teddy.js

Teddy is an easy-to-read, HTML-based, mostly logic-less DOM templating engine with support for both server-side and client-side templating.

It uses new HTML-like <tags> for rudimentary templating logic and Teddy Roosevelt's facial hair for {variables}.

Why yet another templating engine?

Good question.

Here's why:

Other popular templating engines are too cryptic

Are you tired of all those unnecessarily cryptic templating systems out there that look like this?

<h1>{{header}}</h1>
{{#bug}}
{{/bug}}

{{#items}}
  {{#first}}
    <li><strong>{{name}}</strong></li>
  {{/first}}
  {{#link}}
    <li><a href="{{url}}">{{name}}</a></li>
  {{/link}}
{{/items}}

{{#empty}}
  <p>The list is empty.</p>
{{/empty}}

...Or this?

{{#each comments}}
<h2><a href="/posts/{{../permalink}}#{{id}}">{{title}}</a></h2>
<div>{{body}}</div>
{{/each}}

...Or this?

{#names}{.}{@idx}{.}{/idx}{@sep}, {/sep}{/names}

Want something simpler and more readable so you can stop wasting time memorizing what all those funky symbols do and just get work done?

Well you're not the only one.

Teddy, symbol-buster extraordinaire

Teddy the trust-buster was a man of the people, curtailing the abuse of monopolists. As the most badass President of all-time, there's no way he'd put up with all this indecipherable templating nonsense.

As such, Teddy the templating engine is an engine of the people, curtailing the abuse of indecipherable templating systems. Teddy trust-buster extraordinaire is now symbol-buster extraordinaire.

Here's how:

• More concise {variable} syntax. One curly bracket. Not two, not three, no pound signs, no question marks, no backticks, no gang signs, and no hieroglyphs.
• An <include> tag for layout templates and partials which accepts arguments via child <arg> elements.
• Flow control tags: <if>, <unless>, <elseif>, <elseunless>, and <else> for basic templating logic.
• A <foreach> tag for looping.
• Server-side {!comments!} delimited by exclamation points in a fashion similar to <!-- HTML comments -->. Server-side comments are stripped out at the template compilation stage.
• No funky symbols to memorize. Just {variables} for data and new HTML-like <tags> for rudimentary logic.

Documentation

Here's some examples of to write Teddy templates:

Includes

Include another template:

<include src='partial.html'></include>

Or use the no extension shorthand (Teddy will append the .html extension for you):

<include src='partial'></include>

Pass arguments to the template:

<include src='partial.html'>
  <arg firstArgument>Plain text argument</arg>
  <arg secondArgument>
    <span>Argument with HTML in it</span>
  </arg>
</include>

The arguments you've defined will be accessible as {firstArgument} and {secondArgument} in the child template partial.html.

Conditionals

Check for the presence of a variable:

<if something>
  <p>The variable 'something' is present</p>
</if>
<else>
  <p>The variable 'something' is not present</p>
</else>

Check a variable's value:

<if something='hello'>
  <p>The variable 'something' is set to 'hello'</p>
</if>
<else>
  <p>The variable 'something' is not set to 'hello'</p>
</else>

Check for the non-presence of a variable:

<unless something>
  <p>The variable 'something' is not present</p>
</unless>
<else>
  <p>The variable 'something' is present</p>
</else>

An <if> statement structure with an <elseif> tag which is evaluated if the first <if> fails:

<if something>
  <p>The variable 'something' is present</p>
</if>
<elseif somethingElse>
  <p>The variable 'something' is not present, but 'somethingElse' is present</p>
</elseif>
<else>
  <p>The variable 'something' is not present and neither is 'somethingElse'</p>
</else>

An <unless> statement structure with an <elseunless> tag which is evaluated if the first <unless> fails:

<unless something>
  <p>The variable 'something' is not present</p>
</unless>
<elseunless somethingElse>
  <p>The variable 'something' is present, but 'somethingElse' is not present</p>
</elseunless>
<else>
  <p>The variables 'something' and 'somethingElse' are both present</p>
</else>

One line ifs

If you need a more concise conditional just to control which attributes are applied to a given element, then use this syntax:

<p if-something true="class='something-is-present'" false="class='something-is-not-present'">One line if.</p>

In that structure, the attribute if-something checks to see if the variable something is present. If so, the class delcared in the true attribute is written to the element, resulting in the following output:

<p class='something-is-present'>One line if.</p>

If not, the class declared in the false attribute is written to the element, resulting in the following output:

<p class='something-is-not-present'>One line if.</p>

Like the <if> tag you can check for both the presence of a variable as well as its value. To check the value of a variable, use this syntax:

<p if-something='hello' true="class='something-is-hello'" false="class='something-is-not-hello'">One line if.</p>

It's important to note that whichever type of quotes you use on the outside of your true or false attributes must be reversed on the inside. So if you use single quotes on the outside, then you must use double quotes on the inside.

Loops

Assume the following JS model:

letters = ['a', 'b', 'c'];

It can be iterated over like so:

<foreach val='letter' in='letters'>
  <p>{letter}</p> <!-- outputs a, b, c -->
</foreach>

In the above example in='letters' defines the JS model being iterated over and val='letter' defines a local variable for the current letter being iterated over.

When looping over more complex data structures, sometimes you will need access to both the key and the value of your array or object. For instance, suppose this JS model:

names = {jack: 'guy', jill: 'girl', hill: 'landscape'};

It can be iterated over like so:

<foreach key='name' val='description' in='names'>
  <p>{name}</p> <!-- outputs jack, jill, hill -->
  <p>{description}</p> <!-- outputs guy, girl, landscape -->
</foreach>

We once again define an in attribute which we set to in='names' and a val attribute which we set to val='description' similar to the last example. However this time we've iterated over a JS object with named keys instead of a simple indexed array, so it is useful to define a key attribute in the <foreach> tag to gain access to the name of the current iteration variable. We have defined it as key='name' in this example.

Even complex, hierarchical data structures can be iterated over. For instance, suppose this JS model:

objects = [{a:1, b:2, c:3}, {a:4, b:5, c:6}, {a:7, b:8, c:9}];

For the above array of objects, we can combine the techniques illustrated above to display each member of the hierarchy in sequence:

<foreach key='i' val='item' in='objects'>
  <p>{i}</p> <!-- outputs 0, 1, 2 -->
  <p>{item.a}</p> <!-- outputs 1, 4, 7 -->
  <p>{item.b}</p> <!-- outputs 2, 5, 8 -->
  <p>{item.c}</p> <!-- outputs 3, 6, 9 -->
</foreach>

A complex example combining all tag types

Supposing the following JS model again:

objects = [{a:1, b:2, c:3}, {a:4, b:5, c:6}, {a:7, b:8, c:9}];

We could perform many complex operations simultaneously. For instance, we could iterate over it with a <foreach> and then at each iteration perform an <if> statement and <include> a partial:

<foreach val='item' in='objects'>
  <if item.a='4'>
    <p>item.a is 4</p>
  </if>
  <p if-item.b='5' true="class='item-b-is-five'" false='hidden'>item.b is 5</p>
  <include src='partial.html'>
    <arg firstArgument>{item.b}</arg>
    <arg secondArgument>
      <span>{item.c}</span>
    </arg>
  </include>
</foreach>

Using Teddy in Node.js

Teddy is designed for use with Express in Node.

• First require the node modules express, and teddy
• Then initialize express and configure it to your liking
• In your express config, make sure to include this line: app.engine('html', teddy.__express)

For a complete sample implementation, see the sample app here: sampleApps/nodeHelloWorld

Or if you're looking for a more fully-featured web framework to build web apps with using Teddy templates, then try out Teddy's companion, Roosevelt.

Using Teddy with client-side JS

Server-side app:

• Setup a Node.js app using the instructions above
• Precompile templates somewhere in your app using the teddy.compile('templateName.html') method
• Create a route that serves one or more precompiled templates as text/javascript
• Optionally write the precompiled templates somewhere to disk and serve them as statics instead to optimize for performance

Client-side HTML:

• Include teddy.js: <script src='teddy.js'></script>
• Include precompiled templates: <script src='path/to/precompiled/templates'></script>
• Include your client.js: <script src='client.js'></script>

Writing your client.js:

• Define a model: var model = {some: 'data'} or optionally pull the JSON from the server, localStorage, or wherever
• Render a template client-side: var rendered = teddy.render(teddy.compiledTemplates['template.html'], model)
• Render the template into the document somewhere, for instance: document.body.insertAdjacentHTML('beforeend', rendered);

For a complete sample implementation, see the sample app here: sampleApps/client-server

Notable intentional design choices and limitations

• All variables in Teddy templates are case-insensitive because HTML is case-insensitive.
• Unlike some other templating systems, Teddy will not automatically escape HTML entities contained within variables. You should filter such things if desired at the code level and pass the results to the templates through the model as readable variables.
• Teddy adheres to a mostly logic-less templates philosophy. <if> and <unless> statements in Teddy can only check for a single variable's presence or its precise value. They cannot evaluate complex logic; there are no ands, ors, xors, parentheticals, or math operators. Teddy is of the opinion that complex logic doesn't belong in templates. Evaluate such logic at the code level and pass the results to the templates through the model as readable variables.
• Teddy's client-side performance and browser support is largely tied to how well DOMParser and XMLSerializer are implemented (if at all) in the target browser.

Client-side browser support

• Android browser 4.0+
• Chrome 24+
• Firefox 18+
• Internet Explorer 10+
• iOS Safari 5+
• Opera 12+
• Safari 6+

Known issues

• Teddy is beta software. Not many apps have been written using it yet, so it's entirely possible that there will be some significant bugs.
• Source code view on client-side unit tests is broken in IE10 most likely due to a bug in dependency library vkbeautify.
• The unit tests are a bit primitive at the moment. Suggestions for improvement or pull requests with better tests will be much appreciated.

Notable untested browsers

• Android browser 2.2 and below.
• iOS Safari 4 and below.
• Opera Mini.
• BlackBerry browsers.
• Lots of older versions of the desktop versions of Chrome, Firefox, Opera, and Safari probably work well with Teddy, but there hasn't yet been time to test them. If anyone wants to test them and post the results, it will be greatly appreciated.

Browsers known to be partially broken

• Older Android 2.3 browsers and most likely all prior versions fail the <title> and <style> tag tests because doc.documentElement.innerHTML fails to return the <head> tag or its contents. However, if you only intend to use Teddy for partials containing markup fragments rather than full documents, then this bug shouldn't affect you.
• Older Presto-powered Opera Mobile versions fail the conditionals overall unit test due to an as yet uninvestigated bug, but all other tests pass. Such Opera Mobile versions also have overall terrible performance with Teddy, which may be related.

Browsers known to be totally broken

• IE9 is totally broken because IE9's implementation of DOMParser sucks and @eligrey's DOMParser HTML extension (polyfill) doesn't work in IE9 and below.
• Windows Phone 7's browser: same reason as IE9.
• IE8 and below are totally broken because they have no implementation whatsoever of DOMParser.

Helped wanted!

There are many ways to improve Teddy, such as:

• Testing some of the untested browsers.
• Helping to fix some of the things noted above.
• Porting Teddy's JS parser so Teddy templates can be used in other languages, like PHP, Python, Java, etc.
• Submitting pull requests with suggestions for refactoring or performance optimization.

How to run the unit tests

Want to hack teddy's code but don't want to break something and cause a regression in the process? Run the supplied unit tests to sanity check existing features.

Start by cloning the git repo:

git clone git@github.com:kethinov/teddy.git
cd teddy

Install dependencies for the server test:

npm install

Run server test:

npm test

To run the client tests, open unitTests/client/clientTests.html and follow its instructions.

Dependencies

Node.js dependencies:

• xmldom - W3C Standard based (XML DOM Level 2 Core) DOMParser and XMLSerializer for Node.js

Client-side dependencies:

• DOMParser HTML extension (bundled) - polyfill for DOMParser parseFromString for certain browsers

Node.js unit test dependencies:

• None

Client-side unit test dependencies:

• prettify.js (bundled) - used to syntax highlight rendered template in unitTests/client/clientTests.html
• vkBeautify (bundled) - used to indent rendered template in unitTests/client/clientTests.html

License

All original code in Teddy is licensed under the Creative Commons Attribution 4.0 International License. Commercial and noncommercial use is permitted with attribution.