Merge pull request #1280 from Tpt/duckling

Duckling HTTP: uses Message time for parsing and allows to set timezone

rasa_nlu/extractors/duckling_http_extractor.py

@@ -5,6 +5,7 @@
 
 import logging
 import os
+import time
 
 import requests
 import simplejson
@@ -41,6 +42,10 @@ class DucklingHTTPExtractor(EntityExtractor):
 
         # locale - if not set, we will use the language of the model
         "locale": None,
+
+        # timezone like Europe/Berlin
+        # if not set the default timezone of Duckling is going to be used
+        "timezone": None
     }
 
     def __init__(self, component_config=None, language=None):
@@ -72,11 +77,19 @@ def _url(self):
 
         return self.component_config.get("url")
 
-    def _duckling_parse(self, text):
+    def _payload(self, text, reference_time):
+        return {
+            "text": text,
+            "locale": self._locale(),
+            "tz": self.component_config.get("timezone"),
+            "reftime": reference_time
+        }
+
+    def _duckling_parse(self, text, reference_time):
         """Sends the request to the duckling server and parses the result."""
 
         try:
-            payload = {"text": text, "locale": self._locale()}
+            payload = self._payload(text, reference_time)
             headers = {"Content-Type": "application/x-www-form-urlencoded; "
                                        "charset=UTF-8"}
             response = requests.post(self._url() + "/parse",
@@ -99,20 +112,33 @@ def _duckling_parse(self, text):
                          "Error: {}".format(e))
             return []
 
+    @staticmethod
+    def _reference_time_from_message(message):
+        if message.time is not None:
+            try:
+                return int(message.time)
+            except ValueError as e:
+                logging.warning("Could not parse timestamp {}. Instead "
+                                "current UTC time will be passed to "
+                                "duckling. Error: {}".format(message.time, e))
+        # fallbacks to current time
+        return int(time.time())
+
     def process(self, message, **kwargs):
         # type: (Message, **Any) -> None
 
         if self._url() is not None:
-            matches = self._duckling_parse(message.text)
+            reference_time = self._reference_time_from_message(message)
+            matches = self._duckling_parse(message.text, reference_time)
             dimensions = self.component_config["dimensions"]
             relevant_matches = filter_irrelevant_matches(matches, dimensions)
             extracted = convert_duckling_format_to_rasa(relevant_matches)
         else:
             extracted = []
-            logger.warn("Duckling HTTP component in pipeline, but no "
-                        "`url` configuration in the config "
-                        "file nor is `RASA_DUCKLING_HTTP_URL` "
-                        "set as an environment variable.")
+            logger.warning("Duckling HTTP component in pipeline, but no "
+                           "`url` configuration in the config "
+                           "file nor is `RASA_DUCKLING_HTTP_URL` "
+                           "set as an environment variable.")
 
         extracted = self.add_extractor_name(extracted)
         message.set("entities",