Build the index for the geo dataset to speed up the translations lookup

1 files changed, 41 insertions, 17 deletions

diff --git a/gui/scripts/extract-geo-data.py b/gui/scripts/extract-geo-data.py
index 273392e3ad..fa1a094fe6 100644
--- a/gui/scripts/extract-geo-data.py
+++ b/gui/scripts/extract-geo-data.py
@@ -431,11 +431,8 @@ class PlaceTranslator(object):
 
   def __init__(self):
     super(PlaceTranslator, self).__init__()
-    shape_path = get_shape_path("ne_10m_populated_places")
-    self.source = fiona.open(shape_path, "r")
 
-  def __del__(self):
-    self.source.close()
+    self.dataset = self.__build_index()
 
   def translate(self, locale, english_city_name):
     """
@@ -450,25 +447,52 @@ class PlaceTranslator(object):
     preferred_locales = (get_locale_language(locale), convert_locale_ident(locale))
     match_prop_keys = list("name_" + x for x in preferred_locales)
 
-    for feat in self.source:
-      props = lower_dict_keys(feat["properties"])
+    props = self.dataset.get(english_city_name)
 
-      # namepar works for "Wien"
-      # use nameascii to match "Sao Paolo"
-      if props.get("name") == english_city_name or \
-         props.get("namepar") == english_city_name or \
-         props.get("nameascii") == english_city_name:
-        for key in match_prop_keys:
-          value = props.get(key)
+    if props is not None:
+      for key in match_prop_keys:
+        value = props.get(key)
 
-          if value is not None:
-            return value
+        if value is not None:
+          return value
 
-        print c.orange(u"Missing translation for {} ({}). Probe keys: {}".format(
-          english_city_name, locale, match_prop_keys).encode('utf-8'))
+      print c.orange(u"Missing translation for {} ({}). Probe keys: {}".format(
+        english_city_name, locale, match_prop_keys).encode('utf-8'))
 
     return None
 
+  def __build_index(self):
+    """
+    Private helper to build the index for the geo dataset, that can be used to speed up the
+    translations lookup.
+    """
+    shape_path = get_shape_path("ne_10m_populated_places")
+    dataset = dict()
+
+    # build a hash map of the entire datasource in memory
+    with fiona.open(shape_path, "r") as source:
+      for feat in source:
+        props = lower_dict_keys(feat["properties"])
+
+        name = props.get("name")
+
+        # namepar works for "Wien"
+        namepar = props.get("namepar")
+
+        # use nameascii to match "Sao Paolo"
+        nameascii = props.get("nameascii")
+
+        if name is not None:
+          dataset[name] = props
+
+        if namepar is not None:
+          dataset[namepar] = props
+
+        if nameascii is not None:
+          dataset[nameascii] = props
+
+    return dataset
+
 
 def get_shape_path(dataset_name):
   return path.join(SCRIPT_DIR, dataset_name, dataset_name + ".shp")