Recipe Chatbot

Overview

The Recipe Chatbot is a Python-based application that helps users find easy and healthy recipes from allrecipes.com. Users can ask the chatbot for meal ideas, and it will scrape recipes from the website, format the results, and present them in a user-friendly format using Markdown syntax.

The application uses the langchain framework and the OllamaLLM model to generate responses, making it conversational and engaging.

---

Prerequisites

Before you begin, ensure you have the following installed on your system:

• Python 3.9 or later
• Docker (optional, for containerized usage)
• Git (optional, for cloning the repository)
• Ollama (for running the LLM)

---

Installation and Setup Instructions

1. Clone the Repository

git clone https://github.com/chrisprasanna/Local_RAG_Agent_for_Car_Manual.git
cd Local_RAG_Agent_for_Car_Manual

2. Create a Virtual Environment

python -m venv venv
.\venv\Scripts\activate  # For Windows
# source venv/bin/activate  # For macOS/Linux

3. Install Dependencies

Install the required Python libraries using pip:

python -m pip install --upgrade pip
python -m pip install -r requirements.txt

4. Pull the Required Ollama Models

Ensure you have Ollama installed locally, then pull the required model:

ollama pull llama3.2

---

Usage Instructions

Run the Application

To start the chatbot, run the following command:

python main.py

Interact with the Chatbot

• The chatbot will prompt you to ask a question about what you want to cook.
• Example: "What can I make with chicken and broccoli?"
• Type q or quit to exit the chatbot.

---

Overview of Components

1. main.py

   • The main application file that:
     • Accepts user input.
     • Scrapes recipes from allrecipes.com using urllib and BeautifulSoup.
     • Formats the recipe data and generates a response using the OllamaLLM model.
     • Includes a FastAPI-based health check endpoint for monitoring.

2. allrecipes_scraper.py

   • Contains the AllRecipesScraper class for scraping recipes from allrecipes.com.
   • Includes methods for searching recipes and fetching detailed recipe information.

3. requirements.txt

   • Lists all the dependencies required for the project.

4. Dockerfile

   • Defines the Docker image for containerizing the application.

5. tests/test_main.py

   • Contains unit tests for the fetch_recipe_from_allrecipes function.

6. .github/workflows/test.yml

   • GitHub Actions workflow for automated testing and Docker builds.

---

Dependencies

The project uses the following Python libraries:

• langchain-core: Core framework for building conversational AI applications.
• langchain-ollama: Integration with the Ollama LLM.
• beautifulsoup4: For parsing and extracting data from HTML.
• urllib3: For making HTTP requests.
• fastapi: For creating a health check endpoint.
• uvicorn: For running the FastAPI server.
• pytest: For testing.
• flake8 and black: For linting and formatting.

---

Docker Usage

Build the Docker Image

Ensure Docker is installed and running, then build the image:

docker build -t recipe-chatbot .

Run the Docker Container

Run the chatbot in a container:

docker run --rm -p 8080:8080 recipe-chatbot

Stop the Container

The container will stop automatically when the chatbot exits.

---

Limitations

• Web Scraping Dependency: The application relies on the structure of allrecipes.com. If the website changes its structure, the scraper may break.
• Limited Query Understanding: The chatbot may not handle complex or ambiguous queries well.
• Internet Requirement: The application requires an active internet connection to fetch recipes.
• Performance: Web scraping can be slow, especially for large queries or during high traffic on allrecipes.com.

---

Possible Improvements

• Improve query understanding using advanced NLP techniques.
• Cache results to reduce repeated web scraping.
• Add a web-based or mobile interface for easier interaction.
• Enhance error handling for failed web scraping or invalid user queries.
• Add support for multiple recipe websites.