Compute hits per day from site data.

Speeds up parsing a bit, and makes future extensions easier.

apachedex/__init__.py

@@ -392,6 +392,19 @@ class ERP5SiteStats(GenericSiteStats):
     else:
       self.no_module[date].accumulate(match)
 
+  def getApdexData(self):
+    asDict = lambda apdex_data: dict((x, (y, z)) for (x, y, z) in apdex_data)
+    date_dict = asDict(super(ERP5SiteStats, self).getApdexData())
+    for date, (apdex, hit) in asDict(getDataPoints(self.no_module)
+        ).iteritems():
+      super_apdex, super_hit = date_dict.get(date, (0, 100))
+      new_hit = hit + super_hit
+      if super_hit:
+        new_apdex = super_apdex * new_hit / super_hit + apdex * new_hit / hit
+      date_dict[date] = (new_apdex, new_hit)
+    return sorted(((date, apdex, hit)
+      for date, (apdex, hit) in date_dict.iteritems()), key=ITEMGETTER0)
+
   def asHTML(self, date_format, placeholder_delta, graph_period,
       stat_filter=lambda x: x):
     result = []
@@ -695,7 +708,6 @@ def main():
   error_detail = args.error_detail
   file_count = len(infile_list)
   per_site = {}
-  hit_per_day = defaultdict(int)
   skip_user_agent = list(itertools.chain(*args.skip_user_agent))
   malformed_lines = 0
   skipped_lines = 0
@@ -745,7 +757,6 @@ def main():
         skipped_lines += 1
         continue
       date = asDate(match.group('timestamp'))
-      hit_per_day[decimator(date)] += 1
       try:
         site_data = per_site[site]
       except KeyError:
@@ -788,6 +799,10 @@ def main():
         caption, value))
     out.write('</table><h2>Hits per period</h2><table class="stats">'
       '<tr><th>date</th><th>hits</th></tr>')
+    hit_per_day = defaultdict(int)
+    for site_data in per_site.itervalues():
+      for date, _, hit in site_data.getApdexData():
+        hit_per_day[decimator(date)] += hit
     for date, hit in sorted(hit_per_day.iteritems(), key=ITEMGETTER0):
       out.write('<tr><td>%s</td><td>%s</td></tr>' % (date, hit))
     out.write('</table>')