Skip to main content

Pythonic HTML generation/templating (no template files)

Project description

HTML5 Generation with html5tagger: Fast, Pure Python, No Dependencies

If you're looking for a more efficient and streamlined way to generate HTML5, look no further than html5tagger! This module provides a simplified HTML5 syntax, so you can create your entire document template using only Python. Say goodbye to the clunky and error-prone process of manually writing HTML tags.

With html5tagger, you can safely and quickly generate HTML5 without any dependencies, making it the perfect solution for developers who value speed and simplicity. And with its pure Python implementation, you'll never have to worry about compatibility issues or adding extra libraries to your project.

Ready to streamline your page rendering process? It is super fast to get started. Trust us, once you try html5tagger, you'll never go back to Jinja2 or manual HTML writing again!

pip install html5tagger

Intro

html5tagger provides two starting points for HTML generation: E as an empty builder for creating HTML snippets, or Document for generating full HTML documents with a DOCTYPE declaration. Both produce a Builder object, in case you need that for type annotations.

To get started, simply import both of them and start adding elements with dot notation:

from html5tagger import Document, E

# Create a document using template placeholders TitleText and Head
doc = Document(
    E.TitleText_,           # The first argument (optional) to add <title>
    lang="en",              # Keyword arguments for <html>

    # Just list the resources you need, no need to remember link/script tags
    _urls=[ "style.css", "favicon.png", "manifest.json" ]
)

# Upper case names are template variables. You can modify them later.
doc.Head_
doc.h1.TitleText_("Demo")   # Goes inside <h1> and updates <title> as well

# This has been a hard problem for DOM other such generators:
doc.p("A paragraph with ").a("a link", href="/files")(" and ").em("formatting")

# Use with for complex nesting (not often needed)
with doc.table(id="data"):
    doc.tr.th("First").th("Second").th("Third")
    doc.TableRows_

# Let's add something to the template fields
doc.Head._script("console.log('</script> escaping is weird')")

table = doc.TableRows
for row in range(10):
    table.tr
    for col in range(3):
        table.td(row * col)

# Or remove the table data we just added
doc.TableRows = None

You can str(doc) or just print it to get the HTML code. If used in a Jupyter Notebook or with any other system that supports __html__ or _repr_html_, it will render as HTML. When used via repr(doc) the templating tags are visible:

>>> doc
《Document Builder》
<!DOCTYPE html><html lang=en><meta charset="utf-8">
<title>《TitleText:Demo》</title>
<link href="style.css" rel=stylesheet>
<link href="favicon.png" rel=icon type="image/png">
<link href="manifest.json" rel=manifest>
《Head:<script>console.log('<\/script> escaping is weird')</script><h1>《TitleText:Demo》</h1>
<p>A paragraph with <a href="/files">a link</a> and <em>formatting</em>
<table id=data>
  <tr><th>First<th>Second<th>Third
  《TableRows》
</table>

The actual HTML output is similar. No whitespace is added to the document, it is all on one line, unless the content contains newlines. Similarly you may notice that body and other familiar tags are missing and that the escaping is very minimal. This is the HTML5 part: the document is standards-compliant with a lot less cruft.

The Document function creates a full document, and optionally adds typical headers that are hard to remember be yourself. This is in contrast to E that creates bare HTML snippets with no DOCTYPE or anything extra.

Everything generally chains, i.e. returns self, so that you can add more tags on one line.

Templating

Use template variables to build a document once and only update the dynamic parts at render time for faster performance. Access template variables via doc.TitleText and add content in parenthesis after the tag name. The underscore at the end of a tag name indicates the tag is added to the document and can have content in parenthesis, but any further tags on the same line go to the original document, not the template.

Nesting

⚠️ A tag is automatically closed when you add content to it, or prior to another tag.

In HTML5 elements such as <p> do not need any closing tag, so we can keep adding content without worrying of when it should close.

For elements like <table> and <ul>, you can use with blocks, pass sub-snippet arguments, or add a template variable. Unlike adding another tag, adding a template does NOT close its preceding element.

with doc.ul:  # Nest using with
    doc.li("Write HTML in Python")
    doc.li("Simple syntax").ul(id="inner").InnerList_  # Nest using template
    doc.li("No need for brackets or closing tags")
    doc.ul(E.li("Easy").li("Peasy"))  # Nest using (...)

Where unintentional nesting of a following element needs to be prevented, use (None) as content that causes the offending element to be kept empty.

Escaping

All content and attributes are automatically escaped. For instance, we can put the entire document into an iframe's srcdoc attribute where only the minimal but necessary escaping is applied. Use custom methods _script, _style and _comment for corresponding inline formats, to follow their custom escaping rules.

doc = Document("Escaping & Context")
doc._style('h1::after {content: "</Style>"}').h1("<Escape>")
doc._comment("All-->OK")
doc.iframe(srcdoc=Document().p("&amp; is used for &"))
<!DOCTYPE html><meta charset="utf-8"><title>Escaping &amp; Context</title>
<style>h1::after {content: "<\/Style>"}</style><h1>&lt;Escape></h1>
<!--All‒‒>OK-->
<iframe srcdoc="<!DOCTYPE html><p>&amp;amp;amp; is used for &amp;amp;"></iframe>

Works perfectly in browsers.

Name mangling and boolean attributes

Underscore at the end of name is ignored so that class_ and for_ among other attributes may be used despite being reserved words in Python. Other underscores convert into hyphens.

⚠️ The above only is true for HTML elements and attributes, but template placeholders only use an ending underscore to denote that the it is to be placed on the document, rather than be fetched for use.

Boolean values convert into short attributes.

E.input(type="checkbox", id="somebox", checked=True).label(for_="somebox", aria_role="img")("🥳")
<input type=checkbox id=somebox checked><label for=somebox aria-role=img>🥳</label>

Preformatted HTML

All content is automatically escaped, unless it provides an __html__ method that returns a string in HTML format. Similarly, the builder objects of this module expose __html__ and _repr_html_ accessors that allow them to be rendered as HTML in Jupyter Notebooks and various other systems that follow this convention.

Any preformatted HTML may be wrapped in html5tagger.HTML(string_of_html) to avoid it being escaped when included in a document, as the HTML class has those accessors.

⚠️ Do not use HTML() for text, in particular not on messages sent by users, that may contain HTML that you didn't intend to execute as HTML.

Performance

%timeit str(Document("benchmarking", lang="en", _urls=("foo.js", "bar.js")))
14 µs ± 153 ns per loop (mean ± std. dev. of 7 runs, 100,000 loops each)

Jinja2 renders similar document from memory template within about 10 µs but it doesn't need to format any of the HTML. When Templating is similarly used with html5tagger, the rendering times drop to about 4 µs.

In the above benchmark html5tagger created the entire document from scratch, one element and attribute at a time. Unless you are creating very large documents dynamically, this should be quite fast enough.

Further development

There have been no changes to the tagging API since 2018 when this module was brought to production use, and thus the interface is considered stable.

In 2023 support for templating was added, allowing documents to be preformatted for all their static parts (as long strings), with only templates filled in between. This is a work on progress and has not been optimized yet.

Additionally, _script and _style special methods were added in 2023. These may eventually replace also the non-underscored automatic versions but for now a separate method was easier to implement.

Pull requests are still welcome.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

html5tagger-1.2.1.tar.gz (13.3 kB view hashes)

Uploaded source

Built Distribution

html5tagger-1.2.1-py3-none-any.whl (9.9 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page