ai_coder

Developing and Contributing

From the root of this repo:

Create a virtual env and activate.

python3 -m venv .venv

How you activate the environment depends on your operating system and shell. Typically, it involves sourcing an activate script located in the virtual environment's bin (Unix-based) or Scripts (Windows) directory.

For Unix-based systems:

. ./.venv/bin/activate

For Windows Command Prompt:

.\.venv\Scripts\activate

Install ai_coder into your pip environment with:

pip install -e .

This installs the ai_coder package in editable mode. All imports of ai_coder will run the ai_coder source at ./ai_coder directly.

Use ai_coder in another project (Let's see Project B)

Step 1: Identify the Path to ai_coder

First, you need to find the absolute path to ai_coder's directory. This is the directory that contains the setup.py file. You can navigate to this directory in your terminal and use a command to get the path:

• On Unix-based systems (Linux, macOS), you can use pwd.
• On Windows, you can use cd without arguments or echo %cd% in Command Prompt (or echo $PWD in PowerShell).

Step 2: Activate Project B's Virtual Environment

Before you install ai_coder as a dependency in Project B, make sure Project B's virtual environment is activated.

Step 3: Install ai_coder in Editable Mode into Project B's Environment

With Project B's virtual environment activated, use pip to install ai_coder in editable mode using the path you obtained in Step 1. Replace /path/to/projectA with the actual path to ai_coder:

pip install -e /path/to/ai_coder

This command will install ai_coder into Project B's virtual environment in editable mode. Any changes made to ai_coder's code will be reflected in Project B without needing to reinstall ai_coder.

Test

1. Setup OPENAI API KEY

It currently executes the prompt with OpenAI, so you'll need to use your own OpenAI API Key

export OPENAI_API_KEY=sk-abcdefghijklmnopqrstuvwxyz

2. Add below code to prompts/src/my_code.py

from ai_coder.ai_decorators import ai_code

@ai_code
def get_top_hacker_news():
    return """Implement a function that can fetch the top hacker news, which mention certain keywords, e.g. 'gpt', 'llm', 'workflow', 'serverless'.
    You can use the url \"https://hacker-news.firebaseio.com/v0/topstories.json\" to fetch the top hacker news. 
    Please only return the story title and the url.
    The input is the keywords and the output is a list of the news."""

3. Run

ai_coder gen prompts/src/my_code.py