""" *** Original blogpost: http://tomerfiliba.com/blog/Hypertext/ Hypertext: "It's Snakes all the Way Down" ========================================= Who needs ``haml`` or templating languages? Hypertext is an in-language DSL for producing perfectly valid (and safe) XHTML directly from Python code! It's a bit magical, that's true, but it's thread-safe and it makes rendering HTML pages almost as compact as ``haml``. Creating HTML elements is easy and straightforward :: >>> import hypertext as H >>> x = H.h1("hello", class_ = "highlight") >>> x Invoking ``str`` (or via ``print``) on an HTML element renders it textually :: >>> print x

hello

Alternatively, if you're not afraid of polluting your namespace :: >>> from hypertext import * >>> print h1("hello", class_ = "highlight")

hello

And there's a shortcut for specifying the class (a la ``haml``) :: >>> print h1.highlight("hello")

hello

Which can be combined for specifying multiple classes... :: >>> print h1.highlight.important("hello")

hello

Element can be nested :: >>> print div.content(h1.highlight("hello"))

hello

But also using ``with`` statement >>> with div.content as root: # doctest:+ELLIPSIS ... h1.highlight("hello") ... ,,, >>> print root

hello

.. note:: Don't mind the ``,,,``, it's there to make ``doctest`` happy The attributes and text of elements can be specified after their construction :: >>> with div.content as root: ... with h1: ... ATTR(class_ = "highlight") ... TEXT("hello") ... >>> print root

hello

Text is HTML-escaped by default, but you can override it (if you trust its source) :: >>> with div.content as root: ... dangerous = "" ... TEXT("hello %s" % (dangerous,)) ... UNESCAPED("hello %s" % (dangerous,)) ... >>> print root

hello <script>alert('oh no!');</script> hello

Putting it all together:: >>> with html as doc: # doctest:+ELLIPSIS ... with head: ... meta(http_equiv="Content-Type", content="text/html; charset=utf-8") ... title("welcome to my page") ... with body: ... ATTR(id="body") ... with div.content(id = "floop", data_role = "page").pretty: ... TEXT("Hello, my is") ... strong("Bob") ... TEXT("Also known as", em("Robert")) ... UNESCAPED("and I like") ... with ul: ... for item in ["cats", "rats", "hats"]: ... li(item) ... ,,, >>> print doc welcome to my page

Hello, my <name> is Bob Also known as Robert and I like

cats
rats
hats

The use of ``with`` and the fact you're dealing with pure Python code makes things easy - there's no need for templating languages, passing parameters back and forth, extending base templates or including files. It's snakes all the way down! """ import threading __all__ = ["Element", "TEXT", "UNESCAPED", "ATTR", "EMBED", "THIS", "PARENT"] class Unescaped(str): __slots__ = () def __repr__(self): return "Unescaped(%s)" % (str.__repr__(self)) _MAPPING = {"&" : "&", "'" : "'", '"' : """, "<" : "<", ">" : ">"} def xml_escape(text): if isinstance(text, Unescaped): return str(text) else: return "".join(_MAPPING.get(ch, ch) for ch in str(text)) #=================================================================================================== # #=================================================================================================== _per_thread = threading.local() class Element(object): class __metaclass__(type): __slots__ = () def __getattr__(cls, name): return cls(class_ = name) def __enter__(cls): return cls().__enter__() def __exit__(cls, t, v, tb): _per_thread._stack[-1].__exit__(t, v, tb) __slots__ = ["_attrs", "_elems"] TAG = None INLINE = False DOCTYPE = None DEFAULT_ATTRS = {} def __init__(self, *elems, **attrs): self._attrs = self.DEFAULT_ATTRS.copy() self._elems = [] self._parent = None if getattr(_per_thread, "_stack", None): _per_thread._stack[-1]._elems.append(self) self._parent = _per_thread._stack[-1] else: self._parent = None self(*elems, **attrs) def __repr__(self): return "" % (self.TAG if self.TAG else self.__class__.__name__.lower(), ", ".join("%s=%r" % (k, v) for k, v in self._attrs.items()), len(self._elems)) def __str__(self): return render(self) def __enter__(self): if not hasattr(_per_thread, "_stack"): _per_thread._stack = [] _per_thread._stack.append(self) return self def __exit__(self, t, v, tb): _per_thread._stack.pop(-1) def __getitem__(self, name): return self._attrs[name] def __delitem__(self, name): del self._attrs[name] def __setitem__(self, name, value): self._attrs[name] = value def __call__(self, *elems, **attrs): for elem in elems: if isinstance(elem, Element): if elem._parent is not None: elem._parent._elems.remove(elem) elem._parent = self self._elems.append(elem) for k, v in attrs.items(): if k.endswith("_"): k = k[:-1] self._attrs[k.replace("_", "-")] = v return self def __getattr__(self, name): if "class" in self._attrs: self._attrs["class"] += " " + name else: self._attrs["class"] = name return self def THIS(): """Return the current HTML element on the stack; raises ``IndexError`` in case the stack is empty""" return _per_thread._stack[-1] def PARENT(count = 1): """Returns the ``count``-parent of the current HTML element on the stack; raises ``IndexError`` if there's no ``count`` parent. ``count`` defaults to 1, which means the immedaite parent""" return _per_thread._stack[-1 - count] def TEXT(*texts): """Appends the given texts (as well as HTML elements) to the current element""" THIS()(*texts) def UNESCAPED(*texts): """Appends the given texts unescaped to the current element. Note: security risk, use with caution""" THIS()(*(Unescaped(text) for text in texts)) def EMBED(element): """Embeds (appends) the given HTML element into the current element (transfers ownership)""" THIS()(element) def ATTR(**kwargs): """Sets the given keyword-arguments as attributes of the current HTML element""" THIS()(**kwargs) #=================================================================================================== # #=================================================================================================== def _render(elem, level, dont_indent = False): indent = " " * level if not isinstance(elem, Element): result = xml_escape(elem) else: tag = elem.TAG if elem.TAG else elem.__class__.__name__.lower().rstrip("_") attrs = " ".join('%s="%s"' % (k, xml_escape(str(v))) for k, v in elem._attrs.items()) if attrs: attrs = " " + attrs if elem._elems: elements = "\n".join(_render(e, level + 1) for e in elem._elems) if elem.INLINE or not "\n" in elements: elements = _render(elem._elems[0], level + 1, True) result = "<%s%s>%s" % (tag, attrs, elements, tag) else: result = "<%s%s>\n%s\n%s" % (tag, attrs, elements, indent, tag) else: result = "<%s%s/>" % (tag, attrs) if dont_indent: return result else: return indent + result def render(root): """renders the given HTML element and prepends ``DOCTYPE`` if one exists""" raw = _render(root, 0) if root.DOCTYPE: raw = root.DOCTYPE + "\n" + raw return raw #=================================================================================================== # HEAD elements #=================================================================================================== class html(Element): DEFAULT_ATTRS = {"xmlns" : "http://www.w3.org/1999/xhtml"} DOCTYPE = ('') class html5(Element): DOCTYPE = '' TAG = 'html' class head(Element): pass class link(Element): pass class meta(Element): pass class title(Element): INLINE = True class script(Element): DEFAULT_ATTRS = {"type" : "text/javascript"} class style(Element): DEFAULT_ATTRS = {"type" : "text/css"} #=================================================================================================== # BODY elements #=================================================================================================== class body(Element): pass class p(Element): pass class div(Element): pass class blockquote(Element): pass class dl(Element): pass class dt(Element): pass class dd(Element): pass class li(Element): pass class ul(Element): pass class ol(Element): pass class form(Element): pass class input(Element): pass class button(Element): pass class select(Element): pass class label(Element): pass class optgroup(Element): pass class option(Element): pass class textarea(Element): pass class legend(Element): pass class table(Element): pass class tr(Element): pass class th(Element): pass class td(Element): pass class colgroup(Element): pass class thead(Element): pass class tbody(Element): pass class tfoot(Element): pass class frame(Element): pass class iframe(Element): pass class noframe(Element): pass class frameset(Element): pass class pre(Element): INLINE = True class code(Element): INLINE = True class span(Element): INLINE = True class a(Element): INLINE = True class br(Element): INLINE = True class hr(Element): INLINE = True class em(Element): INLINE = True class strong(Element): INLINE = True class cite(Element): INLINE = True class h1(Element): INLINE = True class h2(Element): INLINE = True class h3(Element): INLINE = True class h4(Element): INLINE = True class h5(Element): INLINE = True class h6(Element): INLINE = True class i(Element): INLINE = True class b(Element): INLINE = True class u(Element): INLINE = True class sub(Element): INLINE = True class sup(Element): INLINE = True class big(Element): INLINE = True class small(Element): INLINE = True class img(Element): INLINE = True __all__.extend(cls.__name__ for cls in Element.__subclasses__()) if __name__ == "__main__": import doctest # nice trick from http://stackoverflow.com/a/9400829/434796 doctest.ELLIPSIS_MARKER = ",,," doctest.testmod()