Prompt Frameworks

A practical guide and demonstration of different prompt engineering frameworks using Google's Generative AI.

Features

• Implementation of prompting methodology
• Interactive Jupyter notebook examples
• Ready-to-use prompt templates
• Real-world use cases and examples

Prerequisites

• Python 3.8+
• Google API key for Generative AI

Installation

1. Clone the repository:

git clone https://github.com/yourusername/prompt_frameworks.git
cd prompt_frameworks

2. Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate # on Windows, use `venv\Scripts\activate`

3. Install the required packages:

pip install -r requirements.txt

Configuration

1. Get your Google API key for Generative AI from Google AI Studio.

2. Create a .env file in the root directory of the project.

3. Add the following line to the .env file:

GOOGLE_API_KEY=your_api_key_here

Usage

1. Run the Jupyter notebook:

jupyter notebook # or jupyter lab

2. Open the prompt_frameworks_presentation.ipynb notebook and follow the instructions.

Optional - Running the Jupyter Notebook as a HTML Slideshow with RISE

Run the Classic Jupyter Notebook Interface

1. Launch the classic notebook interface from the terminal:

jupyter notebook

Firebase Studio will provide a public URL in the terminal output (e.g., http://localhost:8888/...). Open it in the browser.

2. Once open, select prompt_frameworks_presentation.ipynb.

3. Click the RISE (bar chart) icon in the toolbar to start the slideshow.

Note: RISE currently works best in the classic notebook interface, not JupyterLab

Export to Reveal.js HTML Slides

1. Convert the Jupyter notebook to a Standalone HTML slideshow using the nbconvert command-line tool:

jupyter nbconvert prompt_frameworks_presentation.ipynb --to slides --reveal-prefix https://unpkg.com/reveal.js@5.1.0/ --output presentation.html

2. Open the generated prompt_frameworks_presentation.slides.html file in a web browser to view the slideshow.

Note for Firebase Studio / Project IDX Users

If you're using Firebase Studio or Google Project IDX, you may encounter the following error when running a cell in Jupyter Notebook:

Running cells with 'venv (Python 3.11.10)' requires the ipykernel package.
Run the following command to install 'ipykernel' into the Python environment. 
Command: '/home/user/prompt_frameworks/fwenv/bin/python -m pip install ipykernel -U --force-reinstall'

This is caused by the absence of the GNU C++ runtime library (libstdc++.so.6), which is required by native Python packages such as pyzmq used by Jupyter.

Solution

1. Open your dev.nix file (located at the root of your project).
2. Add pkgs.gcc13 to the packages list:

{
 channel = "stable-24.05";

 packages = [
   pkgs.python311
   pkgs.python311Packages.pip
   pkgs.gcc13  # ✅ Adds libstdc++.so.6 needed for Jupyter
 ];

 env = {};

 ...
}

Contact / Social Media

• Twitter – @seetechnologic
• GitHub - https://github.com/JavaVista/
• LinkedIn - Javier Carrion
• Website - techno-logic.us

License

This project is licensed under the MIT License.