How can I deploy my chatbot?

Hello all,

I am using java for the NLP part of my chatbot now I need to combine them. Actually, I don’t have experience with this so I don’t know how can I do that. I want to listen to different ports for use this chatbot’s two different parts. (If there is no else better way)

I need your help, how can I combine rasa and java? I need inputs for NLP Part(JAVA). I am open to new ideas. I did not use AWS, AZURE, etc. before so it was hard on my first try.

Sorry if the post is in the wrong place.

Thanks all.

Welcome IIker,

Are you saying that you don’t want to use the Rasa NLU? Or you want to call Rasa from Java?

Greg

Hey,

I can assist you, please reach me on Skype - cis.seth

Regards! Seth

Hey Stephens,

I am sorry for my late response, I want to explain more. I am developing a chatbot for a university and I have a dataset about the university’s rules and so on. I need to use Java for the NLP part of my project as I said before. In the end, I want to take input from the chatbot (I use Telegram now.) to Java as JSON format and then I will find the correct response in my dataset and I need to show the response on Telegram so I don’t have any idea how can I do that.

I searched some papers, in these papers people have used AWS, etc but I am not sure this way is the correct way or not.

I want to use Rasa with my ML system.

Thanks for your answer.

Sounds like you are going to develop your own Java solution and I see no role for Rasa in your project.

I have done a similar project converting a document with rules into a basic FAQ bot with Rasa and this worked well but still involved creating much of the training data. In my project, I wanted the user to be able to ask questions about the rules and provide a response along with a link to the rule document.

In this case, imported the document into an open source wiki tool and put up the Chatroom react widget we reference as part of the wiki site. I created training data for the Rasa NLU and used the Response Selector.

Other than some manipulation of the source document for the wiki and initial conversion of the doc to jumpstart the NLU FAQ training data, there was no code involved in my project.