Dynamic Interactive documentation made easier.

DocEaser is a very simple framework used to render Markdown documents. The twist is that those documents can embed HTMX components written in Python, to add interactivity to your documentation.

$ pip install git+https://github.com/linkdd/doceaser

Create a folder for your content:

$ mkdir content/
$ mkdir content/_static
$ python -m doceaser.cli --root content/
Listening on http://localhost:8080/

The server is based on the production-ready WSGI server waitress.

Static files go in the content/_static/ directory and will be served at /_static/... URLs:

HTMX components go in the content/ directory as Python files, and will be served at /_components/... URLs:

Markdown files go in the content/ directory, and will be served at /... URLs. The special file _index.md can be used as an index section:

Markdown files may have a frontmatter header (using YAML, TOML or JSON):

---
title: Hello world
---

# Hello World

The title property will be used for the <title/> HTML tag.

To embed an HTMX component, you can use this special syntax (an extension of the CommonMark specification):

{< htmx:form >}

{< htmx:hello/world >}

This will render to the following HTML (with some extra irrelevant details):

<div hx-get="/_components/form" hx-trigger="load"></div>

<div hx-get="/_components/hello/world" hx-trigger="load"></div>

An HTMX component is a Python file in the content/ directory. This file may implement a function per HTTP method to handle the request. That function accepts a werkzeug Request and returns a werkzeug Response:

from werkzeug.wrappers import Request, Response
from http import HTTPStatus

from lxml.html.builder import E as e
from lxml import html

import requests


def get(request):
    r = requests.get("https://httpbin.org/anything")
    body = e.pre(e.code(r.text))

    return Response(
        html.tostring(body),
        status=HTTPStatus.OK,
        content_type="text/html; charset=utf-8",
    )

NB:

• You can use any method you like to render the HTML. Internally, DocEaser uses lxml, so you might as well use it. But it is not mandatory.
• Any extra Python dependencies you might want shall be installed by you (like requests in the example above).

None at the moment except this README and a few examples in the examples/ directory.

This project is mostly a Proof of Concept. Nothing has been planned ahead (yet). I will probably not actively develop it, but I can at least guarantee that bugs will be addressed, and feature requests will be discussed (and maybe implemented?).

Here are a few ideas that might be nice, from the top of my head:

• Make the website layout customizable: At the moment, I simply generate a single page with Bulma as the CSS framework, everything is tightly coupled to it (fine for a PoC, not ideal for a finished product)
• Cache the Markdown rendering: Everything is rerendered from scratch on every call. This might not be the best for performance.

Any other ideas, and Pull Requests are more than welcome! Give a look to the CONTRIBUTING file.

This software is released under the terms of the MIT License.