--- a/ckanext/ga_report/controller.py +++ b/ckanext/ga_report/controller.py @@ -1,6 +1,7 @@ import re import csv import sys +import json import logging import operator import collections @@ -21,6 +22,10 @@ d = strptime(strdate, '%Y-%m') return '%s %s' % (calendar.month_name[d.tm_mon], d.tm_year) +def _get_unix_epoch(strdate): + from time import strptime,mktime + d = strptime(strdate, '%Y-%m') + return int(mktime(d)) def _month_details(cls, stat_key=None): ''' @@ -107,11 +112,26 @@ return key, val + # Query historic values for sparkline rendering + sparkline_query = model.Session.query(GA_Stat)\ + .filter(GA_Stat.stat_name=='Totals')\ + .order_by(GA_Stat.period_name) + sparkline_data = {} + for x in sparkline_query: + sparkline_data[x.key] = sparkline_data.get(x.key,[]) + key, val = clean_key(x.key,float(x.value)) + tooltip = '%s: %s' % (_get_month_name(x.period_name), val) + sparkline_data[x.key].append( (tooltip,x.value) ) + # Trim the latest month, as it looks like a huge dropoff + for key in sparkline_data: + sparkline_data[key] = sparkline_data[key][:-1] + c.global_totals = [] if c.month: for e in entries: key, val = clean_key(e.key, e.value) - c.global_totals.append((key, val)) + sparkline = sparkline_data[e.key] + c.global_totals.append((key, val, sparkline)) else: d = collections.defaultdict(list) for e in entries: @@ -121,10 +141,18 @@ v = sum(v) else: v = float(sum(v))/float(len(v)) + sparkline = sparkline_data[k] key, val = clean_key(k,v) - c.global_totals.append((key, val)) - c.global_totals = sorted(c.global_totals, key=operator.itemgetter(0)) + c.global_totals.append((key, val, sparkline)) + # Sort the global totals into a more pleasant order + def sort_func(x): + key = x[0] + total_order = ['Total page views','Total visits','Pages per visit'] + if key in total_order: + return total_order.index(key) + return 999 + c.global_totals = sorted(c.global_totals, key=sort_func) keys = { 'Browser versions': 'browser_versions', @@ -161,7 +189,29 @@ for k, v in keys.iteritems(): q = model.Session.query(GA_Stat).\ - filter(GA_Stat.stat_name==k) + filter(GA_Stat.stat_name==k).\ + order_by(GA_Stat.period_name) + # Run the query on all months to gather graph data + series = {} + x_axis = set() + for stat in q: + x_val = _get_unix_epoch(stat.period_name) + series[ stat.key ] = series.get(stat.key,{}) + series[ stat.key ][x_val] = float(stat.value) + x_axis.add(x_val) + # Common x-axis for all series. Exclude this month (incomplete data) + x_axis = sorted(list(x_axis))[:-1] + # Buffer a rickshaw dataset from the series + def create_graph(series_name, series_data): + return { + 'name':series_name, + 'data':[ {'x':x,'y':series_data.get(x,0)} for x in x_axis ] + } + rickshaw = [ create_graph(name,data) for name, data in series.items() ] + rickshaw = sorted(rickshaw,key=lambda x:x['data'][-1]['y']) + setattr(c, v+'_graph', json.dumps(rickshaw)) + + # Buffer the tabular data if c.month: entries = [] q = q.filter(GA_Stat.period_name==c.month).\ @@ -178,7 +228,7 @@ # 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']) + total = sum([x for n,x,graph 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 ]) @@ -284,8 +334,9 @@ filter(GA_Stat.key==package.name) if month != 'All': # Fetch everything unless the month is specific dls = dls.filter(GA_Stat.period_name==month) - - downloads = sum(int(d.value) for d in dls.all()) + downloads = 0 + for x in dls: + downloads += int(x.value) else: downloads = 'No data' top_packages.append((package, entry.pageviews, entry.visits, downloads)) @@ -335,6 +386,26 @@ c.publisher_page_views = entry.pageviews if entry else 0 c.top_packages = self._get_packages(c.publisher, 20) + + # Graph query + top_package_names = [ x[0].name for x in c.top_packages ] + graph_query = model.Session.query(GA_Url,model.Package)\ + .filter(model.Package.name==GA_Url.package_id)\ + .filter(GA_Url.url.like('/dataset/%'))\ + .filter(GA_Url.package_id.in_(top_package_names)) + graph_data = {} + for entry,package in graph_query: + if not package: continue + if entry.period_name=='All': continue + graph_data[package.id] = graph_data.get(package.id,{ + 'name':package.title, + 'data':[] + }) + graph_data[package.id]['data'].append({ + 'x':_get_unix_epoch(entry.period_name), + 'y':int(entry.pageviews), + }) + c.graph_data = json.dumps(graph_data.values()) return render('ga_report/publisher/read.html')