Error while trying to create my own task

[VishalPallagani]

Hi Ji,
Thanks a lot for the code and explanation to create our own task in ParlAI.
I am trying to follow suite and create my own task called as DialogPlan. The code files for the same are accessible at: https://drive.google.com/drive/folders/1aumHBWTyXVeTc9XGOqOJqLMz8GOFRwKS?usp=sharing

However, when trying to execute the task, I am getting an error saying the dialogplan agent isn't available.
I would be grateful if you can help me out with debugging this issue. Thanks in advance.

[GAONNR]

Hello, thank you for the issue.
Did you try the pip install . command after you implemented the task and the agent?
I think you should execute the command above to re-register the task and the agent to ParlAI for every edit in the code.

[VishalPallagani]

Yes, I did pip install . in the parlai folder. I have followed the same steps as mentioned by you in this echobot tutorial. Echobot worked perfectly fine but not sure why this error is occuring with this.

[VishalPallagani]

This is the exact error. I have added init.py files as well.

[GAONNR]

Maybe try to change the agent class's name in the form of (Folder name + Agent).
For your case, change the class name into class DialogplanAgent(Agent), not class Dialogplan(Agent).
I'm sorry I did not write the text tutorial in English.

[VishalPallagani]

Thank you, I will try that.
No, your tutorial was a lot helpful. Thank you for that.

[VishalPallagani]

Renaming it to DialogplanAgent(Agent) worked for me. Thank you very much.
However, I had one more question regarding how to do we exit from the conversation. For example, if I input "quit" or "bye" as a message, I want the execution to be stopped. Any help on how to do this would be appreciated.
Thanks in advance and also sorry for bothering you by reopening the issue

[GAONNR]

Hello, it's a pleasure that the renaming worked for the problem.
However, you have to add some scripts to exit the conversation.
I haven't made a script yet, but I think the file parlai/chat_service/tasks/overworld_demo/worlds.py will help you.
In line 57, there is a code for the message [DONE].
Adopting and editing this code seems to be able to solve the problem.
Thank you very much.
(Actually, I have been used Ctrl+C in the terminal to end the conversation.)

[VishalPallagani]

Yes, I have been using Ctrl + C but right now, I am working on making a chatbot using ParlAI which performs multi-turn conversation. The chatbot I am developing is for querying contacts in a phone directory. Firstly, I am unable to achieve multi-turn dialogue conversation with ParlAI, as with each input of text, the program is being executed from the start. However, my aim is to design conversations that are dependent on the history of the chat as well, and then when the user enters exit, stop the execution.

This is the basic code I have currently written for the Agent:

class DialogplanAgent(Agent):
    def __init__(self, opt):
        super().__init__(opt)
        self.id = self.__class__.__name__
        self.heard = None
        self.flag = True
        self.GREETING_INPUTS = ("hello", "hi", "greetings", "sup", "what's up","hey",)
        self.GREETING_RESPONSES = ["hi", "hey", "*nods*", "hi there", "hello", "I am glad! You are talking to me"]
        self.EXACTMATCH_INPUTS = ("call")
        self.EXACTMATCH_RESPONSES = ["exists in the directory with number", "is present in the directory with number"]
        self.df = pd.read_csv('/Users/vishal/ParlAI/parlai_internal/Directory.csv', header=None)

    def greeting(self, sentence):
        for word in sentence.split():
            if word.lower() in self.GREETING_INPUTS:
                return random.choice(self.GREETING_RESPONSES)

    def exactMatch(self, sentence):
        temp = sentence.split()
        name = temp[1]
        for word in sentence.split():
            if word.lower() in self.EXACTMATCH_INPUTS:
                return random.choice(self.EXACTMATCH_RESPONSES), name

    
    def feedback(self):
        answer = input('🧍 Enter 1 if you are satisfied with the output: ')
        return answer

    def observe(self, observation):
        flag = self.flag
        while(flag == True):
            self.heard = observation['text']
            user_response = self.heard.lower()
            if user_response != 'bye':
                if user_response == 'thanks' or user_response == 'thank you':
                    flag = False
                    print("🤖 You are welcome..")
                else:
                    if self.greeting(user_response) != None:
                        print("🤖 "+self.greeting(user_response))
                    if self.exactMatch(user_response) != None:
                        response, target_name = self.exactMatch(user_response)
                        for i in range(len(self.df)):
                            if self.df.iloc[i,0] == target_name:
                                print(f"🤖 {target_name} {response} {self.df.iloc[i,1]}")
                                feedback_exact = self.feedback() 
                                if feedback_exact == '1':
                                    print("🤖 Goal Achieved.")
                                    flag = False
                                else:
                                    print(f"🤖 Giving close match. Are you looking for {self.df.iloc[i+1,0]}")
                                    feedback_close = self.feedback()
                                    if feedback_close == '1':
                                        print(f"🤖 Number of {self.df.iloc[i+1,0]} is {self.df.iloc[i+1,1]}")
                                        print("🤖 Goal Achieved.")
                                        flag = False
                                    else:
                                        print("🤖 Desired entry doesn't exist in the directory. Please add it to directory")
                                        name_to_add = input("🤖 Enter the name of the person: ")
                                        num_to_add = int(input(f"🤖 Enter the number of {name_to_add} "))
                                        self.df = self.df.append(pd.DataFrame([[name_to_add, num_to_add]]), ignore_index=True)
                                        print("🤖 Entry added.")
                                        print("🤖 Goal Achieved.")
                                        flag = False
            else:
                flag = False
                print("🤖 Bye! take care..")

    def act(self):
        return {
            'id': self.id
        }

Please let me know if you can suggest me something in this code that will help me with what I am actually trying to implement. Also, thank you for giving me a heads up, I will surely check the worlds.py and see what I can do. Thank you very much.