docullim 0.5.0

auto-generate documentation for python code using llms

📝 docullim

auto-generate documentation for python code using llms.

⚙️ installation

pip install docullim

poetry add docullim

💻 usage

add @docullim to the function or class you want to generate documention for. you can also pass a tag to the annotation like @docullim("custom_tag").

from docullim import docullim

@docullim
def add(a, b):
    return a + b

@docullim("custom_tag")
def sub(a, b):
    return a - b

docullim file1.py file2.py
docullim "src/**/*.py"
docullim --config docullim.json --model gpt-4 "src/**/*.py"
docullim --reset-cache --concurrency 3 --write file1.py "src/**/*.py"

config file is a json file that can have model as string, max_concurrency as number or prompts as string pairs.

docullim.json

{
  "model": "gpt-4",
  "max_concurrency": 5,
  "prompts": {
    "default": "Generate short and simple documentation explaing the code and include sample usage.",
    "custom_tag": "this is a a diffrent propmt being passed to the llm when @docullim('custom_tag') is passed"
  }
}

you should provide your llm api key provided as enviroment variable by default it requires OPENAI_API_KEY. you can switch models and provide other llm api keys. See supported llms

🛠️ development

Install the following tools are needed to have this project running

• devbox manages all the packages needed by the project
• direnv loads env variables and run devbox when in project directory

also rename .env.example to .env and add your llm api key.