Sign inDemoInstall


Package Overview
File Explorer

Install Socket

Detect and block malicious and high-risk dependencies



A reliable markup language with the essentials for HTML typography. Fast to type and memorable. Includes optional UML, for unicode shortcuts.




:warning: new version of TML has changed API in some places.

Note: Due to a name clash This code is in Python repositories named as "django-ttml" ("ttml" is not a mistype).

A markup language with the essentials for HTML typography. Fast to type and memorable. Includes optional UML, for unicode shortcuts.

This version of TML provides some error reporting.

This app is called 'django-tml', but internally the module is called 'tml'. It has no connection with, and different intentions to, the pypi package called [tml]{, and it's Github repository. But this code has been called TML for years, in several computer languages, and so the rename.

Why you may not want this app

Where people don't use Markdown, they use reStructuredText. This is the only current implementation of TML. And this version is deliberately a web/HTML tool. TML is a very flexible language, but this app generates basic HTML. It is not intended to generate academic-level citations, micro-formating etc.

Why you may be interested

Like me, you feel markup is essential, interesting, but don't want to spend time on it.


The main code in the package is a parser. There are also some small pieces of support code. In 'tml/io' is a snippet to feed and present the parser in a suitable way for all kinds of code, not only Django. And a few template tags.

There is a basic attitude for use in Django which is, if users send markup, code should save markup. So there is no code for a transforming 'TML field'. If this bothers you, run a lazy conversion somewhere and cache. There are many ways to do that, and the web itself does that too.

If you have done this before

  • install
  • pick or create a TextField for TML storage.
  • Add the template tag to the rendering of the textfield



pip install django-ttml

or download the app code to Django.

If you do not need the Django integration, you do not need to install. Otherwise, declare in Django settings,


No migrations needed.

Template tags

Most of the time, this is all you need. Assuming a model called Page, with a TextField called 'text' holding the main body of text, in a template,

 {% load tml %}
 {{ page.text|tml }}

Done. Any TML code posted into the field becomes converted to HTML on rendering.

Other template tags

This will run the unicode converters,

 {{ page.text|uml }}

This runs the HTML generation of TML, followed by the Unicode conversion of UML,

 {{ page.text|tml_uml }}

See UML/TML integration for a description.

The next tag runs the TML and UML converters, and converts 'pre' areas into codeblocks configured for the code highlighter prismjs,

 {{ page.text|tml_uml_prism }}

Yes, you can create a parser for your choice of code-highlighter in a few lines. You need a Parser override to generate the HTML. Look at the code in 'tml/'. Then you need a template tag. Look at the code in 'tml/templatetags/'.

View-context conversion

If you are looking at TML semi-permanently, you could convert in a View context. Seems rather involved, but will spare you a template tag. The code to use will depend on the source data and the target template, but something like,

import tml

def get_context_data(self, **kwargs):
    ctx = super().get_context_data(**kwargs)
    obj = kwargs['object']
    ctx['content'] = tml.html(object.body)
    return ctx

This, at least, demonstrates how to convert in open code.

Details of operation


Django is Python, everything can be overridden. But I will not try to sell or fool, the structure of Django makes TML plugin tricky to override.

You need to make a custom parser. Often, this would only be a small override, For example, to respond to PRE marks with special HTML layout. The parser is structured to make these kinds of changes easy (it uses structured calls to render).

Then, you will need to make a new template tag. After all, those template tag calls are the easy, dramtic effect of this module.

In summary, overrides are awkward, and will cost you an afternoon to understand what is happening, then where and how to trigger the effect you want. But to get what you want may, in the end, only need five lines of code (if anyoine is interested in making this easier, then talk to me).


The code in 'tml/' is worth mentioning. It shows how to handle the parser.

The Parser is a feed parser. It does not take a 'source', it accepts a feed of lines, and needs a builder for output. The lines must be in a particular form, stripped of trailing newlines. This can be done using a io.StringIO object, which can normalise the line-ends, then feed the source lines as chunks.


Current attitude of the support code is to trust input, but escape 'pre' formatted areas for code display. I can see cases where you may wish for more or less escaping. We shall see.

Form support

The widget

A form widget for TextFields that contains a simple button-click-for-help. Please note this widget will not transform TML into HTML.

We try

It only adds a help button.

You can do a quick override in an admin field,

from django.db import models
from tml import TMLTextarea
class FireworkAdmin(admin.ModelAdmin):
    fields = ('title', 'slug', 'body', 'author')
    formfield_overrides = {
        models.TextField: {'widget': TMLTextarea},

NB: if used in admin, the button is positioned to the right. But in other forms it can be positioned elsewhere.

TML Intro

The API in Version 1

The main changes are:

The image mark ('*') has been expanded to cover the idea of self-closing elements. The new mark is parsed as a block opeining mark, which needs a tagname. The shortcut of this mark is 'img'. So 'img' now needs two asterisks, not one,

**(/images/tux.jpg)"The little guy"


<img src="/images/tux.jpg"/>
<figcaption>The little guy</figcaption>

Less elegant, but means if you want to add self-closing HTML elements, you can. For example this version of TML can now generate a horizontal rule,

I've finished what I was saying, am out of breath, so,


Ok, I'm back. 

or other self-closing elements like BR and WBR.

There's an addition. '#' in the aattribute list will generate an 'id' attribute in the HTML. In the same way '.' will generate a 'class' attribute. If both are present, 'id' must come before 'class',

I am me!


<div id="iamwhatiam" class="simplelifestyle">I am me!</div>

The special inline shortcut for 'span' is dropped. Inline attribute handling is now like block opeing marks. Needs a tag or shortcut. So, if you need a 'span' shortcut, double up,

I give warning {{.warning can be risky}


I give warning <span class="warning">can be risky</span>

But inlines marks are usually used with custom tags,

{cite The Encyclopedia}


<cite>The Encyclopedia</cite>

Less elegant, but more consistent with block opening marks (the 'span' shortcut seems to be rarely, if ever used).

Note that Version 1 does not refer to app testing. It's version 1 because there was a major API change. Semantic naming.


It would take pages of specification to cover the ins and outs of TML. But here is a short into to this reduced version.

Unlike several text-markup languages, TML does not aim to be 'invisible' to readers. It aims to be robust. Two results of robustness---within it's defined scope of typeography (ok, is that a cheat?) TML never needs to be escaped, and can give useful error messages. Sub-aims are to be fast to type (you only ever use top keyboard chars for shortcut cases), and memorable (I hate wasting time relearning markup).

The output from TML parsers usually paragraphs everything, automatically.

Block control codes are placed hard left in text. These are followed by a tagname, then closed by a repetition of the control code. These 'block' codes correspond to HTML 'structural' tags,

This is inside the div

There are shortcuts made by repeating the control code,

This is inside a div

There is also an inline tag. This can also be followed by a tagname, The tagname must be followed by a space or spaces to the enclosed text. This 'inline' control corresponds to HTML 'phrase content' tags,

So {cite this is inside} a citation tag.

There is a shortcut made by doubling the opening curly bracket,

So {{ this is inside a span} followed by other text

The tagname can be followed with a hash to start an id, then a full-stop/period to start a classname,

This is inside a div


<div id="anIdName" class="aWarningClassName"><p>Lorem ipsum dolor</p></div>

That is more than enough explanation in a README devoted to a module. I give up. Try pasting this into a TML field,


= About TML
== h2 headline
=== h3 headline
==== etc.
Newlines are reduced to spaces, like in TeX. 
That's not all.

Like TeX, two newlines make a paragrah break.


- List item
- List item 2

~ term 1
: definition 1
~ term 2
: definition 2

If you only need one line list elements, lists can be anonymous
- item 1
- item 2
- item 3

And finally, a couple of useful tricks {a( here is an anchor}. 

If you use the Prism parser (or the template tag ''tml_uml_prism'), the ''pre' block is tricked into being a codeblock, and the tagname is tricked into being the language specifier. Like this,

    c = a + b

To make code rendering in ''pre' blocks highlight you neet to link the webpage to the [a( prismjs] code.

Self-closing tags, and images

There is a special, self-closing tag mark at block level. It's an asterisk. It's shortcut implementation is an 'img' tag,

**.wide(/images/tux.jpg)"The little guy"

Renders unusual HTML, because this is what I use for block/structural-placed images. If you want a more conventional render, make or ask me for an override. Currently, this mark,

**#tux-image.wide(/images/tux.jpg)"The little guy"


<figure class="wide"">
    <img id="tux-image"  src="/images/tux.jpg" alt="image of tux">
    <figcaption>The little guy</figcaption>

Yes, this is a special render. The image is wrapped in a FIGURE element. The FIGURE element takes any marked classname. The IMG takes any marked idname. Optionally, the text attribute is used to generate a caption. And the HTML 'alt' attribute is auto-generated from the filename.

Tables, a quick word

Page layout aside (in which TML has no interest) tables are not used often. So there is no TML implementation. If you do need tables, this version of TML can invert the Django template policy of ''everything is escaped' and rage onwards with raw HTML,

<table class=ruled&#32;wide>

However, you can utilise the ability of the TML mark forms to define HTML element names,

{th manufacturer}{th model}
{th purchase-date}{th type}{th cordless}{th washable}
{td MicroShave}{td XFP456SC}
{td  2/9/2021}{td foil}{td y}{td y}

This may not be a dedicated solution, but I find it easier to handle than raw HTML. And consistent.

UML Intro

UML is a small set of conversions from keyboard-accessible codepoint sequences into inaccessible-but-useful-or-common codepoints. For example, it can turn shortcut inverted commas into open/close curly inverted commas. It can render a copyright sign etc.

Some people don't believe in doing this, for reasons I can't argue with. On the other hand, 19 out of 20 computer users can't find a copyright symbol. That's my excuse.

Have a look at the UML file for the details (of a somewhat reduced implementation). Or paste the following sentence into a textield rendered by a UML template tag,

So ''this' will generate curly quotes, as ""this" will generate curly double quotes, ''---' makes an em-dash, ''(c)' makes copyright, and this ''110:mo' will display your temperature as running at 110 degrees.

UML/TML integration

UML code stands alone. However, you may notice that in the templatetags UML is integrated into the TML parser,

{% load tml %}

    {{ article.text|tml_uml }}

If you look at the parser code, there's a paremeter to enable UML. Is that a convenience? No. The TML parser integrates UML so UML is never applied to TML codes. This is useful. If UML is applied to text generally, it may do unintended conversions to URLs or TML codes. Running UML after TML (or any other text processing) is risky too, it may do unintended conversions to the generated HTML. But the integrated UML only converts TML-designated free text.


Did you know?

Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.


Related posts

SocketSocket SOC 2 Logo


  • Package Alerts
  • Integrations
  • Docs
  • Pricing
  • FAQ
  • Roadmap

Stay in touch

Get open source security insights delivered straight into your inbox.

  • Terms
  • Privacy
  • Security

Made with ⚡️ by Socket Inc