ArtworkBookingChatbot-Sinhala

This repository contain source code for Sinhala artwork booking chatbot created using RASA.

Directory Structure

The important files and directories of the repository

├── data :                  
    ├── nlu.yml : training data for intent classification and entity extraction
    ├── rules.yml : rules for rasa core model training
    ├── stories.yml : sample stories for rasa core training
├── models :
    ├── Trained models (in .tar.gz format) 
├── actions :
    ├── action.py : custom actions
├── results :
    ├── results of model evaluations (confusion matrices, histograms, etc)
├── domain.yml : training data for action prediction
├── config.yml : model configurations
├── endpoints.yml : action server endpoint
├── credentials.yml
├── index.html : chatbot api interface

Dataset properties

Intents:

• service, inform_medium, inform_style, inform_figures, inform_background, inform_email, ask_samples, out_of_scope, greet, thank_you, goodbye, bot_challenge

Stories:

• ask service path, ask samples path

Getting started

Run these commands in seperate terminals/command promts:

To initiate the project & run rasa server

pip install rasa
python3 -m venv ./venv
.\venv\Scripts\activate
git clone https://github.com/AnuttharadeSilva/ArtworkBookingChatbot-Sinhala
cd ArtworkBookingChatbot-Sinhala
rasa run --enable-api -cors "*" 

To run action server

rasa run actions 

To run web api

python -m http.server 

Navigate to http://localhost:8000/