html templating with python (now http://github.com/mdamien/lys)

example.py

from html42 import H, render

def _base(content='', title_prefix=''):
    return render(H.html() / (
      H.head() / (
        H.meta(charset="utf-8"),
        H.meta(content="width=device-width, initial-scale=1", name="viewport"),
        H.title() / (title_prefix + "Chrome Extensions Archive"),
        H.link(href="/style.css", media="screen", rel="stylesheet", type="text/css"),
      ),
      H.body() / (
        H.a(href='/') / (H.h1() / "Chrome Extensions Archive"),
        H.div(style='text-align: right') /
          H.a(href="https://github.com/mdamien/chrome-extensions-archive") /
            "github.com/mdamien/chrome-extensions-archive",
        H.hr(),
        content,
      ),
    ))


print(render(_base('hello world'))

html42.py

import html, types

def render(node):
    assert node is not None  # TODO: None => '' ?

    if type(node) in (tuple, list, types.GeneratorType):
        return ''.join(render(child) for child in node)

    if type(node) == str:
        return html.escape(node)

    children_rendered = ''
    if node.children:
        children_rendered = render(node.children)

    attrs_rendered = ''
    if node.attrs:
        # TODO: really, should I ignore the None values ?
        # TODO: class_, is_, ...
        attrs_rendered = ' ' + ' '.join(
            html.escape(key.replace('class_', 'class')) + '="' + html.escape(value) + '"'
            for key, value in node.attrs.items()
            if value is not None)

    if node.tag in ('br', 'img', 'hr'):
        assert not node.children
        return '<{tag}{attrs} />'.format(tag=node.tag, attrs=attrs_rendered)

    return '<{tag}{attrs}>{children}</{tag}>'.format(
        tag=node.tag, children=children_rendered, attrs=attrs_rendered)

class Node:
    def __init__(self, tag, attrs=None, children=None):
        assert tag
        self.tag = tag
        self.attrs = attrs
        self.children = children

    def __truediv__(self, children):
        if type(children) not in (tuple, list):
            children = (children,)
        self.children = children
        return self

    def __str__(self):
        return render(self)

class _H:
    def __getattr__(self, tag):
        return lambda **attrs: Node(tag, attrs)
H = _H()

TODO.md

• django integration example
• auto-converter from HTML and Django Templates (and other templating languages)
• self-closing tags (br, img,..)
• special attributes (class, is,..)
• custom anchor point for children
• support escaped string (markdown rendered, javascript,...)

other ideas:

• use beautifulsoup to store the node ? nice querying, prettify,..

• what to do with the a() / b() / c() trap ?

• also a() / b() + c() trap, change operator to high priority instead ?

  • content can go as first arg too (H.span('the dress is green'))
    • if this is used, forbid a new / since it's counter-intuitive

• less typing

  • H.span is valid too
  • H('#main.red.big')to create <div id='main' class='red big'>

• change main letter to L ?

• safe execution (sandbox) so user write the stuff directly

---

• clear error & clear when a / b / c
• custome node
• good repr()
• import * des balises
• https://github.com/blakev/lys/commit/f8374b4d76dc28e00e77f869700d2bcc6a44d032

---

ideas after making a website:

TypeError: call() takes from 1 to 2 positional arguments but 3 were given

=> do not return the Lys object directly, do str(node) befores

BASIC "HELP ME PLEASE" MODE WHERE IT PREVENT BASIC ERROR (mystyped css, atributes,...)

would be nice to have live editing (save the round trip from chrome)

autogenerated class names