From 973466bb1370fc20f54b6acb8209c46b2994b8a3 Mon Sep 17 00:00:00 2001
From: Al <albarrentine@gmail.com>
Date: Fri, 9 Dec 2016 19:48:59 -0500
Subject: [PATCH] [states] adding multiple state abbreviations for states that
 can have periods in the naem like D.C., D.F. in Mexico and Brasil, etc.

---
 resources/states/au.yaml                      | 28 +++++++++++++++++++
 resources/states/br.yaml                      |  4 +++
 resources/states/mx.yaml                      |  4 +++
 resources/states/us.yaml                      |  4 +++
 scripts/geodata/states/state_abbreviations.py | 25 ++++++++++-------
 5 files changed, 55 insertions(+), 10 deletions(-)

diff --git a/resources/states/au.yaml b/resources/states/au.yaml
index 68b02f0a..c8019f13 100644
--- a/resources/states/au.yaml
+++ b/resources/states/au.yaml
@@ -1,20 +1,48 @@
+"A.A.T":
+    en: Australian Antarctic Territory
+"A.A.T.":
+    en: Australian Antarctic Territory
 AAT:
     en: Australian Antarctic Territory
+"A.C.T":
+    en: Australian Capital Territory
+"A.C.T.":
+    en: Australian Capital Territory
 ACT:
     en: Australian Capital Territory
+"J.B.T.":
+    en: Jervis Bay Territory
+"J.B.T":
+    en: Jervis Bay Territory
 JBT:
     en: Jervis Bay Territory
+"N.S.W":
+    en: New South Wales
+"N.S.W.":
+    en: New South Wales
 NSW:
     en: New South Wales
+"N.T":
+    en: Northern Territory
+"N.T.":
+    en: Northern Territory
 NT:
     en: Northern Territory
 QLD:
     en: Queensland
+"S.A":
+    en: South Australia
+"S.A.":
+    en: South Australia
 SA:
     en: South Australia
 TAS: 
     en: Tasmania
 VIC:
     en: Victoria
+"W.A":
+    en: Western Australia
+"W.A.":
+    en: Western Australia
 WA:
     en: Western Australia
\ No newline at end of file
diff --git a/resources/states/br.yaml b/resources/states/br.yaml
index 752c875f..f64c0445 100644
--- a/resources/states/br.yaml
+++ b/resources/states/br.yaml
@@ -10,6 +10,10 @@ BA:
     pt: Bahia
 CE:
     pt: Ceará
+"D.F":
+    pt: Distrito Federal
+"D.F.":
+    pt: Distrito Federal
 DF:
     pt: Distrito Federal
 ES:
diff --git a/resources/states/mx.yaml b/resources/states/mx.yaml
index 7412db6c..dfd69811 100644
--- a/resources/states/mx.yaml
+++ b/resources/states/mx.yaml
@@ -14,6 +14,10 @@ CA:
     es: Coahuila
 CL:
     es: Colima
+"D.F.":
+    es: Distrito Federal
+"D.F":
+    es: Distrito Federal
 DF:
     es: Distrito Federal
 DU:
diff --git a/resources/states/us.yaml b/resources/states/us.yaml
index 5e3862e0..7e971448 100644
--- a/resources/states/us.yaml
+++ b/resources/states/us.yaml
@@ -12,6 +12,10 @@ CO:
     en: Colorado
 CT:
     en: Connecticut
+"D.C":
+    en: District of Columbia
+"D.C.":
+    en: District of Columbia
 DC:
     en: District of Columbia
 DE:
diff --git a/scripts/geodata/states/state_abbreviations.py b/scripts/geodata/states/state_abbreviations.py
index 07d980be..e3ad15d1 100644
--- a/scripts/geodata/states/state_abbreviations.py
+++ b/scripts/geodata/states/state_abbreviations.py
@@ -1,4 +1,5 @@
 import os
+import random
 import six
 import sys
 import yaml
@@ -24,30 +25,34 @@ class StateAbbreviations(object):
             country = filename.split('.yaml')[0]
             country_config = yaml.load(open(os.path.join(base_dir, filename)))
 
-            country_abbreviations = defaultdict(dict)
-            country_full_names = defaultdict(dict)
+            country_abbreviations = defaultdict(list)
+            country_full_names = defaultdict(list)
 
             for abbreviation, vals in six.iteritems(country_config):
                 for language, full_name in six.iteritems(vals):
                     full_name = safe_decode(full_name)
                     abbreviation = safe_decode(abbreviation)
-                    country_abbreviations[full_name.lower()][language] = abbreviation
-                    country_full_names[abbreviation.lower()][language] = full_name
+                    country_abbreviations[(full_name.lower(), language)].append(abbreviation)
+                    country_full_names[(abbreviation.lower(), language)].append(full_name)
 
             self.abbreviations[country] = dict(country_abbreviations)
             self.full_names[country] = dict(country_full_names)
 
     def get_abbreviation(self, country, language, state, default=None):
-        value = nested_get(self.abbreviations, (country.lower(), state.lower(), language.lower()))
-        if value is DoesNotExist:
+        values = nested_get(self.abbreviations, (country.lower(), (state.lower(), language.lower())))
+        if values is DoesNotExist:
             return default
-        return value
+        if len(values) == 1:
+            return values[0]
+        return random.choice(values)
 
     def get_full_name(self, country, language, state, default=None):
-        value = nested_get(self.full_names, (country.lower(), state.lower(), language.lower()))
-        if value is DoesNotExist:
+        values = nested_get(self.full_names, (country.lower(), (state.lower(), language.lower())))
+        if values is DoesNotExist:
             return default
-        return value
+        if len(values) == 1:
+            return values[0]
+        return random.choice(values)
 
 
 state_abbreviations = StateAbbreviations()