Make the guessing a bit more sophisticated

Things we’ve noticed from looking at real data that we could handle in a
smarter way:
- removing numbers (there might be a tom.smith2@dept.gov.uk if tom.smith
  is already taken)
- removing middle initials (again, these tend to be used for
  disambiguation and aren’t included when we ask people for their names)
- ignoring email addresses which only have someone’s initial, not their
  first name (because we can’t make a decent guess in this case)

tests/app/main/views/test_register.py

@@ -171,13 +171,23 @@ def test_register_with_existing_email_sends_emails(
 
 
 @pytest.mark.parametrize('email_address, expected_value', [
-    ("example123@example.com", ""),
     ("first.last@example.com", "First Last"),
     ("first.middle.last@example.com", "First Middle Last"),
+    ("first.m.last@example.com", "First Last"),
     ("first.last-last@example.com", "First Last-Last"),
     ("first.o'last@example.com", "First O’Last"),
     ("first.last+testing@example.com", "First Last"),
     ("first.last+testing+testing@example.com", "First Last"),
+    ("first.last6@example.com", "First Last"),
+    ("first.last.212@example.com", "First Last"),
+    ("first.2.last@example.com", "First Last"),
+    ("first.2b.last@example.com", "First Last"),
+    ("first.1.2.3.last@example.com", "First Last"),
+    ("first.last.1.2.3@example.com", "First Last"),
+    # Instances where we can’t make a good-enough guess:
+    ("example123@example.com", ""),
+    ("f.last@example.com", ""),
+    ("f.m.last@example.com", ""),
 ])
 def test_shows_registration_page_from_invite(
     client_request,