From 54041078426a0c51ed7de6ee12fd2ac4e81ec7e8 Mon Sep 17 00:00:00 2001
From: Chris Hill-Scott <me@quis.cc>
Date: Tue, 6 Feb 2018 16:26:02 +0000
Subject: [PATCH] Split email list out again

The list of email domains is a different list from the list of all
government domains. And because the list of all government domains is
really long now, it could be unnecessarily slow to search through when
(a lot of the time) all we care about is whether the email address ends
with `.gov.uk`.
---
 app/email_domains.yml | 30 ++++++++++++++++++++++++++++++
 app/utils.py          | 12 ++++++++----
 2 files changed, 38 insertions(+), 4 deletions(-)
 create mode 100644 app/email_domains.yml

diff --git a/app/email_domains.yml b/app/email_domains.yml
new file mode 100644
index 000000000..bb90d459f
--- /dev/null
+++ b/app/email_domains.yml
@@ -0,0 +1,30 @@
+---
+- gov.uk
+- mod.uk
+- mil.uk
+- ddc-mod.org
+- slc.co.uk
+- gov.scot
+- parliament.uk
+- nhs.uk
+- nhs.net
+- police.uk
+- dclgdatamart.co.uk
+- ucds.email
+- naturalengland.org.uk
+- hmcts.net
+- scotent.co.uk
+- assembly.wales
+- cjsm.net
+- cqc.org.uk
+- bl.uk
+- stfc.ac.uk
+- wmfs.net
+- bbsrc.ac.uk
+- acas.org.uk
+- gov.wales
+- biglotteryfund.org.uk
+- marinemanagement.org.uk
+- britishmuseum.org
+- derrystrabane.com
+- highwaysengland.co.uk
diff --git a/app/utils.py b/app/utils.py
index 9874e1ec7..e0bf116c7 100644
--- a/app/utils.py
+++ b/app/utils.py
@@ -284,7 +284,7 @@ def get_help_argument():
 
 def is_gov_user(email_address):
     try:
-        GovernmentDomain(email_address)
+        GovernmentDomain(email_address, domain_list='email_domain_names')
         return True
     except NotGovernmentDomain:
         return False
@@ -447,17 +447,21 @@ class GovernmentDomain:
         domains = yaml.safe_load(domains)
         domain_names = sorted(domains.keys(), key=len)
 
-    def __init__(self, email_address_or_domain):
+    with open('{}/email_domains.yml'.format(_dir_path)) as email_domains:
+        email_domain_names = yaml.safe_load(email_domains)
+
+    def __init__(self, email_address_or_domain, domain_list='domain_names'):
 
         try:
             self._match = next(filter(
                 self.get_matching_function(email_address_or_domain),
-                self.domain_names,
+                getattr(self, domain_list),
             ))
         except StopIteration:
             raise NotGovernmentDomain()
 
-        self.owner, self.sector, self.agreement_signed = self._get_details_of_domain()
+        if domain_list == 'domain_names':
+            self.owner, self.sector, self.agreement_signed = self._get_details_of_domain()
 
     @staticmethod
     def get_matching_function(email_address_or_domain):