Load gazetteer lazily in NgramFactory

snips_nlu/slot_filler/feature_factory.py

@@ -213,7 +213,7 @@ def __init__(self, factory_config, **shared):
         self.use_stemming = self.args["use_stemming"]
         self.common_words_gazetteer_name = self.args[
             "common_words_gazetteer_name"]
-        self.gazetteer = None
+        self._gazetteer = None
         self._language = None
         self.language = self.args.get("language_code")
 
@@ -226,9 +226,16 @@ def language(self, value):
         if value is not None:
             self._language = value
             self.args["language_code"] = self.language
-            if self.common_words_gazetteer_name is not None:
-                self.gazetteer = get_gazetteer(
-                    self.resources, self.common_words_gazetteer_name)
+
+    @property
+    def gazetteer(self):
+        # Load the gazetteer lazily
+        if self.common_words_gazetteer_name is None:
+            return None
+        if self._gazetteer is None:
+            self._gazetteer = get_gazetteer(
+                self.resources, self.common_words_gazetteer_name)
+        return self._gazetteer
 
     @property
     def feature_name(self):

snips_nlu/tests/test_crf_slot_filler.py

@@ -235,6 +235,44 @@ def test_should_not_log_weights_when_not_fitted(self):
         with self.assertRaises(NotTrained):
             slot_filler.log_weights()
 
+    def test_refit(self):
+        # Given
+        dataset_stream = io.StringIO("""
+---
+type: intent
+name: my_intent
+utterances:
+- this is [entity1](my first entity)""")
+        dataset = Dataset.from_yaml_files("en", [dataset_stream]).json
+
+        updated_dataset_stream = io.StringIO("""
+---
+type: intent
+name: my_intent
+utterances:
+- this is [entity1](my first entity)
+- this is [entity1](my first entity) again""")
+        updated_dataset = Dataset.from_yaml_files(
+            "en", [updated_dataset_stream]).json
+
+        config = CRFSlotFillerConfig(feature_factory_configs=[
+            {
+                "args": {
+                    "common_words_gazetteer_name": "top_10000_words_stemmed",
+                    "use_stemming": True,
+                    "n": 1
+                },
+                "factory_name": "ngram",
+                "offsets": [-2, -1, 0, 1, 2]
+            },
+        ])
+
+        # When
+        slot_filler = CRFSlotFiller(config).fit(dataset, "my_intent")
+
+        # Then
+        slot_filler.fit(updated_dataset, "my_intent")
+
     def test_should_fit_with_naughty_strings_no_tags(self):
         # Given
         naughty_strings_path = TEST_PATH / "resources" / "naughty_strings.txt"
@@ -755,7 +793,7 @@ def test_should_fit_and_parse_empty_intent(self):
         slot_filler.fit(dataset, "dummy_intent")
         slot_filler.get_slots("ya")
 
-    def test___ensure_safe(self):
+    def test_ensure_safe(self):
         unsafe_examples = [
             ([[]], [[]]),
             ([[], []], [[], []]),