aydie-prompt-forge

A modern toolkit for creating, managing, and testing LLM prompts as code.

aydie-prompt-forge is a lightweight but powerful Python library designed to bring standard software engineering discipline to the world of prompt engineering. It helps you move away from messy, hardcoded strings and manage your prompts in a structured, version-controlled, and testable way.

This library is part of the Aydie family of developer tools.

Why aydie-prompt-forge?

In modern AI applications, prompts are not just text—they are a critical part of your application's logic. They deserve to be treated like code. aydie-prompt-forge provides the framework to do just that.

• Reliability & Reproducibility: Stop guessing how a prompt change will affect your app. With a structured repository, you can version and test your prompts to ensure consistent, high-quality outputs.
• Improved Collaboration: Enable product managers, developers, and QA engineers to collaborate seamlessly. Prompts are defined in simple, human-readable YAML files, separate from the application code.
• Increased Efficiency: Automate the loading, templating, and management of your prompts. Spend less time on boilerplate and more time on crafting the perfect prompt.
• Robust Error Handling: With custom, branded exceptions (AydieException), you can write clean, predictable error-handling logic for all prompt-related operations.

Core Features

• Structured Prompt Repository: Define all your prompts in a clean, simple YAML file with rich metadata like version, author, tags, and description.
• Powerful Templating: Easily load prompts and fill them with dynamic data using a simple, intuitive .fill() method.
• Version Control Friendly: Because your prompts are stored as text, you can use Git to track changes, review diffs, and manage your prompt history just like any other codebase.
• Lightweight & Dependency-Free: With only PyYAML as a dependency, aydie-prompt-forge is easy to add to any project without unnecessary bloat.

Installation

You can install aydie-prompt-forge directly from PyPI:

pip install aydie-prompt-forge

Quick Start Guide

Getting started with aydie-prompt-forge is easy.

1. Create your prompts.yml file

First, create a prompts.yml file to store your prompts. This file should contain a list of prompt objects.

prompts.yml:

- id: summarize_article_v1.1
  author: "Aydie"
  version: "1.1"
  description: "Summarizes a long article into three concise bullet points."
  tags: [summarization, production]
  model_parameters:
    temperature: 0.6
    max_tokens: 150
  template: |
    You are a world-class editor. Summarize the following article into
    exactly three concise bullet points.

    ARTICLE:
    ---
    {article_text}
    ---

- id: sentiment_analysis_v1
  author: "Aydie"
  version: "1.0"
  description: "Analyzes the sentiment of a user's comment."
  tags: [classification, beta]
  template: "Is the sentiment of the following comment positive, negative, or neutral?\n\nCOMMENT: {user_comment}"

2. Load and use your prompts in Python

Now, you can use the load() function to parse your repository and use your prompts.

main.py:

import aydie_prompt_forge

# Define the path to your prompt repository
PROMPT_FILE = "prompts.yml"

try:
    # 1. Load the entire repository from the file.
    repo = aydie_prompt_forge.load(PROMPT_FILE)
    print("Repository loaded successfully!")

    # 2. Get a specific prompt by its unique ID.
    summarize_prompt = repo.get("summarize_article_v1.1")

    if summarize_prompt:
        # 3. Prepare your dynamic data.
        long_article = "Generative AI is a type of artificial intelligence technology that can produce various types of content, including text, imagery, audio and synthetic data."

        # 4. Fill the prompt template with your data.
        final_prompt = summarize_prompt.fill(article_text=long_article)

        # The final_prompt is now ready to be sent to an LLM API!
        print("\n--- Generated Prompt ---")
        print(final_prompt)
        print("------------------------")

        # You can also access the metadata
        print(f"\nModel parameters to use: {summarize_prompt.model_parameters}")

except aydie_prompt_forge.AydieException as e:
    # Catch any errors from the aydie-prompt-forge library, like a malformed file.
    print(f"An error occurred with aydie-prompt-forge: {e}")
except FileNotFoundError:
    print(f"Error: The file '{PROMPT_FILE}' was not found.")

Contributing

Contributions are welcome! If you have an idea for a new feature, find a bug, or want to improve the documentation, please open an issue or submit a pull request on our GitHub repository.

License

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

Connect with Me