Using Rasa 2.4.0.
I have a Rasa server running locally and I am trying to train a model ONLY using this api call: Rasa Open Source Documentation
For some reason I cannot train the model to incorporate the DucklingHTTPExtractor. Everything else works but the entity extraction. I tried to take out the quotation marks, but nothing worked. Additionally, when I train an exact model through rasa train everything works fine.
Here is my API call command:
curl -X PUT http://localhost:5005/model -H ‘Content-type: application/json’ --data-binary "@newModel.txt"
Here is my training data:
{
"config":"",
"nlu":"",
"stories":"",
"domain": "version: \"2.0\"
intents:
- greet
- goodbye
- bot_challenge
- callback
entities:
- \"time\"
- \"phone-number\"
slots:
phone_no:
type: text
influence_conversation: false
time_of_callback:
type: text
influence_conversation: false
responses:
utter_greet:
- text: \"Hello, this is the callback bot service, powered by Rasa. How could I help?\"
utter_goodbye:
- text: \"Goodbye\"
utter_iamabot:
- text: \"I am a bot, powered by Rasa. How can I help?\"
utter_ask_phone_no:
- text: \"What is your phone number?\"
utter_ask_time_of_callback:
- text: \"What date and time would you like to be called at?\"
utter_callback:
- text: \"Alright, lets get you booked in straight away. I would require a few details from you please, in order to schedule the call.\"
actions:
- utter_goodbye
- utter_greet
- utter_iamabot
forms:
callback_scheduling_form:
phone_no:
- type: from_entity
entity: \"phone-number\"
time_of_callback:
- type: from_entity
entity: \"time\"
session_config:
session_expiration_time: 60 # value in minutes
carry_over_slots_to_new_session: true
pipeline:
- name: WhitespaceTokenizer
- name: RegexFeaturizer
- name: LexicalSyntacticFeaturizer
- name: CountVectorsFeaturizer
- name: CountVectorsFeaturizer
analyzer: char_wb
min_ngram: 1
max_ngram: 4
- name: DIETClassifier
epochs: 100
constrain_similarities: true
- name: EntitySynonymMapper
- name: ResponseSelector
epochs: 100
constrain_similarities: true
- name: FallbackClassifier
threshold: 0.3
ambiguity_threshold: 0.1
- name: \"DucklingHTTPExtractor\"
url: \"http://localhost:8000\"
dimensions: [\"time\", \"phone-number\"]
policies:
- name: MemoizationPolicy
- name: TEDPolicy
max_history: 5
epochs: 10
nlu:
- intent: greet
examples: |
- hi
- hello
- how are you
- good morning
- good evening
- hey
- mornin
- hey there
- aloha
- hey dude
- good afternoon
- intent: goodbye
examples: |
- bye
- goodbye
- ciao
- cu
- cya
- good bye
- cee you later
- see ya
- good night
- intent: bot_challenge
examples: |
- are you a bot?
- are you a human?
- am I talking to a bot?
- am I talking to a human?
- I want pizza
- what time is it?
- what is the weather
- tell me the weather
- never mind
- what is 2 + 2?
- intent: callback
examples: |
- I would like to arrange a callback at 3
- I would like to arrange a callback
- call me back tomorrow at 1 pm
- arrange call Monday the 13th at 13:20
- I want to arrange a callback
- rearrange call for tomorrow at 10am
- Can you arrange a callback?
- I want a call at 3.00
- I want a call on Monday 13th
- I request to arrange a call on Friday, the 13th of July at 10:45am
- Can I get a callback?
- Schedule callback
- Can you schedule a callback
- call me back
- Call me back at 3 please
- 07288838922
- 07792992452
- My number is 07792452923
rules:
- rule: activate callback scheduling form
steps:
- intent: callback
- action: callback_scheduling_form
- active_loop: callback_scheduling_form
- rule: submit form callback
condition:
- active_loop: callback_scheduling_form
steps:
- action: callback_scheduling_form
- active_loop: null
- action: utter_goodbye
stories:
- story: say hi
steps:
- intent: greet
- action: utter_greet
- story: challenge the bot
steps:
- intent: bot_challenge
- action: utter_iamabot
- story: say goodbye
steps:
- intent: goodbye
- action: utter_goodbye"
}
I can communicate with the duckling server running on port 8000. So I am not quite sure what my issue is.
(This api call has an unusual format due to a bug in this version, so that shouldn’t be the issue)