Arabic-Conversational-Assistant: A Python repository from skawy

ARCA

Arabic Conversational Assistant

🏄 Introduction

The purpose of this repository is to showcase a contextual AI assistant built with the open source Rasa framework.

ARCA is an alpha version giving FCAI students Academic Advising. It aims at helping each student identify his/her strengths and abilities to assist him/her in making decisions relevant to his/her studies and specialization and in overcoming any impediments at college.

The chatbot should be able to the following basic functionalities

Helping you to make your own schedule
knowing whether you can study this subject or not
which major is most suitable to you based on your grades

👉 Installation

To install ARCA, please clone the repository and run:

cd FCAI_CU_Chatbot
pip install -r requirements.txt | pip3 install -r requirements.txt

This will install the bot and all of its requirements. Note that this bot should be used with python 3.6 or 3.7.

📂 To Add Components

We Have custom components to add to our pipeline for preprocessing incoming input, but we need first to add these components to file called registry.py in rasa package files

so open **your_virtual_environment_path **/lib/python3.7/site-packages/rasa/nlu/registry.py

add the following code:

import sys

folder_path = "your_full_path/FCAI_CU_Chatbot/components"
sys.path.insert(1, folder_path)

from input_preprocessing import Cleaning_Arabic_Text,Print_Clean_Text

add in the component_classes list in the same file "regisry.py" add the imported classes:

component_classes = [
    # Arabic Preprocessing
    Cleaning_Arabic_Text,
    Print_Clean_Text,
    # utils
    SpacyNLP,

Now all files is ready to be trained.

🤖 To run ARCA

Use rasa train to train a model

Then, to run, first set up your action server in one terminal window:

rasa run actions

Finally, in the second terminal window you can run and start your conversation

rasa shell

Note that this bot should be used with Rasa 2.8

📘 Overview of the files

data/stories/ - contains stories

data/rules/ - contains rules

data/nlu - contains NLU training data

actions - contains custom action code

domain.yml - the domain file, including bot response templates

config.yml - training configurations for the NLU pipeline and policy ensemble

🔉 To Run Voice Bot (optional)

Use rasa train to train a model

These are the needed dependencies

pip install SpeechRecognition
pip install playsound
pip install pipwin
pipwin install pyaudio
pip install gTTS
pip install requests

First Terminal

rasa run -m models --endpoints endpoints.yml --port 5002 --credentials credentials.yml

Second Terminals

rasa run actions

Finally you can run the voice_bot.py from IIDE or terminal.

Note that the bot must be trained before running this python file.

skawy/Arabic-Conversational-Assistant

ARCA