[423] Handle many more graphing edge-cases.
[ckanext-ga-report.git] / ckanext / ga_report / helpers.py
blob:a/ckanext/ga_report/helpers.py -> blob: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') == 'organization' ]
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, preview_image=None): def most_popular_datasets(publisher, count=20, preview_image=None):
   
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,
'preview_image': preview_image 'preview_image': preview_image
} }
   
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: if not p:
_log.warning("Could not find Package for {url}".format(url=entry.url)) _log.warning("Could not find Package for {url}".format(url=entry.url))
continue continue
   
  if not p.state == 'active':
  _log.warning("Package {0} is not active, it is {1}".format(p.name, p.state))
  continue
   
  if not p.private == False:
  _log.warning("Package {0} is private {1}".format(p.name, p.state))
  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)
   
def month_option_title(month_iso, months, day): def month_option_title(month_iso, months, day):
month_isos = [ iso_code for (iso_code,name) in months ] month_isos = [ iso_code for (iso_code,name) in months ]
try: try:
index = month_isos.index(month_iso) index = month_isos.index(month_iso)
except ValueError: except ValueError:
_log.error('Month "%s" not found in list of months.' % month_iso) _log.error('Month "%s" not found in list of months.' % month_iso)
return month_iso return month_iso
month_name = months[index][1] month_name = months[index][1]
if index==0: if index==0:
return month_name + (' (up to %s)'%day) return month_name + (' (up to %s)'%day)
return month_name return month_name