Conflicts in execution of form

Rasa Open Source
1

Hi!
Let me describe the case

case_code : bug ID
emp_id : Id of the employee

Now i have two functions to do -
1.) translate the bug id
for this i will require only case_code
2.) assign the case_code to another emp_id.
for this i requuire both case_code and emp_id

i am giving u the snippets of code\

Stories.md-

interactive_story_1

  • translate{“case_code”: “p191107-07787”}
    • slot{“case_code”: “p191107-07787”}
    • action_translate_issue
    • slot{“case_code”: null}

story_1

  • ask_id{“emp_id”: “k.saurabh”}

    • action_id

  • ask_assign

    • bot_form
    • form{“name”:“bot_form”}
    • slot{“case_code”:“P12331-98450”}
    • slot{“emp_id”:“pratush.kr”}
    • form{“name”:null}

  • thankyou

    • utter_happy

    and more examples …

#nlu.md:

intent:plm_case_code

  • [P12345-09876] (case_code)
  • [P23409-76462] (case_code)
  • case code is [P12345-09876] (case_code)
  • case code [P23409-76462] (case_code)
  • casecode [P65108-76109] (case_code)
  • plm id is [P08186-60721] (case_code)
  • plm case code [P69852-73145] (case_code)
  • defect id is [P32198-65931] (case_code)
  • defect id - [P32198-47963] (case_code)

intent:translate

  • translate issues id [P190100-12345] (case_code)
  • translate issue [P190100-12345] (case_code)
  • translate the issue in english on [P190100-12345] (case_code)
  • traslate [P190100-12345] (case_code) issue
  • translate issue [P190100-12345] (case_code) to english
  • translate [P190100-12345] (case_code) issue to english
  • Please translate issue [P190100-12345] (case_code)

intent: ask_id

  • [vi.kumar] (emp_id) … a few more examples

actions.py

class BotForm(FormAction):

def name(self):
    return "bot_form"
@staticmethod
def required_slots(tracker):
    return["case_code", "id"]

def slot_mappings(self) -> Dict[Text, Union[Dict, List[Dict]]]:
    """A dictionary to map required slots to
        - an extracted entity
        - intent: value pairs
        - a whole message
        or a list of them, where a first match will be picked"""

    return {
        "case_code": [
            self.from_entity(entity = "case_code", intent = "plm_case_code"),
            self.from_text(intent = None)
        ],
        "id": [
            self.from_entity(entity = "id" , intent = "ask_knox_id_part_group"),
            self.from_text(intent = None)
        ]
    }

def validate_case_code(
    self,
    value: Text,
    dispatcher: CollectingDispatcher,
    tracker: Tracker,
    domain: Dict[Text, Any],
) -> Dict[Text, Any]:
    print("validate_case_code")
    # code for validation

def validate_id(
    self,
    value: Text,
    dispatcher: CollectingDispatcher,
    tracker: Tracker,
    domain: Dict[Text, Any],
) -> Dict[Text, Any]:
    try:
        print("validate_id")
        # code for validation

def request_next_slot(
    self,
    dispatcher: "CollectingDispatcher",
    tracker: "Tracker",
    domain: Dict[Text, Any],
) -> Dict[Text, Any]:
    intent = tracker.latest_message.get("intent", {}).get("name")
    print("intent name = ", intent)
    for slot in self.required_slots(tracker):
        if self._should_request_slot(tracker, slot):
            ## Condition of validated slot that triggers deactivation
            if intent == "stop":
                print("form will be stopped")
                dispatcher.utter_message("Form will be stopped")
                return self.deactivate()
            print("Request next slot '{}'".format(slot))
            dispatcher.utter_template("utter_ask_{}".format(slot),
                                    tracker,
                                    silent_fail=False,
                                    **tracker.slots)
            return [SlotSet(REQUESTED_SLOT, slot)]
    print("No slots left to request")
    return None

def submit(
    self,
    dispatcher: CollectingDispatcher,
    tracker: Tracker,
    domain: Dict[Text, Any],
) -> List[Dict]:
    # code for execution

My doubt is -
once intent “assign” is predicted -> bot_form is executed ->
utter_ask_case_code runs -> i provide the case_code. ->
intent “plm_case_code” is predicted when i provide case_code , it enters into the translate intent and execute action_translate. How to solve this so that i can execute my form successfully without a break in the flow of story.

2

How does your training data for the intent ask_assign look like? Also does it make sense to extend the slot mapping of the form to include more intents?

3

Hi, ask_assign will be a payload .

i will give a button “Assign” for this button , payload intent = ask_Assign.

once ask_assign is picked , i have written stories for the activation of form.

4

One more thing there are two slots which are marked required {case_code , emp_id}.

as i have described , there is a intent “plm_case_code” for picking case_code but once plm_case_code is predicted, instead of asking emp_id (going with the flow of form) , action_translate_issue is predicted .

i want to avoid this . kindly guide me through this

5

Could you please upload the complete nlu.md and stories.md files? It is hard to see what is going wrong without those files. Thanks!