count())
q.delete()	q.delete()

model.Session.flush()	model.Session.flush()
model.Session.commit()	model.Session.commit()
model.repo.commit_and_remove()	model.repo.commit_and_remove()
log.debug('...done')	log.debug('...done')

def post_update_url_stats():	def post_update_url_stats():

""" Check the distinct url field in ga_url and make sure	""" Check the distinct url field in ga_url and make sure
it has an All record. If not then create one.	it has an All record. If not then create one.

After running this then every URL should have an All	After running this then every URL should have an All
record regardless of whether the URL has an entry for	record regardless of whether the URL has an entry for
the month being currently processed.	the month being currently processed.
"""	"""
log.debug('Post-processing "All" records...')	log.debug('Post-processing "All" records...')
query = """select url, pageviews::int, visits::int	query = """select url, pageviews::int, visits::int
from ga_url	from ga_url
where url not in (select url from ga_url where period_name ='All')"""	where url not in (select url from ga_url where period_name ='All')"""
connection = model.Session.connection()	connection = model.Session.connection()
res = connection.execute(query)	res = connection.execute(query)

views, visits = {}, {}	views, visits = {}, {}
# url, views, visits	# url, views, visits
for row in res:	for row in res:
views[row[0]] = views.get(row[0], 0) + row[1]	views[row[0]] = views.get(row[0], 0) + row[1]
visits[row[0]] = visits.get(row[0], 0) + row[2]	visits[row[0]] = visits.get(row[0], 0) + row[2]

progress_total = len(views.keys())	progress_total = len(views.keys())
progress_count = 0	progress_count = 0
for key in views.keys():	for key in views.keys():
progress_count += 1	progress_count += 1
if progress_count % 100 == 0:	if progress_count % 100 == 0:
log.debug('.. %d/%d done so far', progress_count, progress_total)	log.debug('.. %d/%d done so far', progress_count, progress_total)

package, publisher = _get_package_and_publisher(key)	package, publisher = _get_package_and_publisher(key)

values = {'id': make_uuid(),	values = {'id': make_uuid(),
'period_name': "All",	'period_name': "All",
'period_complete_day': 0,	'period_complete_day': 0,
'url': key,	'url': key,
'pageviews': views[key],	'pageviews': views[key],
'visits': visits[key],	'visits': visits[key],
'department_id': publisher,	'department_id': publisher,
'package_id': publisher	'package_id': package
}	}
model.Session.add(GA_Url(**values))	model.Session.add(GA_Url(**values))
model.Session.commit()	model.Session.commit()
log.debug('..done')	log.debug('..done')


def update_url_stats(period_name, period_complete_day, url_data):	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,	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	stores them in GA_Url under the period and recalculates the totals for
the 'All' period.	the 'All' period.
'''	'''
progress_total = len(progress_data)	progress_total = len(url_data)
progress_count = 0	progress_count = 0
for url, views, visits in url_data:	for url, views, visits in url_data:
progress_count += 1	progress_count += 1
if progress_count % 100 == 0:	if progress_count % 100 == 0:
log.debug('.. %d/%d done so far', progress_count, progress_total)	log.debug('.. %d/%d done so far', progress_count, progress_total)

package, publisher = _get_package_and_publisher(url)	package, publisher = _get_package_and_publisher(url)

item = model.Session.query(GA_Url).\	item = model.Session.query(GA_Url).\
filter(GA_Url.period_name==period_name).\	filter(GA_Url.period_name==period_name).\
filter(GA_Url.url==url).first()	filter(GA_Url.url==url).first()
if item:	if item:
item.pageviews = item.pageviews + views	item.pageviews = item.pageviews + views
item.visits = item.visits + visits	item.visits = item.visits + visits
if not item.package_id:	if not item.package_id:
item.package_id = package	item.package_id = package
if not item.department_id:	if not item.department_id:
item.department_id = publisher	item.department_id = publisher
model.Session.add(item)	model.Session.add(item)
else:	else:
values = {'id': make_uuid(),	values = {'id': make_uuid(),
'period_name': period_name,	'period_name': period_name,
'period_complete_day': period_complete_day,	'period_complete_day': period_complete_day,
'url': url,	'url': url,
'pageviews': views,	'pageviews': views,
'visits': visits,	'visits': visits,
'department_id': publisher,	'department_id': publisher,
'package_id': package	'package_id': package
}	}
model.Session.add(GA_Url(**values))	model.Session.add(GA_Url(**values))
model.Session.commit()	model.Session.commit()

if package:	if package:
old_pageviews, old_visits = 0, 0	old_pageviews, old_visits = 0, 0
old = model.Session.query(GA_Url).\	old = model.Session.query(GA_Url).\
filter(GA_Url.period_name=='All').\	filter(GA_Url.period_name=='All').\
filter(GA_Url.url==url).all()	filter(GA_Url.url==url).all()
old_pageviews = sum([int(o.pageviews) for o in old])	old_pageviews = sum([int(o.pageviews) for o in old])
old_visits = sum([int(o.visits) for o in old])	old_visits = sum([int(o.visits) for o in old])

entries = model.Session.query(GA_Url).\	entries = model.Session.query(GA_Url).\
filter(GA_Url.period_name!='All').\	filter(GA_Url.period_name!='All').\
filter(GA_Url.url==url).all()	filter(GA_Url.url==url).all()
values = {'id': make_uuid(),	values = {'id': make_uuid(),
'period_name': 'All',	'period_name': 'All',
'period_complete_day': 0,	'period_complete_day': 0,
'url': url,	'url': url,
'pageviews': sum([int(e.pageviews) for e in entries]) + int(old_pageviews),	'pageviews': sum([int(e.pageviews) for e in entries]) + int(old_pageviews),
'visits': sum([int(e.visits or 0) for e in entries]) + int(old_visits),	'visits': sum([int(e.visits or 0) for e in entries]) + int(old_visits),
'department_id': publisher,	'department_id': publisher,
'package_id': package	'package_id': package
}	}

model.Session.add(GA_Url(**values))	model.Session.add(GA_Url(**values))
model.Session.commit()	model.Session.commit()




def update_social(period_name, data):	def update_social(period_name, data):
# Clean up first.	# Clean up first.
model.Session.query(GA_ReferralStat).\	model.Session.query(GA_ReferralStat).\
filter(GA_ReferralStat.period_name==period_name).delete()	filter(GA_ReferralStat.period_name==period_name).delete()

for url,data in data.iteritems():	for url,data in data.iteritems():
for entry in data:	for entry in data:
source = entry[0]	source = entry[0]
count = entry[1]	count = entry[1]

item = model.Session.query(GA_ReferralStat).\	item = model.Session.query(GA_ReferralStat).\
filter(GA_ReferralStat.period_name==period_name).\	filter(GA_ReferralStat.period_name==period_name).\
filter(GA_ReferralStat.source==source).\	filter(GA_ReferralStat.source==source).\
filter(GA_ReferralStat.url==url).first()	filter(GA_ReferralStat.url==url).first()
if item:	if item:
item.count = item.count + count	item.count = item.count + count
model.Session.add(item)	model.Session.add(item)
else:	else:
# create the row	# create the row
values = {'id': make_uuid(),	values = {'id': make_uuid(),
'period_name': period_name,	'period_name': period_name,
'source': source,	'source': source,
'url': url,	'url': url,
'count': count,	'count': count,
}	}
model.Session.add(GA_ReferralStat(**values))	model.Session.add(GA_ReferralStat(**values))
model.Session.commit()	model.Session.commit()

def update_publisher_stats(period_name):	def update_publisher_stats(period_name):
"""	"""
Updates the publisher stats from the data retrieved for /dataset/*	Updates the publisher stats from the data retrieved for /dataset/*
and /publisher/*. Will run against each dataset and generates the	and /publisher/*. Will run against each dataset and generates the
totals for the entire tree beneath each publisher.	totals for the entire tree beneath each publisher.
"""	"""
toplevel = get_top_level()	toplevel = get_top_level()
publishers = model.Session.query(model.Group).\	publishers = model.Session.query(model.Group).\
filter(model.Group.type=='publisher').\	filter(model.Group.type=='publisher').\
filter(model.Group.state=='active').all()	filter(model.Group.state=='active').all()
for publisher in publishers:	for publisher in publishers:
views, visits, subpub = update_publisher(period_name, publisher, publisher.name)	views, visits, subpub = update_publisher(period_name, publisher, publisher.name)
parent, parents = '', publisher.get_groups('publisher')	parent, parents = '', publisher.get_groups('publisher')
if parents:	if parents:
parent = parents[0].name	parent = parents[0].name
item = model.Session.query(GA_Publisher).\	item = model.Session.query(GA_Publisher).\
filter(GA_Publisher.period_name==period_name).\	filter(GA_Publisher.period_name==period_name).\
filter(GA_Publisher.publisher_name==publisher.name).first()	filter(GA_Publisher.publisher_name==publisher.name).first()
if item:	if item:
item.views = views	item.views = views
item.visits = visits	item.visits = visits
item.publisher_name = publisher.name	item.publisher_name = publisher.name
item.toplevel = publisher in toplevel	item.toplevel = publisher in toplevel
item.subpublishercount = subpub	item.subpublishercount = subpub
item.parent = parent	item.parent = parent
model.Session.add(item)	model.Session.add(item)
else:	else:
# create the row	# create the row
values = {'id': make_uuid(),	values = {'id': make_uuid(),
'period_name': period_name,	'period_name': period_name,
'publisher_name': publisher.name,	'publisher_name': publisher.name,
'views': views,	'views': views,
'visits': visits,	'visits': visits,
'toplevel': publisher in toplevel,	'toplevel': publisher in toplevel,
'subpublishercount': subpub,	'subpublishercount': subpub,
'parent': parent	'parent': parent
}	}
model.Session.add(GA_Publisher(**values))	model.Session.add(GA_Publisher(**values))
model.Session.commit()	model.Session.commit()


def update_publisher(period_name, pub, part=''):	def update_publisher(period_name, pub, part=''):
views,visits,subpub = 0, 0, 0	views,visits,subpub = 0, 0, 0
for publisher in go_down_tree(pub):	for publisher in go_down_tree(pub):
subpub = subpub + 1	subpub = subpub + 1
items = model.Session.query(GA_Url).\	items = model.Session.query(GA_Url).\
filter(GA_Url.period_name==period_name).\	filter(GA_Url.period_name==period_name).\
filter(GA_Url.department_id==publisher.name).all()	filter(GA_Url.department_id==publisher.name).all()
for item in items:	for item in items:
views = views + int(item.pageviews)	views = views + int(item.pageviews)
visits = visits + int(item.visits)	visits = visits + int(item.visits)

return views, visits, (subpub-1)	return views, visits, (subpub-1)


def get_top_level():	def get_top_level():
'''Returns the top level publishers.'''	'''Returns the top level publishers.'''
return model.Session.query(model.Group).\	return model.Session.query(model.Group).\
outerjoin(model.Member, model.Member.table_id == model.Group.id and \	outerjoin(model.Member, model.Member.table_id == model.Group.id and \
model.Member.table_name == 'group' and \	model.Member.table_name == 'group' and \
model.Member.state == 'active').\	model.Member.state == 'active').\
filter(model.Member.id==None).\	filter(model.Member.id==None).\
filter(model.Group.type=='publisher').\	filter(model.Group.type=='publisher').\
order_by(model.Group.name).all()	order_by(model.Group.name).all()

def get_children(publisher):	def get_children(publisher):
'''Finds child publishers for the given publisher (object). (Not recursive)'''	'''Finds child publishers for the given publisher (object). (Not recursive)'''
from ckan.model.group import HIERARCHY_CTE	from ckan.model.group import HIERARCHY_CTE
return model.Session.query(model.Group).\	return model.Session.query(model.Group).\
from_statement(HIERARCHY_CTE).params(id=publisher.id, type='publisher').\	from_statement(HIERARCHY_CTE).params(id=publisher.id, type='publisher').\
all()	all()

def go_down_tree(publisher):	def go_down_tree(publisher):
'''Provided with a publisher object, it walks down the hierarchy and yields each publisher,	'''Provided with a publisher object, it walks down the hierarchy and yields each publisher,
including the one you supply.'''	including the one you supply.'''
yield publisher	yield publisher
for child in get_children(publisher):	for child in get_children(publisher):
for grandchild in go_down_tree(child):	for grandchild in go_down_tree(child):
yield grandchild	yield grandchild

def delete(period_name):	def delete(period_name):
'''	'''
Deletes table data for the specified period, or specify 'all'	Deletes table data for the specified period, or specify 'all'
for all periods.	for all periods.
'''	'''
for object_type in (GA_Url, GA_Stat, GA_Publisher, GA_ReferralStat):	for object_type in (GA_Url, GA_Stat, GA_Publisher, GA_ReferralStat):
q = model.Session.query(object_type)	q = model.Session.query(object_type)
if period_name != 'All':	if period_name != 'All':
q = q.filter_by(period_name=period_name)	q = q.filter_by(period_name=period_name)
q.delete()	q.delete()
model.repo.commit_and_remove()	model.repo.commit_and_remove()

def get_score_for_dataset(dataset_name):	def get_score_for_dataset(dataset_name):
'''	'''
Returns a "current popularity" score for a dataset,	Returns a "current popularity" score for a dataset,
based on how many views it has had recently.	based on how many views it has had recently.
'''	'''
import datetime	import datetime
now = datetime.datetime.now()	now = datetime.datetime.now()
last_month = now - datetime.timedelta(days=30)	last_month = now - datetime.timedelta(days=30)
period_names = ['%s-%02d' % (last_month.year, last_month.month),	period_names = ['%s-%02d' % (last_month.year, last_month.month),
'%s-%02d' % (now.year, now.month),	'%s-%02d' % (now.year, now.month),
]	]

score = 0	score = 0
for period_name in period_names:	for period_name in period_names:
score /= 2 # previous periods are discounted by 50%	score /= 2 # previous periods are discounted by 50%
entry = model.Session.query(GA_Url)\	entry = model.Session.query(GA_Url)\
.filter(GA_Url.period_name==period_name)\	.filter(GA_Url.period_name==period_name)\
.filter(GA_Url.package_id==dataset_name).first()	.filter(GA_Url.package_id==dataset_name).first()
# score	# score
if entry:	if entry:
views = float(entry.pageviews)	views = float(entry.pageviews)
if entry.period_complete_day:	if entry.period_complete_day:
views_per_day = views / entry.period_complete_day	views_per_day = views / entry.period_complete_day
else:	else:
views_per_day = views / 15 # guess	views_per_day = views / 15 # guess
score += views_per_day	score += views_per_day

score = int(score * 100)	score = int(score * 100)
log.debug('Popularity %s: %s', score, dataset_name)	log.debug('Popularity %s: %s', score, dataset_name)
return score	return score

file:a/ckanext/ga_report/helpers.py -> file:b/ckanext/ga_report/helpers.py

import logging	import logging
import operator	import operator

import ckan.lib.base as base	import ckan.lib.base as base
import ckan.model as model	import ckan.model as model
from ckan.logic import get_action	from ckan.logic import get_action

from ckanext.ga_report.ga_model import GA_Url, GA_Publisher	from ckanext.ga_report.ga_model import GA_Url, GA_Publisher
from ckanext.ga_report.controller import _get_publishers	from ckanext.ga_report.controller import _get_publishers
_log = logging.getLogger(__name__)	_log = logging.getLogger(__name__)

def popular_datasets(count=10):	def popular_datasets(count=10):
import random	import random

publisher = None	publisher = None
publishers = _get_publishers(30)	publishers = _get_publishers(30)
total = len(publishers)	total = len(publishers)
while not publisher or not datasets:	while not publisher or not datasets:
rand = random.randrange(0, total)	rand = random.randrange(0, total)
publisher = publishers[rand][0]	publisher = publishers[rand][0]
if not publisher.state == 'active':	if not publisher.state == 'active':
publisher = None	publisher = None
continue	continue
datasets = _datasets_for_publisher(publisher, 10)[:count]	datasets = _datasets_for_publisher(publisher, 10)[:count]

ctx = {	ctx = {
'datasets': datasets,	'datasets': datasets,
'publisher': publisher	'publisher': publisher
}	}
return base.render_snippet('ga_report/ga_popular_datasets.html', **ctx)	return base.render_snippet('ga_report/ga_popular_datasets.html', **ctx)

def single_popular_dataset(top=20):	def single_popular_dataset(top=20):
'''Returns a random dataset from the most popular ones.	'''Returns a random dataset from the most popular ones.

:param top: the number of top datasets to select from	:param top: the number of top datasets to select from
'''	'''
import random	import random

top_datasets = model.Session.query(GA_Url).\	top_datasets = model.Session.query(GA_Url).\
filter(GA_Url.url.like('/dataset/%')).\	filter(GA_Url.url.like('/dataset/%')).\
order_by('ga_url.pageviews::int desc')	order_by('ga_url.pageviews::int desc')
num_top_datasets = top_datasets.count()	num_top_datasets = top_datasets.count()

dataset = None	dataset = None
if num_top_datasets:	if num_top_datasets:
count = 0	count = 0
while not dataset:	while not dataset:
rand = random.randrange(0, min(top, num_top_datasets))	rand = random.randrange(0, min(top, num_top_datasets))
ga_url = top_datasets[rand]	ga_url = top_datasets[rand]
dataset = model.Package.get(ga_url.url[len('/dataset/'):])	dataset = model.Package.get(ga_url.url[len('/dataset/'):])
if dataset and not dataset.state == 'active':	if dataset and not dataset.state == 'active':
dataset = None	dataset = None
# When testing, it is possible that top datasets are not available	# When testing, it is possible that top datasets are not available
# so only go round this loop a few times before falling back on	# so only go round this loop a few times before falling back on
# a random dataset.	# a random dataset.
count += 1	count += 1
if count > 10:	if count > 10:
break	break
if not dataset:	if not dataset:
# fallback	# fallback
dataset = model.Session.query(model.Package)\	dataset = model.Session.query(model.Package)\
.filter_by(state='active').first()	.filter_by(state='active').first()
if not dataset:	if not dataset:
return None	return None
dataset_dict = get_action('package_show')({'model': model,	dataset_dict = get_action('package_show')({'model': model,
'session': model.Session,	'session': model.Session,
'validate': False},	'validate': False},
{'id':dataset.id})	{'id':dataset.id})
return dataset_dict	return dataset_dict

def single_popular_dataset_html(top=20):	def single_popular_dataset_html(top=20):
dataset_dict = single_popular_dataset(top)	dataset_dict = single_popular_dataset(top)
groups = package.get('groups', [])	groups = package.get('groups', [])
publishers = [ g for g in groups if g.get('type') == 'publisher' ]	publishers = [ g for g in groups if g.get('type') == 'publisher' ]
publisher = publishers[0] if publishers else {'name':'', 'title': ''}	publisher = publishers[0] if publishers else {'name':'', 'title': ''}
context = {	context = {
'dataset': dataset_dict,	'dataset': dataset_dict,
'publisher': publisher_dict	'publisher': publisher_dict
}	}
return base.render_snippet('ga_report/ga_popular_single.html', **context)	return base.render_snippet('ga_report/ga_popular_single.html', **context)


def most_popular_datasets(publisher, count=20):	def most_popular_datasets(publisher, count=20):

if not publisher:	if not publisher:
_log.error("No valid publisher passed to 'most_popular_datasets'")	_log.error("No valid publisher passed to 'most_popular_datasets'")
return ""	return ""

results = _datasets_for_publisher(publisher, count)	results = _datasets_for_publisher(publisher, count)

ctx = {	ctx = {
'dataset_count': len(results),	'dataset_count': len(results),
'datasets': results,	'datasets': results,

'publisher': publisher	'publisher': publisher
}	}

return base.render_snippet('ga_report/publisher/popular.html', **ctx)	return base.render_snippet('ga_report/publisher/popular.html', **ctx)

def _datasets_for_publisher(publisher, count):	def _datasets_for_publisher(publisher, count):
datasets = {}	datasets = {}
entries = model.Session.query(GA_Url).\	entries = model.Session.query(GA_Url).\
filter(GA_Url.department_id==publisher.name).\	filter(GA_Url.department_id==publisher.name).\
filter(GA_Url.url.like('/dataset/%')).\	filter(GA_Url.url.like('/dataset/%')).\
order_by('ga_url.pageviews::int desc').all()	order_by('ga_url.pageviews::int desc').all()
for entry in entries:	for entry in entries:
if len(datasets) < count:	if len(datasets) < count:
p = model.Package.get(entry.url[len('/dataset/'):])	p = model.Package.get(entry.url[len('/dataset/'):])
	if not p:
	_log.warning("Could not find Package for {url}".format(url=entry.url))
	continue

if not p in datasets:	if not p in datasets:
datasets[p] = {'views':0, 'visits': 0}	datasets[p] = {'views':0, 'visits': 0}
datasets[p]['views'] = datasets[p]['views'] + int(entry.pageviews)	datasets[p]['views'] = datasets[p]['views'] + int(entry.pageviews)
datasets[p]['visits'] = datasets[p]['visits'] + int(entry.visits)	datasets[p]['visits'] = datasets[p]['visits'] + int(entry.visits)

results = []	results = []
for k, v in datasets.iteritems():	for k, v in datasets.iteritems():
results.append((k,v['views'],v['visits']))	results.append((k,v['views'],v['visits']))

return sorted(results, key=operator.itemgetter(1), reverse=True)	return sorted(results, key=operator.itemgetter(1), reverse=True)

file:a/ckanext/ga_report/templates/ga_report/notes.html -> file:b/ckanext/ga_report/templates/ga_report/notes.html

<html xmlns:py="http://genshi.edgewall.org/"	<html xmlns:py="http://genshi.edgewall.org/"
xmlns:i18n="http://genshi.edgewall.org/i18n"	xmlns:i18n="http://genshi.edgewall.org/i18n"
xmlns:xi="http://www.w3.org/2001/XInclude"	xmlns:xi="http://www.w3.org/2001/XInclude"
py:strip="">	py:strip="">

<li class="widget-container boxed widget_text">	<li class="widget-container boxed widget_text">
<h4>Notes</h4>	<h4>Notes</h4>
<ul>	<ul>
<li>"Views" is the number of times a page was loaded in users' browsers.</li>	<li>"Views" is the number of times a page was loaded in users' browsers.</li>
<li>"Downloads" is the number of times a user has clicked to download either an original or cached resource for a particular dataset. Download information is only available from 2nd December 2012, 'No data' is shown for records before that date.</li>	<li>"Downloads" is the number of times a user has clicked to download either an original or cached resource for a particular dataset. Download information is only available from 2nd December 2012; 'No data' is shown for records before that date.</li>
<li>These usage statistics are confined to users with javascript enabled, which excludes web crawlers and API calls.</li>	<li>These usage statistics are confined to users with javascript enabled, which excludes web crawlers and API calls.</li>
<li>The results are not shown when the number of views/visits is tiny. Where these relate to site pages, results are available in full in the CSV download. Where these relate to users' web browser information, results are not disclosed, for privacy reasons.</li>	<li>The results are not shown when the number of views/visits is tiny. Where these relate to site pages, results are available in full in the CSV download. Where these relate to users' web browser information, results are not disclosed, for privacy reasons.</li>
</ul>	</ul>
</li>	</li>
</html>	</html>