doguda 0.1.1

Add your description here

Doguda quickstart

Define your commands in doguda_app.py so the doguda CLI can find them:

# doguda_app.py
from pydantic import BaseModel

from doguda import DogudaApp, doguda


class UrlToMarkdownResponse(BaseModel):
    markdown: str


app = DogudaApp()


@doguda  # registers on the default Doguda app
async def url_to_markdown(url: str) -> UrlToMarkdownResponse:
    return UrlToMarkdownResponse(markdown=f"received: {url}")

# Doguda will use your function's return annotation as the FastAPI response model,
# so you can shape responses with your own Pydantic models.

CLI

doguda url_to_markdown --url "https://example.com"

Set a different module with DOGUDA_MODULE=my_module or --module my_module.

HTTP

Start the FastAPI server (defaults to doguda_app in the current directory):

doguda serve

Then POST to your function at /v1/doguda/<function_name>:

curl -X POST http://localhost:8000/v1/doguda/url_to_markdown \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com"}'

Organizing commands in a package

You can also make doguda_app a package to auto-load submodules that register commands:

doguda_app/
  __init__.py
  urls.py       # contains @doguda functions
  reports.py    # contains @doguda functions

The loader imports all submodules, so any @doguda usages inside doguda_app/* are registered without extra wiring. Keep DOGUDA_MODULE=doguda_app (or --module doguda_app) and run doguda serve or doguda <command> as usual.

Running with uv (and flox)

If you're using uv from the flox env, activate it (or prefix with flox run) and keep the cache inside the repo to avoid permission issues:

UV_CACHE_DIR=$PWD/.flox/cache/uv uv run python -m doguda serve --module doguda_app