Indy-Bot

A Rasa-based conversational AI assistant built in Python and HTML.

---

Table of Contents

• Overview
• Features
• Project Structure
• Installation
• Usage
• Training
• Custom Actions
• Configuration
• Troubleshooting
• Contributing
• License

---

Overview

Indy-Bot is a chatbot built with the Rasa framework. It leverages intent recognition, dialogue management, and customizable actions to engage with users in a conversational experience.

---

Features

• Intent-driven conversations powered via Rasa NLU (see domain.yml)
• Stories or conversation flows defined to guide dialogs (see config.yml)
• Custom actions implemented in actions.py to integrate dynamic functionality or backend logic
• HTML templates for crafting response messages (folder: templates/)
• Pre-trained or sample models stored in the models/ directory

---

Project Structure

Indy-Bot/
├── .rasa/             # Rasa internal caching (auto-generated)
├── data/              # Training data (e.g., NLU and Stories)
├── models/            # Saved Rasa model files
├── templates/         # HTML or response templates
├── actions.py         # Custom action code
├── app.py             # Main application file to run the bot
├── config.yml         # Rasa configuration (pipeline & policies)
├── domain.yml         # Intents, entities, slots, responses, actions
├── indiana_chatbot.db # Optional: sample or persistent database
├── LICENSE            # MIT open-source license
└── README.md          # Placeholder — to be replaced with this version

---

Installation

1. Clone the repository:

   git clone https://github.com/nehajosephh/Indy-Bot.git
   cd Indy-Bot

2. Set up a Python virtual environment (recommended):

   python3 -m venv venv
   source venv/bin/activate

3. Install dependencies:

   pip install rasa

   Add any other required packages as needed (e.g., rasa-sdk, database drivers, etc.).

---

Usage

Training the Model

If training data exists in data/:

rasa train

This generates a new model inside the models/ directory.

Running the Bot

To launch the Rasa server with the action server:

rasa run actions &
rasa shell

Alternatively, use app.py if it provides a custom interface:

python app.py

---

Training

Training configurations are defined in config.yml:

• Pipeline: Defines NLU components (e.g., tokenizers, intent classifiers)
• Policies: Dictates dialog management behavior

Make sure to adjust hyperparameters or components as needed for your use case.

---

Custom Actions

Custom logic goes in actions.py. Typical structure:

from rasa_sdk import Action, Tracker
from rasa_sdk.executor import CollectingDispatcher

class YourCustomAction(Action):
    def name(self):
        return "action_name"

    def run(self, dispatcher, tracker, domain):
        # Your action logic here
        dispatcher.utter_message(text="Your response")
        return []

To register actions:

• Ensure domain.yml lists the action in the actions: section.
• Run the action server alongside the Rasa server (rasa run actions).

---

Configuration

Key files:

• domain.yml: Defines intents, responses, slots, entities, and actions.
• config.yml: Setup of NLU pipeline and dialog policies.
• templates/: Stores HTML or rich response templates.

Make any necessary edits to align with your conversational goals.

---

Troubleshooting

• Training errors: Ensure training data formats and domain definitions are correct.
• Actions not working: Double-check naming, ensure action server is running.
• Template issues: Confirm proper paths and template syntax.

---

Contributing

Contributions are welcome! Feel free to:

• Add new intents, stories, or responses
• Enhance custom actions
• Improve conversational flows or integrations
• Submit pull requests or file issues for improvements

---

License

This project is licensed under the MIT License. See the LICENSE file for more details.

---