RasaHQ/rasa rasa chatbot duckling implementation issue

Rasa Open Source
1

I’m currently working on rasa chatbot , I can install duckling but can not implement it in for time and date entity so pls help me if any one had already use it. thank you

2

Hi @IKhushi - I’m afraid I’ll need a bit more information to help you. What have you tried so far? What are the steps you are following, and what is the error you encounter?

3

I’ve added time entity in my data and train it. after that when I test it with message of time in my API it return empty entity. ########## nlu.md

intent:block_unblock

########### config.py pipeline: supervised_embeddings

-name: ner_duckling_http

when I’ll test the api it will give warning :

rasa.nlu.extractors.crf_entity_extractor - extracted in crf_entity extarctor

4

the problem is with your config.py - since you provide the value supervised_embeddings, this specifies a whole pipeline. Try removing that so it’s just:

pipeline:
- name: ner_duckling_http
  - url: http://localhost:8000
  - dimensions: ["time", "day"]
  - locale: en_US
  - timezone: Europe/Berlin

and then add back the other pieces you need,

- name: "CountVectorsFeaturizer"
- name: "EmbeddingIntentClassifier"
- name: "ner_duckling_http"
...
5

I’ve tried

pipeline:

  • name: ner_duckling_http

in config file but gives error :

The config file ‘config.yml’ is missing mandatory parameters: ‘pipeline’. Add missing parameters to config file and try again.

6

after modification in config file

{ “intent”: { “name”: “block_unblock”, “confidence”: 0.9578648806 }, “entities”: [ { “start”: 0, “end”: 5, “value”: “block”, “entity”: “blockType”, “confidence”: 0.9478110737, “extractor”: “CRFEntityExtractor” }, { “start”: 6, “end”: 14, “value”: “tommorow”, “entity”: “targetTime”, “confidence”: 0.8920832596, “extractor”: “CRFEntityExtractor” }, { “start”: 6, “end”: 14, “text”: “tommorow”, “value”: “2019-08-13T00:00:00.000-07:00”, “confidence”: 1, “additional_info”: { “values”: [ { “value”: “2019-08-13T00:00:00.000-07:00”, “grain”: “day”, “type”: “value” } ], “value”: “2019-08-13T00:00:00.000-07:00”, “grain”: “day”, “type”: “value” }, “entity”: “time”, “extractor”: “DucklingHTTPExtractor” } ], “intent_ranking”: [ { “name”: “block_unblock”, “confidence”: 0.9578648806 } }

but duckling returns entity name as time instead of targetTime of that value from trained data.So, now what should I do to get entity value from duckling?

7

duckling is a static tool, so it doesn’t look at your training data at all. You can map the time to targetTime using an entity synonym. You'll have to addner_synonyms` to your pipeline as well

1 Like
8

Alan, thank you so much for your support and guidance.