Declarative HTML templating system with lazy rendering
Project description
HTML Generator
A python package to generate HTML from a template which is defined through a tree of render-elements.
Why
-
Guaranteed correct markup: Don't worry about correct markup-structure anymore, never again count closing divs or get upset about a missing slash.
-
Code formatting included (if you format your Python code): Don't worry what would be the correct and most consistent way of formatting your HTML, never again search desperately for a Django-template-language-formatter for your IDE. (:heart: black).
-
Easily generate and modify any template: Don't worry about overwritting 3rd-party html templates just to change a single letter anymore, never again cramp your templates with an infinite number of
{% if feat %}-statements trying to cover all possible use cases. -
Generate HTML in the same manner that you write Python code: Use functions to generate parameterized html objects or build a custom declarativ-like system to compose output
-
Keep the advantages of lazy rendering: Render contexts and lazy values allow for implicit dependencies, like traditional templates.
-
Define your own components: Subclassing the base elements allows for easy implementation of e.g. a custom design or layout system.
For more in depth reasoning and thoughts check out this great articel: https://www.devever.net/~hl/stringtemplates
Getting started
Installing:
pip install htmlgenerator
A simple example:
import htmlgenerator as hg
my_page = hg.HTML(hg.HEAD(), hg.BODY(hg.H1("It works!")))
print(hg.render(my_page, {}))
This will print the following HTML-document:
<!DOCTYPE html><html ><head ><meta charset="utf-8" /><meta name="viewport" content="width=device-width, initial-scale=1.0" /></head><body ><h1 >It works!</h1></body></html>
Note that the provided implementation of the HTML tag and the HEAD tag come with sensible defaults for DOCTYPE and META tags.
HTML elements
The package htmlgenerator defines all standard HTML tags with uppercase classnames, e.g. BODY, DIV, SECTION, etc.
The init method of HTML elements will treat all passed arguments as child elements and keyword arguments as attributes on the HTML element. Leading underscores of attribute names will be removed (to allow class and for attributes to be specified) and underscores will be replaced by dashes because python does not allow identifiers to have a dash and HTML attributes normally do not use underscores.
Example:
from htmlgenerator import render, DIV, OL, LI
print(
render(
DIV(
"My list is:",
OL(
LI("not very long"),
LI("having a good time"),
LI("rendered with htmlgenerator"),
LI("ending with this item"),
),
),
{},
)
)
print(
render(
DIV(
"Hello world!",
_class="success-message",
style="font-size: 2rem",
data_status="ok",
),
{},
)
)
Output:
<div class="success-message" style="font-size: 2rem" data-status="ok">Hello world!</div>
<div>My list is:<ol><li>not very long</li><li>having a good time</li><li>rendered with htmlgenerator</li><li>ending with this item</li></ol></div>
Rendering
The method htmlgenerator.render should be used to render an element tree. All nodes in the tree should inherit from htmlgenerator.BaseElement. Leaves in the tree can be arbitrary python objects. The render function expects the root element of the tree and a context dictionary as arguments.
The output is generated by rendering the tree recursively. If an object in the tree is an instance of BaseElement its render method will be called with the context as single argument and it must return a generator which yields objects of type str. Otherwise the object will be converted to a string and escaped for use in HTML. In order to prevent a string from beeing escaped htmlgenerator.mark_safe or django.utils.html.mark_safe from the django package can be used.
Example python object:
import datetime
from htmlgenerator import render, DIV
print(render(DIV("Hello, here is some date: ", datetime.date.today()), {}))
print(
render(
DIV(
"Hello, here is some data: ",
{"fingers": [1, 2, 3, 4, 5], "stuff": {"set": {1, 2, 3, 3, 3, 3, 3}}},
),
{},
)
)
Output:
<div>Hello, here is some date: 2020-11-20</div>
<div>Hello, here is some data: {'fingers': [1, 2, 3, 4, 5], 'stuff': {'set': {1, 2, 3}}}</div>
Example render object:
from htmlgenerator import render, DIV
class DoStuff:
# be aware that all yielded strings will not be separated by spaces but concatenated directly
def render(self, context):
yield "eat "
yield "sleep "
yield "program"
print(render(DIV("I like to ", DoStuff()), {}))
Output:
<div>I like to eat sleep program</div>
Example escaping:
from htmlgenerator import mark_safe, render, DIV
print(
render(
DIV(
"Hello, here is some HTML: ",
'<div style="font-size: 2rem">Hello world!</div>',
),
{},
)
)
print(
render(
DIV(
"Hello, here is NOT some HTML: ",
mark_safe('<div style="font-size: 2rem">Hello world!</div>'),
),
{},
)
)
Output:
<div>Hello, here is some HTML: <div style="font-size: 2rem">Hello world!</div></div>
<div>Hello, here is NOT some HTML: <div style="font-size: 2rem">Hello world!</div></div>
Lazy values
In order to allow rendering values which are not yet known at construction time but only at render time lazy values can be used. By default htmlgenerator comes with the following lazy values:
htmlgenerator.ContextFunction: Calls a function with the context as argument and renders the returned value (shortcuthtmlgenerator.F)htmlgenerator.ContextValue: Renders a variable from the context, can use . to access nested attributes or dictionary keys (shortcuthtmlgenerator.C)
A lazy value will be resolved just before it is rendered. Custom implementations of lazy values can be added by inheriting from htmlgenerator.Lazy.
Example:
from htmlgenerator import render, DIV, C, F, ATTR
print(
render(
DIV("Hello, ", C("person.name")),
{"person": {"name": "Alice", "occupation": "Writer"}},
)
)
print(render(DIV("Crazy calculation: 4 + 2 = ", F(lambda context: 4 + 2)), {}))
Output:
<div>Hello, Alice</div>
<div>Crazy calculation: 4 + 2 = 6</div>
<div>This text is wrapped inside a div element</div>
Virtual elements
In order to allow the building of a dynamic page virtual elements need to be used. The following virtual elements exist:
htmlgenerator.BaseElement: The base for all elements, can also be used to group elements without generating output by itselfhtmlgenerator.If: Lazy evaluates the first argument at render time and returns the first child on true and the second child on falsehtmlgenerator.Iterator: Takes an iterator which can be a lazy value and renders the child element for each iterationhtmlgenerator.Fragment: Allows to name a part of the tree which can then selectively be rendered by setting thefragmentparameter when callinghg.render(inspired by https://htmx.org/essays/template-fragments/)
Example:
from htmlgenerator import render, SPAN, BaseElement, If, C, Iterator, Fragment, DIV, BODY
print(
render(BaseElement("Just", SPAN("some"), "elements", SPAN("without"), "parent"), {})
)
print(render(If(C("cold"), "It is cold", "It is not cold"), {"cold": True}))
print(render(If(C("cold"), "It is cold", "It is not cold"), {"cold": False}))
print(render(Iterator(range(7), SPAN("I love loops ")), {}))
print(render(BODY("HTML body", Fragment("content", DIV("Content of my page"))), {}))
print(render(BODY("HTML body", Fragment("content", DIV("Content of my page"))), {}, fragment="content"))
Output:
Just<span>some</span>elements<span>without</span>parent
It is cold
It is not cold
<span>I love loops </span><span>I love loops </span><span>I love loops </span><span>I love loops </span><span>I love loops </span><span>I love loops </span><span>I love loops </span>
<body>HTML body<div>Content of my page</div></body>
<div>Content of my page</div>
Converting existing HTML source
htmlgenerator comes with a handy commandline tool to convert existing HTML-code into htmlgenerator python objects.
In order to get all the dependencies for the tool, install it with pip install htmlgenerator[all]
It can be used with standard input or with a list of files as arguments:
echo '<div class="btn" style="padding: 2rem">Click me</div>' | convertfromhtml > mytemplate.py
convertfromhtml template1.html template2.html # will result in template1.html.py and template2.html.py
By default the generated python files are formatted with black.
Python code which has been generated from very large files, e.g. complete websites, might take multiple minutes to be formated.
In order to get unformatted but still valid python code add the flag --no-formatting.
This will not run black on the generated python code and therefore be very fast.
In order to generate more compact code the flag --compact can be passed to convertfromhtml.
This will generate the most compact python code and works with and without --no-formatting.
In order to get human readable code the flag --compact is recommended.
In order to get code fast (especially for big pages) the flag --no-formatting is recommended.
In order to get a one-liner use --compact and --no-formatting.
Notes regarding the encoding of the input file:
In case the encoding of the input HTML source file differs from the default
encoding of the operating system, there is the flag --encoding <encoding>
which allows to enforce the given encoding when reading the file. This is
likely the case on Windows systems. It is therefore recommended to use
--encoding utf8 when using convertfromhtml on windows.
Django integration
In order to use the element tree renderer in django html templates it is necessary to add a template tag which calls the render function.
@register.simple_tag(takes_context=True)
def render_document(context, root):
return mark_safe(layout.render(root, context.flatten()))
The render method of any object may also be directly passed to a HttpResponse object. This is useful if htmlgenerator should generate the complete page in function based views.
Example of a helper function to render an element tree to a response (layout is the element tree):
from django.http import HttpResponse
def render_layout_to_response(request, layout, context):
return HttpResponse(layout.render(context))
Rational
Notes:
- Want internal/embeded DSL, access to document object model, not string interpolation
- https://wiki.python.org/moin/Templating ==> All dead links
- Existing HTML-DSL have no support for more abstract concepts like iterators and lazy values
- We want something inspired by FRP (Functional reactive programming) and LISP like XAML (but without XML) or React/Vue (but not on the client-side)
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file htmlgenerator-1.2.30.tar.gz.
File metadata
- Download URL: htmlgenerator-1.2.30.tar.gz
- Upload date:
- Size: 24.1 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.2 CPython/3.12.1
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | 4153c37ae5d6971ff3f4d52fb236865bbc592ef24605dc075cef704a0f76bdc4 |
|
| MD5 | e9ae28b46b7004bc2c94c014a0d2819e |
|
| BLAKE2b-256 | 9b687822202de3ed75565bce9096bcfd1210514f590c3382c1af165001de2ca9 |
File details
Details for the file htmlgenerator-1.2.30-py3-none-any.whl.
File metadata
- Download URL: htmlgenerator-1.2.30-py3-none-any.whl
- Upload date:
- Size: 22.7 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.2 CPython/3.12.1
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | 31d7837f59998c9da5c54f01639e90ec9b380a6ea4031641675a06e24d111408 |
|
| MD5 | 1f7c76391024eb1f702c3e519ca49d32 |
|
| BLAKE2b-256 | 65a6c8bb39809b29d001fd126d6c40a0a141c58af7e01149b01a288574aebbeb |
