Reduce transaction scope

rbhl/management/commands/generate_clinic_facts.py

@@ -2,17 +2,16 @@
 Generates the clinic activity facts
 """
 import datetime
-from collections import defaultdict
 from django.core.management.base import BaseCommand
 from django.db import transaction
 from opal.models import Episode
-from rbhl.models import ClinicLog, Diagnosis, Fact, Referral
+from rbhl.models import ClinicLog, Fact
 import logging
 logger = logging.getLogger('commands')
 
 
 class Command(BaseCommand):
-    def get_five_year_episodes(cls):
+    def get_five_year_episodes(self):
         today = datetime.date.today()
         if today.month > 9:
             five_year_range = (
@@ -32,8 +31,6 @@ def get_five_year_episodes(cls):
             cliniclog__clinic_date__lt=five_year_range[1]
         ).distinct().prefetch_related(
             "cliniclog_set",
-            "referral_set",
-            "diagnosis_set"
         )
 
     def handle(self, *args, **kwargs):
@@ -44,24 +41,27 @@ def handle(self, *args, **kwargs):
             raise
 
     @transaction.atomic
+    def save_facts(self, mean_clinic_patients_per_year, mean_known_diagnosis):
+        Fact.objects.filter(
+            label__in=[
+                Fact.MEAN_CLINIC_PATIENTS_PER_YEAR,
+                Fact.FIVE_YEAR_MEAN_KNOWN_DIAGNOSIS
+            ]
+        ).delete()
+        Fact.objects.create(
+            label=Fact.MEAN_CLINIC_PATIENTS_PER_YEAR,
+            value_int=mean_clinic_patients_per_year
+        )
+        Fact.objects.create(
+            label=Fact.FIVE_YEAR_MEAN_KNOWN_DIAGNOSIS,
+            value_int=mean_known_diagnosis
+        )
+
     def process(self, *args, **options):
         five_year_episodes = self.get_five_year_episodes()
         episode_count = five_year_episodes.count()
 
         referral_mean = round(episode_count/5)
-        Fact.objects.create(
-            label=Fact.MEAN_CLINIC_PATIENTS_PER_YEAR,
-            value_int=referral_mean
-        )
-        episode_ids = set(five_year_episodes.values_list('id', flat=True))
-        diagnosis = Diagnosis.objects.filter(
-            episode_id__in=episode_ids
-        ).exclude(
-            date=None
-        ).order_by("date")
-        episode_id_to_diagnosis = defaultdict(list)
-        for i in diagnosis:
-            episode_id_to_diagnosis[i.episode_id].append(i)
         diagnosed_count = 0
 
         # Diagnosis % is calclated by the
@@ -73,26 +73,4 @@ def process(self, *args, **options):
                 diagnosed_count += 1
 
         mean_diagnosis_percent = round(diagnosed_count/episode_count * 100)
-        Fact.objects.filter(label=Fact.FIVE_YEAR_MEAN_KNOWN_DIAGNOSIS).delete()
-        Fact.objects.create(
-            label=Fact.FIVE_YEAR_MEAN_KNOWN_DIAGNOSIS,
-            value_int=mean_diagnosis_percent
-        )
-
-        # Diagnosis time is calculated the time from our first diagnosis
-        # after the referral date
-        total_days = 0
-        count_with_diagnosis = 0
-        for episode in five_year_episodes:
-            referral = Referral.get_recent_referral_for_episode(episode)
-            if referral and referral.date_of_referral:
-                diagnoses = episode.diagnosis_set.all()
-                diagnoses = [i for i in diagnoses if i.date]
-                diagnosis_after_referral = [
-                    i for i in diagnoses if i.date > referral.date_of_referral
-                ]
-                if diagnosis_after_referral:
-                    diagnosis = diagnosis_after_referral[0]
-                    days_to_diagnosis = diagnosis.date - referral.date_of_referral
-                    total_days += days_to_diagnosis.days
-                    count_with_diagnosis += 1
+        self.save_facts(referral_mean, mean_diagnosis_percent)