This repository contains a simple Spoken Dialogue System regarding the COVID-19 Emergency. The chatbot task is to reason and argue with the user regarding the pandemic and eventually update its belief of the world through spoken interaction. The chatbot follows a pipeline structure in which voice input is transformed into text, later processed. Via regular expressions, input phrases are categorized and subsequently handled by the internal class Agent.
The project is developed by:
- Rossella Bonuomo
- Emanuele Giacomini
- Veronica Vulcano
For further information regarding the implementation, you may take a look here.
First install libespeak1 for the speech synthetizer and python-piaudio for the Automatic Speech Recognition module.
In Linux Terminal
sudo apt install libespeak1 python-pyaudio python3-piaudio
If any problem occurs with the installation of python(3)-pyaudio, you may follow this guide.
If you do not use Anaconda/Miniconda, you can install the project dependencies through pip by running the following command
pip3 install SpeechRecognition pyttsx3
On the other hand, you can use Anaconda to first build a custom environment, like the following
conda create -n shri python=3.8
conda activate shri
pip install SpeechRecognition pyttsx3
Once all dependencies are installed, you can clone this repository
git clone https://github.com/EmanueleGiacomini/SHRI_chatbot.git
cd ./SHRI_chatbot/
At this point you can launch the chatbot by running the following code:
python ./main.py