chatbot based on rasa framework as an assessment NLP task for Master-linux company jan-2022
- Note:- now the chatbot have an issue just because the endpoint that was provided by master-linux firm is not valid anymore, please replace the existing endpoint with a working one, and the model will work properly
The goal is to create Rasa Chatbot that can provide answers for users questions about the Population and the capitel city of collecton of countries, where the chatbot use the restful API to get that information from third-party website
to clone the repository:
git clone https://github.com/salama4ai/salama4ai-chatbot.git
to install rasa visit https://www.youtube.com/playlist?list=PL75e0qA87dlEWUA5ToqLLR026wIkk2evk``` install anaconda or miniconda
to ensure that you installed anaconda successfully
conda -V
then open anaconda prompt, and update conda
conda update conda
to create new virtual environment
conda create -n yourenvname python=3.8
you must install python3.8 or 3.7 as rasa compatable with these two
create new folder and go to it and activate you environment
conda activate test_env
install rasa
pip install rasa
initialize you first project
rasa init
if you need to train your bot after modifing files
rasa train
if you need to interact with your bot
rasa shell
or
rasa interactive
if you use rasa interactive
so you can train the bot by chatting and interacting with it
if you write custom action so you need to open anaconda prompt and activate the bot and in the prompt run
rasa run actions
and leave it open in the background to be able to use the custom action with another prompt
you also can run shell in debug mode using rasa shell --debug
and you can use rasa train --force
to enforce retraining
if you make changes in action.py file you don't need to retrain again just rerun rasa run actions
this work done by Ahmed Salama at JAN-2022