web2py™ 

Search

Table of Contents

Chapter 5: 视图

The views

views
template language
HTML

web2py uses Python for its models, controllers, and views, although it uses a slightly modified Python syntax in the views to allow more readable code without imposing any restrictions on proper Python usage.

The purpose of a view is to embed code (Python) in an HTML document. In general, this poses some problems:

  • How should embedded code be escaped?
  • Should indenting be based on Python or HTML rules?

web2py uses {{ ... }} to escape Python code embedded in HTML. The advantage of using curly brackets instead of angle brackets is that it's transparent to all common HTML editors. This allows the developer to use those editors to create web2py views.

Since the developer is embedding Python code into HTML, the document should be indented according to HTML rules, and not Python rules. Therefore, we allow unindented Python inside the {{ ... }} tags. Since Python normally uses indentation to delimit blocks of code, we need a different way to delimit them; this is why the web2py template language makes use of the Python keyword pass.

A code block starts with a line ending with a colon and ends with a line beginning with pass. The keyword pass is not necessary when the end of the block is obvious from the context.

Here is an example:

{{
if i == 0:
response.write('i is 0')
else:
response.write('i is not 0')
pass
}}

Note that pass is a Python keyword, not a web2py keyword. Some Python editors, such as Emacs, use the keyword pass to signify the division of blocks and use it to re-indent code automatically.

The web2py template language does exactly the same. When it finds something like:

<html><body>
{{for x in range(10):}}{{=x}}hello<br />{{pass}}
</body></html>

it translates it into a program:

response.write("""<html><body>""", escape=False)
for x in range(10):
    response.write(x)
    response.write("""hello<br />""", escape=False)
response.write("""</body></html>""", escape=False)

response.write writes to the response.body.

When there is an error in a web2py view, the error report shows the generated view code, not the actual view as written by the developer. This helps the developer debug the code by highlighting the actual code that is executed (which is something that can be debugged with an HTML editor or the DOM inspector of the browser).

Also note that:

{{=x}}

generates

response.write
escape

response.write(x)

Variables injected into the HTML in this way are escaped by default. The escaping is ignored if x is an XML object, even if escape is set to True.

Here is an example that introduces the H1 helper:

{{=H1(i)}}

which is translated to:

response.write(H1(i))

upon evaluation, the H1 object and its components are recursively serialized, escaped and written to the response body. The tags generated by H1 and inner HTML are not escaped. This mechanism guarantees that all text --- and only text --- displayed on the web page is always escaped, thus preventing XSS vulnerabilities. At the same time, the code is simple and easy to debug.

The method response.write(obj, escape=True) takes two arguments, the object to be written and whether it has to be escaped (set to True by default). If obj has an .xml() method, it is called and the result written to the response body (the escape argument is ignored). Otherwise it uses the object's __str__ method to serialize it and, if the escape argument is True, escapes it. All built-in helper objects (H1 in the example) are objects that know how to serialize themselves via the .xml() method.

This is all done transparently. You never need to (and never should) call the response.write method explicitly.

Basic syntax

The web2py template language supports all Python control structures. Here we provide some examples of each of them. They can be nested according to usual programming practice.

for...in

for

In templates you can loop over any iterable object:

{{items = ['a', 'b', 'c']}}
<ul>
{{for item in items:}}<li>{{=item}}</li>{{pass}}
</ul>

which produces:

<ul>
<li>a</li>
<li>b</li>
<li>c</li>
</ul>

Here item is any iterable object such as a Python list, Python tuple, or Rows object, or any object that is implemented as an iterator. The elements displayed are first serialized and escaped.

while

while

You can create a loop using the while keyword:

{{k = 3}}
<ul>
{{while k > 0:}}<li>{{=k}}{{k = k - 1}}</li>{{pass}}
</ul>

which produces:

<ul>
<li>3</li>
<li>2</li>
<li>1</li>
</ul>

if...elif...else

if
elif
else

You can use conditional clauses:

{{
import random
k = random.randint(0, 100)
}}
<h2>
{{=k}}
{{if k % 2:}}is odd{{else:}}is even{{pass}}
</h2>

which produces:

<h2>
45 is odd
</h2>

Since it is obvious that else closes the first if block, there is no need for a pass statement, and using one would be incorrect. However, you must explicitly close the else block with a pass.

Recall that in Python "else if" is written elif as in the following example:

{{
import random
k = random.randint(0, 100)
}}
<h2>
{{=k}}
{{if k % 4 == 0:}}is divisible by 4
{{elif k % 2 == 0:}}is even
{{else:}}is odd
{{pass}}
</h2>

It produces:

<h2>
64 is divisible by 4
</h2>

try...except...else...finally

try
except
else
finally

It is also possible to use try...except statements in views with one caveat. Consider the following example:

{{try:}}
Hello {{= 1 / 0}}
{{except:}}
division by zero
{{else:}}
no division by zero
{{finally}}
<br />
{{pass}}

It will produce the following output:

Hello
division by zero
<br />

This example illustrates that all output generated before an exception occurs is rendered (including output that preceded the exception) inside the try block. "Hello" is written because it precedes the exception.

def...return

def
return

The web2py template language allows the developer to define and implement functions that can return any Python object or a text/html string. Here we consider two examples:

{{def itemize1(link): return LI(A(link, _href="http://" + link))}}
<ul>
{{=itemize1('www.google.com')}}
</ul>

produces the following output:

<ul>
<li><a href="http:/www.google.com">www.google.com</a></li>
</ul>

The function itemize1 returns a helper object that is inserted at the location where the function is called.

Consider now the following code:

{{def itemize2(link):}}
<li><a href="http://{{=link}}">{{=link}}</a></li>
{{return}}
<ul>
{{itemize2('www.google.com')}}
</ul>

It produces exactly the same output as above. In this case, the function itemize2 represents a piece of HTML that is going to replace the web2py tag where the function is called. Notice that there is no '=' in front of the call to itemize2, since the function does not return the text, but it writes it directly into the response.

There is one caveat: functions defined inside a view must terminate with a return statement, or the automatic indentation will fail.

HTML helpers

helpers

Consider the following code in a view:

{{=DIV('this', 'is', 'a', 'test', _id='123', _class='myclass')}}

it is rendered as:

<div id="123" class="myclass">thisisatest</div>

DIV is a helper class, i.e., something that can be used to build HTML programmatically. It corresponds to the HTML <div> tag.

Positional arguments are interpreted as objects contained between the open and close tags. Named arguments that start with an underscore are interpreted as HTML tag attributes (without the underscore). Some helpers also have named arguments that do not start with underscore; these arguments are tag-specific.

Instead of a set of unnamed arguments, a helper can also take a single list or tuple as its set of components using the * notation and it can take a single dictionary as its set of attributes using the **, for example:

{{
contents = ['this','is','a','test']
attributes = {'_id':'123', '_class':'myclass'}
=DIV(*contents,**attributes)
}}

(produces the same output as before).

The following set of helpers:

A, B, BEAUTIFY, BODY, BR, CAT, CENTER, CODE, COL, COLGROUP, DIV, EM, EMBED, FIELDSET, FORM, H1, H2, H3, H4, H5, H6, HEAD, HR, HTML, I, IFRAME, IMG, INPUT, LABEL, LEGEND, LI, LINK, MARKMIN, MENU, META, OBJECT, ON, OL, OPTGROUP, OPTION, P, PRE, SCRIPT, SELECT, SPAN, STYLE, TABLE, TAG, TBODY, TD, TEXTAREA, TFOOT, TH, THEAD, TITLE, TR, TT, UL, URL, XHTML, XML, embed64, xmlescape

can be used to build complex expressions that can then be serialized to XML[xml-w] [xml-o]. For example:

{{=DIV(B(I("hello ", "<world>"))), _class="myclass")}}

is rendered:

<div class="myclass"><b><i>hello &lt;world&gt;</i></b></div>

Helpers can also be serialized into strings, equivalently, with the __str__ and the xml methods:

>>> print str(DIV("hello world"))
<div>hello world</div>
>>> print DIV("hello world").xml()
<div>hello world</div>
Document Object Model (DOM)

The helpers mechanism in web2py is more than a system to generate HTML without concatenating strings. It provides a server-side representation of the Document Object Model (DOM).

Components of helpers can be referenced via their position, and helpers act as lists with respect to their components:

>>> a = DIV(SPAN('a', 'b'), 'c')
>>> print a
<div><span>ab</span>c</div>
>>> del a[1]
>>> a.append(B('x'))
>>> a[0][0] = 'y'
>>> print a
<div><span>yb</span><b>x</b></div>

Attributes of helpers can be referenced by name, and helpers act as dictionaries with respect to their attributes:

>>> a = DIV(SPAN('a', 'b'), 'c')
>>> a['_class'] = 's'
>>> a[0]['_class'] = 't'
>>> print a
<div class="s"><span class="t">ab</span>c</div>

Note, the complete set of components can be accessed via a list called a.components, and the complete set of attributes can be accessed via a dictionary called a.attributes. So, a[i] is equivalent to a.components[i] when i is an integer, and a[s] is equivalent to a.attributes[s] when s is a string.

Notice that helper attributes are passed as keyword arguments to the helper. In some cases, however, attribute names include special characters that are not allowed in Python identifiers (e.g., hyphens) and therefore cannot be used as keyword argument names. For example:

DIV('text', _data-role='collapsible')

will not work because "_data-role" includes a hyphen, which will produce a Python syntax error.

In such cases, you can instead pass the attributes as a dictionary and make use of Python's ** function arguments notation, which map a dictionary of (key:value) pairs into a set of keyword arguments:

>>> print DIV('text', **{'_data-role': 'collapsible'})
<div data-role="collapsible">text</div>

You can also dynamically create special TAGs:

>>> print TAG['soap:Body']('whatever',**{'_xmlns:m':'http://www.example.org'})
<soap:Body xmlns:m="http://www.example.org">whatever</soap:Body>

XML

XML

XML is an object used to encapsulate text that should not be escaped. The text may or may not contain valid XML. For example, it could contain JavaScript.

The text in this example is escaped:

>>> print DIV("<b>hello</b>")
&lt;b&gt;hello&lt;/b&gt;

by using XML you can prevent escaping:

>>> print DIV(XML("<b>hello</b>"))
<b>hello</b>

Sometimes you want to render HTML stored in a variable, but the HTML may contain unsafe tags such as scripts:

>>> print XML('<script>alert("unsafe!")</script>')
<script>alert("unsafe!")</script>

Un-escaped executable input such as this (for example, entered in the body of a comment in a blog) is unsafe, because it can be used to generate Cross Site Scripting (XSS) attacks against other visitors to the page.

sanitize

The web2py XML helper can sanitize our text to prevent injections and escape all tags except those that you explicitly allow. Here is an example:

>>> print XML('<script>alert("unsafe!")</script>', sanitize=True)
&lt;script&gt;alert(&quot;unsafe!&quot;)&lt;/script&gt;

The XML constructors, by default, consider the content of some tags and some of their attributes safe. You can override the defaults using the optional permitted_tags and allowed_attributes arguments. Here are the default values of the optional arguments of the XML helper.

XML(text, sanitize=False,
    permitted_tags=['a', 'b', 'blockquote', 'br/', 'i', 'li',
       'ol', 'ul', 'p', 'cite', 'code', 'pre', 'img/'],
    allowed_attributes={'a':['href', 'title'],
       'img':['src', 'alt'], 'blockquote':['type']})

Built-in helpers

A

This helper is used to build links.

A
>>> print A('<click>', XML('<b>me</b>'),
            _href='http://www.web2py.com')
<a href='http://www.web2py.com'>&lt;click&gt;<b>me/b></a>

Instead of _href you can pass the URL using the callback argument. For example in a view:

{{=A('click me', callback=URL('myaction'))}}

and the effect of pressing the link will be an ajax call to "myaction" instead of a redirection. In this case, optionally you can specify two more arguments: target and delete:

{{=A('click me', callback=URL('myaction'), target="t")}}
<div id="t"><div>

and the response of the ajax callback will be stored in the DIV with id equal to "t".

<div id="b">{{=A('click me', callback=URL('myaction'), delete='div#b")}}</div>

and upon response, the closest tag matching "div#b" will be deleted. In this case, the button will be deleted. A typical application is:

{{=A('click me', callback=URL('myaction'), delete='tr")}}

in a table. Pressing the button will perform the callback and delete the table row.

callback and delete can be combined.

Tha A helper takes a special argument called cid. It works as follows:

{{=A('linked page', _href='http://example.com', cid='myid')}}
<div id="myid"></div>

and a click on the link causes the content to be loaded in the div. This is similar but more powerful than the above syntax since it is designed to refresh page components. We discuss applications of cid in more detail in Chapter 12, in the context of components.

These ajax features require jQuery and "static/js/web2py_ajax.js", which are automatically included by placing {{include 'web2py_ajax.html'}} in the layout head. "views/web2py_ajax.html" defines some variables based on request and includes all necessary js and css files.

B
B

This helper makes its contents bold.

>>> print B('<hello>', XML('<i>world</i>'), _class='test', _id=0)
<b id="0" class="test">&lt;hello&gt;<i>world</i></b>
BODY
BODY

This helper makes the body of a page.

>>> print BODY('<hello>', XML('<b>world</b>'), _bgcolor='red')
<body bgcolor="red">&lt;hello&gt;<b>world</b></body>
BR
BR

This helper creates a line break.

>>> print BR()
<br />
CAT (1.98.1 and up)
CAT

This helper concatenates other helpers, same as TAG[].

>>> print CAT('Here is a ', A('link',_href=URL()), ', and here is some ', B('bold text'), '.')
Here is a <a href="/app/default/index">link</a>, and here is some <b>bold text</b>.
CENTER
CENTER

This helper centers its content.

>>> print CENTER('<hello>', XML('<b>world</b>'),
>>>              _class='test', _id=0)
<center id="0" class="test">&lt;hello&gt;<b>world</b></center>
CODE
CODE

This helper performs syntax highlighting for Python, C, C++, HTML and web2py code, and is preferable to PRE for code listings. CODE also has the ability to create links to the web2py API documentation.

Here is an example of highlighting sections of Python code.

>>> print CODE('print "hello"', language='python').xml()
<table><tr valign="top"><td style="width:40px; text-align: right;"><pre style="
        font-size: 11px;
        font-family: Bitstream Vera Sans Mono,monospace;
        background-color: transparent;
            margin: 0;
            padding: 5px;
            border: none;
        background-color: #E0E0E0;
        color: #A0A0A0;
    ">1.</pre></td><td><pre style="
        font-size: 11px;
        font-family: Bitstream Vera Sans Mono,monospace;
        background-color: transparent;
            margin: 0;
            padding: 5px;
            border: none;
            overflow: auto;
    "><span style="color:#185369; font-weight: bold">print </span>
    <span style="color: #FF9966">"hello"</span></pre></td></tr>
</table>

Here is a similar example for HTML

>>> print CODE(
>>>   '<html><body>{{=request.env.remote_add}}</body></html>',
>>>   language='html')

<table>...<code>...
<html><body>{{=request.env.remote_add}}</body></html>
...</code>...</table>

These are the default arguments for the CODE helper:

CODE("print 'hello world'", language='python', link=None, counter=1, styles={})

Supported values for the language argument are "python", "html_plain", "c", "cpp", "web2py", and "html". The "html" language interprets {{ and }} tags as "web2py" code, while "html_plain" doesn't.

If a link value is specified, for example "/examples/global/vars/", web2py API references in the code are linked to documentation at the link URL. For example "request" would be linked to "/examples/global/vars/request". In the above example, the link URL is handled by the "vars" action in the "global.py" controller that is distributed as part of the web2py "examples" application.

The counter argument is used for line numbering. It can be set to any of three different values. It can be None for no line numbers, a numerical value specifying the start number, or a string. If the counter is set to a string, it is interpreted as a prompt, and there are no line numbers.

The styles argument is a bit tricky. If you look at the generated HTML above, it contains a table with two columns, and each column has its own style declared inline using CSS. The styles attributes allows you to override those two CSS styles. For example:

{{=CODE(...,styles={'CODE':'margin: 0;padding: 5px;border: none;'})}}

The styles attribute must be a dictionary, and it allows two possible keys: CODE for the style of the actual code, and LINENUMBERS for the style of the left column, which contains the line numbers. Mind that these styles completely replace the default styles and are not simply added to them.

COL
COL
>>> print COL('a','b')
<col>ab</col>
COLGROUP
COLGROUP
>>> print COLGROUP('a','b')
<colgroup>ab</colgroup>
DIV

All helpers apart from XML are derived from DIV and inherit its basic methods.

DIV
>>> print DIV('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<div id="0" class="test">&lt;hello&gt;<b>world</b></div>
EM

Emphasizes its content.

EM
>>> print EM('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<em id="0" class="test">&lt;hello&gt;<b>world</b></em>
FIELDSET
FIELDSET

This is used to create an input field together with its label.

>>> print FIELDSET('Height:', INPUT(_name='height'), _class='test')
<fieldset class="test">Height:<input name="height" /></fieldset>
FORM
FORM

This is one of the most important helpers. In its simple form, it just makes a <form>...</form> tag, but because helpers are objects and have knowledge of what they contain, they can process submitted forms (for example, perform validation of the fields). This will be discussed in detail in Chapter 7.

>>> print FORM(INPUT(_type='submit'), _action='', _method='post')
<form enctype="multipart/form-data" action="" method="post">
<input type="submit" /></form>

The "enctype" is "multipart/form-data" by default.

hidden

The constructor of a FORM, and of SQLFORM, can also take a special argument called hidden. When a dictionary is passed as hidden, its items are translated into "hidden" INPUT fields. For example:

>>> print FORM(hidden=dict(a='b'))
<form enctype="multipart/form-data" action="" method="post">
<input value="b" type="hidden" name="a" /></form>
H1, H2, H3, H4, H5, H6
H1

These helpers are for paragraph headings and subheadings:

>>> print H1('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<h1 id="0" class="test">&lt;hello&gt;<b>world</b></h1>
HEAD

For tagging the HEAD of an HTML page.

HEAD
>>> print HEAD(TITLE('<hello>', XML('<b>world</b>')))
<head><title>&lt;hello&gt;<b>world</b></title></head>
HTML

HTML
XHTML

This helper is a little different. In addition to making the <html> tags, it prepends the tag with a doctype string[xhtml-w,xhtml-o,xhtml-school] .

>>> print HTML(BODY('<hello>', XML('<b>world</b>')))
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
                      "http://www.w3.org/TR/html4/loose.dtd">
<html><body>&lt;hello&gt;<b>world</b></body></html>

The HTML helper also takes some additional optional arguments that have the following default:

HTML(..., lang='en', doctype='transitional')

where doctype can be 'strict', 'transitional', 'frameset', 'html5', or a full doctype string.

XHTML
XHTML

XHTML is similar to HTML but it creates an XHTML doctype instead.

XHTML(..., lang='en', doctype='transitional', xmlns='http://www.w3.org/1999/xhtml')

where doctype can be 'strict', 'transitional', 'frameset', or a full doctype string.

HR
HR

This helper creates a horizontal line in an HTML page

>>> print HR()
<hr />
I
I

This helper makes its contents italic.

>>> print I('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<i id="0" class="test">&lt;hello&gt;<b>world</b></i>
INPUT
INPUT

Creates an <input.../> tag. An input tag may not contain other tags, and is closed by /> instead of >. The input tag has an optional attribute _type that can be set to "text" (the default), "submit", "checkbox", or "radio".

>>> print INPUT(_name='test', _value='a')
<input value="a" name="test" />

It also takes an optional special argument called "value", distinct from "_value". The latter sets the default value for the input field; the former sets its current value. For an input of type "text", the former overrides the latter:

>>> print INPUT(_name='test', _value='a', value='b')
<input value="b" name="test" />

For radio buttons, INPUT selectively sets the "checked" attribute:

radio
>>> for v in ['a', 'b', 'c']:
>>>     print INPUT(_type='radio', _name='test', _value=v, value='b'), v
<input value="a" type="radio" name="test" /> a
<input value="b" type="radio" checked="checked" name="test" /> b
<input value="c" type="radio" name="test" /> c

and similarly for checkboxes:

checkbox
>>> print INPUT(_type='checkbox', _name='test', _value='a', value=True)
<input value="a" type="checkbox" checked="checked" name="test" />
>>> print INPUT(_type='checkbox', _name='test', _value='a', value=False)
<input value="a" type="checkbox" name="test" />
IFRAME

This helper includes another web page in the current page. The url of the other page is specified via the "_src" attribute.

IFRAME
>>> print IFRAME(_src='http://www.web2py.com')
<iframe src="http://www.web2py.com"></iframe>
IMG
IMG

It can be used to embed images into HTML:

>>> IMG(_src='http://example.com/image.png',_alt='test')
<img src="http://example.com/image.ong" alt="rest" />

Here is a combination of A, IMG, and URL helpers for including a static image with a link:

>>> A(IMG(_src=URL('static','logo.png'), _alt="My Logo"),
      _href=URL('default','index'))
<a href="/myapp/default/index">
  <img src="/myapp/static/logo.png" alt="My Logo" />
</a>
LABEL

It is used to create a LABEL tag for an INPUT field.

LABEL
>>> print LABEL('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<label id="0" class="test">&lt;hello&gt;<b>world</b></label>
LEGEND

It is used to create a legend tag for a field in a form.

LEGEND
>>> print LEGEND('Name', _for='myfield')
<legend for="myfield">Name</legend>
LI

It makes a list item and should be contained in a UL or OL tag.

LI
>>> print LI('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<li id="0" class="test">&lt;hello&gt;<b>world</b></li>
META

To be used for building META tags in the HTML head. For example:

META
>>> print META(_name='security', _content='high')
<meta name="security" content="high" />
MARKMIN

Implements the markmin wiki syntax. It converts the input text into output html according to the markmin rules described in the example below:

MARKMIN
>>> print MARKMIN("this is **bold** or ''italic'' and this [[a link http://web2py.com]]")
<p>this is <b>bold</b> or <i>italic</i> and
this <a href="http://web2py.com">a link</a></p>

The markmin syntax is described in this file that ships with web2py:

http://127.0.0.1:8000/examples/static/markmin.html

and some examples are shown in chapter 12 in the context of plugin_wiki, which uses MARKMIN extensively.

You can use markmin to generate HTML, LaTeX and PDF documents:

m = "Hello **world** [[link http://web2py.com]]"
from gluon.contrib.markmin.markmin2html import markmin2html
print markmin2html(m)
from gluon.contrib.markmin.markmin2latex import markmin2latex
print markmin2latex(m)
from gluon.contrib.markmin.markmin2pdf import markmin2pdf
print markmin2pdf(m) # requires pdflatex

(the MARKMIN helper is a shortcut for markmin2html)

Here is a basic syntax primer:

SOURCEOUTPUT
# titletitle
## sectionsection
### subsectionsubsection
**bold**bold
''italic''italic
``verbatim``verbatim
http://google.comhttp://google.com
http://...<a href="http://...">http:...</a>
http://...png<img src="http://...png" />
http://...mp3<audio src="http://...mp3"></audio>
http://...mp4<video src="http://...mp4"></video>
qr:http://...<a href="http://..."><img src="qr code"/></a>
embed:http://...<iframe src="http://..."></iframe>
[[click me #myanchor]]click me
$$\int_a^b sin(x)dx$$

Simply including a link to an image, a videos or an audio files without markup result in the corresponding image, video or audio file beling included automatically (for audio and video it uses html <audio> and <video> tags).

Adding a link with the qr: prefix such as

qr:http://web2py.com

results in the corresponding QR code being embedded and linking the said URL.

Adding a link tith the embed: prefix such as

embed:http://www.youtube.com/embed/x1w8hKTJ2Co

results in the page being embedded, in this case a youtube video is embedded.

Images can also be embedded with the following syntax:

[[image-description http://.../image.png right 200px]]

Unordered lists with:

- one
- two
- three

Ordered lists with:

+ one
+ two
+ three

and tables with:

----------
 X | 0 | 0
 0 | X | 0
 0 | 0 | 1
----------

The MARKMIN syntax also supports blockquotes, HTML5 audio and video tags, image alignment, custom css, and it can be extended:

MARKMIN("``abab``:custom", extra=dict(custom=lambda text: text.replace('a','c'))

generates

'cbcb'

Custom blocks are delimited by ``...``:<key> and they are rendered by the function passed as value for the corresponding key in the extra dictionary argument of MARKMIN. Mind that the function may need to escape the output to prevent XSS.

OBJECT

Used to embed objects (for example, a flash player) in the HTML.

OBJECT
>>> print OBJECT('<hello>', XML('<b>world</b>'),
>>>              _src='http://www.web2py.com')
<object src="http://www.web2py.com">&lt;hello&gt;<b>world</b></object>
OL

It stands for Ordered List. The list should contain LI tags. OL arguments that are not LI objects are automatically enclosed in <li>...</li> tags.

OL
>>> print OL('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<ol id="0" class="test"><li>&lt;hello&gt;</li><li><b>world</b></li></ol>
ON

This is here for backward compatibility and it is simply an alias for True. It is used exclusively for checkboxes and deprecated since True is more Pythonic.

ON
>>> print INPUT(_type='checkbox', _name='test', _checked=ON)
<input checked="checked" type="checkbox" name="test" />
OPTGROUP

Allows you to group multiple options in a SELECT and it is useful to customize the fields using CSS.

OPTGROUP
>>> print SELECT('a', OPTGROUP('b', 'c'))
<select>
  <option value="a">a</option>
  <optgroup>
    <option value="b">b</option>
    <option value="c">c</option>
  </optgroup>
</select>
OPTION

This should only be used as part of a SELECT/OPTION combination.

OPTION
>>> print OPTION('<hello>', XML('<b>world</b>'), _value='a')
<option value="a">&lt;hello&gt;<b>world</b></option>

As in the case of INPUT, web2py make a distinction between "_value" (the value of the OPTION), and "value" (the current value of the enclosing select). If they are equal, the option is "selected".

selected
>>> print SELECT('a', 'b', value='b'):
<select>
<option value="a">a</option>
<option value="b" selected="selected">b</option>
</select>
P
P

This is for tagging a paragraph.

>>> print P('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<p id="0" class="test">&lt;hello&gt;<b>world</b></p>
PRE
PRE

Generates a <pre>...</pre> tag for displaying pre-formatted text. The CODE helper is generally preferable for code listings.

>>> print PRE('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<pre id="0" class="test">&lt;hello&gt;<b>world</b></pre>
SCRIPT
SCRIPT

This is include or link a script, such as JavaScript. The content between the tags is rendered as an HTML comment, for the benefit of really old browsers.

>>> print SCRIPT('alert("hello world");', _type='text/javascript')
<script type="text/javascript"><!--
alert("hello world");
//--></script>
SELECT
SELECT

Makes a <select>...</select> tag. This is used with the OPTION helper. Those SELECT arguments that are not OPTION objects are automatically converted to options.

>>> print SELECT('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<select id="0" class="test">
   <option value="&lt;hello&gt;">&lt;hello&gt;</option>
   <option value="&lt;b&gt;world&lt;/b&gt;"><b>world</b></option>
</select>
SPAN
SPAN

Similar to DIV but used to tag inline (rather than block) content.

>>> print SPAN('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<span id="0" class="test">&lt;hello&gt;<b>world</b></span>
STYLE
STYLE

Similar to script, but used to either include or link CSS code. Here the CSS is included:

>>> print STYLE(XML('body {color: white}'))
<style><!--
body { color: white }
//--></style>

and here it is linked:

>>> print STYLE(_src='style.css')
<style src="style.css"><!--
//--></style>
TABLE, TR, TD

TABLE
TR
TD

These tags (along with the optional THEAD, TBODY and TFOOTER helpers) are used to build HTML tables.

>>> print TABLE(TR(TD('a'), TD('b')), TR(TD('c'), TD('d')))
<table><tr><td>a</td><td>b</td></tr><tr><td>c</td><td>d</td></tr></table>

TR expects TD content; arguments that are not TD objects are converted automatically.

>>> print TABLE(TR('a', 'b'), TR('c', 'd'))
<table><tr><td>a</td><td>b</td></tr><tr><td>c</td><td>d</td></tr></table>

It is easy to convert a Python array into an HTML table using Python's * function arguments notation, which maps list elements to positional function arguments.

Here, we will do it line by line:

>>> table = [['a', 'b'], ['c', 'd']]
>>> print TABLE(TR(*table[0]), TR(*table[1]))
<table><tr><td>a</td><td>b</td></tr><tr><td>c</td><td>d</td></tr></table>

Here we do all lines at once:

>>> table = [['a', 'b'], ['c', 'd']]
>>> print TABLE(*[TR(*rows) for rows in table])
<table><tr><td>a</td><td>b</td></tr><tr><td>c</td><td>d</td></tr></table>
TBODY
TBODY

This is used to tag rows contained in the table body, as opposed to header or footer rows. It is optional.

>>> print TBODY(TR('<hello>'), _class='test', _id=0)
<tbody id="0" class="test"><tr><td>&lt;hello&gt;</td></tr></tbody>
TEXTAREA
TEXTAREA

This helper makes a <textarea>...</textarea> tag.

>>> print TEXTAREA('<hello>', XML('<b>world</b>'), _class='test')
<textarea class="test" cols="40" rows="10">&lt;hello&gt;<b>world</b></textarea>

The only caveat is that its optional "value" overrides its content (inner HTML)

>>> print TEXTAREA(value="<hello world>", _class="test")
<textarea class="test" cols="40" rows="10">&lt;hello world&gt;</textarea>
TFOOT
TFOOT

This is used to tag table footer rows.

>>> print TFOOT(TR(TD('<hello>')), _class='test', _id=0)
<tfoot id="0" class="test"><tr><td>&lt;hello&gt;</td></tr></tfoot>
TH
TH

This is used instead of TD in table headers.

>>> print TH('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<th id="0" class="test">&lt;hello&gt;<b>world</b></th>
THEAD
THEAD

This is used to tag table header rows.

>>> print THEAD(TR(TH('<hello>')), _class='test', _id=0)
<thead id="0" class="test"><tr><th>&lt;hello&gt;</th></tr></thead>
TITLE
TITLE

This is used to tag the title of a page in an HTML header.

>>> print TITLE('<hello>', XML('<b>world</b>'))
<title>&lt;hello&gt;<b>world</b></title>
TR
TR

Tags a table row. It should be rendered inside a table and contain <td>...</td> tags. TR arguments that are not TD objects will be automatically converted.

>>> print TR('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<tr id="0" class="test"><td>&lt;hello&gt;</td><td><b>world</b></td></tr>
TT
TT

Tags text as typewriter (monospaced) text.

>>> print TT('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<tt id="0" class="test">&lt;hello&gt;<b>world</b></tt>
UL

Signifies an Unordered List and should contain LI items. If its content is not tagged as LI, UL does it automatically.

UL
>>> print UL('<hello>', XML('<b>world</b>'), _class='test', _id=0)
<ul id="0" class="test"><li>&lt;hello&gt;</li><li><b>world</b></li></ul>
embed64

embed64(filename=None, file=None, data=None, extension='image/gif') encodes the provided (binary) data into base64.

filename: if provided, opens and reads this file in 'rb' mode. file: if provided, reads this file. data: if provided, uses the provided data.

embed64
xmlescape

xmlescape(data, quote=True) returns an escaped string of the provided data.

xmlescape
>>> print xmlescape('<hello>')
&lt;hello&gt;

Custom helpers

TAG
TAG

Sometimes you need to generate custom XML tags. web2py provides TAG, a universal tag generator.

{{=TAG.name('a', 'b', _c='d')}}

generates the following XML

<name c="d">ab</name>

Arguments "a", "b", and "d" are automatically escaped; use the XML helper to suppress this behavior. Using TAG you can generate HTML/XML tags not already provided by the API. TAGs can be nested, and are serialized with str(). An equivalent syntax is:

{{=TAG['name']('a', 'b', c='d')}}

If the TAG object is created with an empty name, it can be used to concatenate multiple strings and HTML helpers together without inserting them into a surrounding tag, but this use is deprecated. Use the CAT helper instead.

Notice that TAG is an object, and TAG.name or TAG['name'] is a function that returns a temporary helper class.

MENU
MENU

The MENU helper takes a list of lists or of tuples of the form of response.menu (as described in Chapter 4) and generates a tree-like structure using unordered lists representing the menu. For example:

>>> print MENU([['One', False, 'link1'], ['Two', False, 'link2']])
<ul class="web2py-menu web2py-menu-vertical">
  <li><a href="link1">One</a></li>
  <li><a href="link2">Two</a></li>
</ul>

Each menu item can have a fourth argument that is a nested submenu (and so on recursively):

>>> print MENU([['One', False, 'link1', [['Two', False, 'link2']]]])
<ul class="web2py-menu web2py-menu-vertical">
  <li class="web2py-menu-expand">
     <a href="link1">One</a>
     <ul class="web2py-menu-vertical">
        <li><a href="link2">Two</a></li>
     </ul>
  </li>
</ul>

The MENU helper takes the following optional arguments:

  • _class: defaults to "web2py-menu web2py-menu-vertical" and sets the class of the outer UL elements.
  • ul_class: defaults to "web2py-menu-vertical" and sets the class of the inner UL elements.
  • li_class: defaults to "web2py-menu-expand" and sets the class of the inner LI elements.
mobile

MENU takes an optional argument mobile. When set to True instead of building a recursive UL menu structure it returns a SELECT dropdown with all the menu options and a onchange attribute that redirects to the page corresponding to the selected option. This is designed an an alterantive menu representation that increases usability on small mobile devices such as phones.

Normally the menu is used in a layout with the following syntax:

{{=MENU(response.menu, mobile=request.user_agent().is_mobile)}}

In this way a mobile device is automatically detected and the menu is rendered accordingly.

BEAUTIFY

BEAUTIFY is used to build HTML representations of compound objects, including lists, tuples and dictionaries:

{{=BEAUTIFY({"a": ["hello", XML("world")], "b": (1, 2)})}}

BEAUTIFY returns an XML-like object serializable to XML, with a nice looking representation of its constructor argument. In this case, the XML representation of:

{"a": ["hello", XML("world")], "b": (1, 2)}

will render as:

<table>
<tr><td>a</td><td>:</td><td>hello<br />world</td></tr>
<tr><td>b</td><td>:</td><td>1<br />2</td></tr>
</table>

Server-side DOM'' and parsing

element
elements

elements

The DIV helper and all derived helpers provide the search methods element and elements.

element returns the first child element matching a specified condition (or None if no match).

elements returns a list of all matching children.

element and elements use the same syntax to specify the matching condition, which allows for three possibilities that can be mixed and matched: jQuery-like expressions, match by exact attribute value, match using regular expressions.

Here is a simple example:

>>> a = DIV(DIV(DIV('a', _id='target',_class='abc')))
>>> d = a.elements('div#target')
>>> d[0][0] = 'changed'
>>> print a
<div><div><div id="target" class="abc">changed</div></div></div>

The un-named argument of elements is a string, which may contain: the name of a tag, the id of a tag preceded by a pound symbol, the class preceded by a dot, the explicit value of an attribute in square brackets.

Here are 4 equivalent ways to search the previous tag by id:

>>> d = a.elements('#target')
>>> d = a.elements('div#target')
>>> d = a.elements('div[id=target]')
>>> d = a.elements('div',_id='target')

Here are 4 equivalent ways to search the previous tag by class:

>>> d = a.elements('.abc')
>>> d = a.elements('div.abc')
>>> d = a.elements('div[class=abc]')
>>> d = a.elements('div',_class='abc')

Any attribute can be used to locate an element (not just id and class), including multiple attributes (the function element can take multiple named arguments), but only the first matching element will be returned.

Using the jQuery syntax "div#target" it is possible to specify multiple search criteria separated by a space:

>>> a = DIV(SPAN('a', _id='t1'),DIV('b',_class='c2'))
>>> d = a.elements('span#t1, div#c2')

or equivalently

>>> a = DIV(SPAN('a', _id='t1'),DIV('b',_class='c2'))
>>> d = a.elements('span#t1','div#c2')

If the value of an attribute is specified using a name argument, it can be a string or a regular expression:

>>> a = DIV(SPAN('a', _id='test123'), DIV('b',_class='c2'))
>>> d = a.elements('span', _id=re.compile('test\d{3}')

A special named argument of the DIV (and derived) helpers is find. It can be used to specify a search value or a search regular expression in the text content of the tag. For example:

>>> a = DIV(SPAN('abcde'), DIV('fghij'))
>>> d = a.elements(find='bcd')
>>> print d[0]
<span>abcde</span>

or

>>> a = DIV(SPAN('abcde'), DIV('fghij'))
>>> d = a.elements(find=re.compile('fg\w{3}'))
>>> print d[0]
<div>fghij</div>

components

Here's an example of listing all elements in an html string:

html = TAG('<a>xxx</a><b>yyy</b>')
for item in html.components: print item

parent

parent returns the parent of the current element.

>>> a = DIV(SPAN('a'),DIV('b'))
>>> d = a.element('a').parent()
>>> d['_class']='abc'
>>> print a
<div class="abc"><span>a</span><div>b</div></div>

flatten

The flatten method recursively serializes the content of the children of a given element into regular text (without tags):

>>> a = DIV(SPAN('this', DIV('is', B('a'))), SPAN('test'))
>>> print a.flatten()
thisisatest

Flatten can be passed an optional argument, render, i.e. a function that renders/flattens the content using a different protocol. Here is an example to serialize some tags into Markmin wiki syntax:

>>> a = DIV(H1('title'), P('example of a ', A('link', _href='#test')))
>>> from gluon.html import markmin_serializer
>>> print a.flatten(render=markmin_serializer)
## titles

example of [[a link #test]]

At the time of writing we provide markmin_serializer and markdown_serializer.

Parsing

The TAG object is also an XML/HTML parser. It can read text and convert into a tree structure of helpers. This allows manipulation using the API above:

>>> html = '<h1>Title</h1><p>this is a <span>test</span></p>'
>>> parsed_html = TAG(html)
>>> parsed_html.element('span')[0]='TEST'
>>> print parsed_html
<h1>Title</h1><p>this is a <span>TEST</span></p>

Page layout

page layout
layout.html
extent
include

Views can extend and include other views in a tree-like structure.

For example, we can think of a view "index.html" that extends "layout.html" and includes "body.html". At the same time, "layout.html" may include "header.html" and "footer.html".

The root of the tree is what we call a layout view. Just like any other HTML template file, you can edit it using the web2py administrative interface. The file name "layout.html" is just a convention.

Here is a minimalist page that extends the "layout.html" view and includes the "page.html" view:

{{extend 'layout.html'}}
<h1>Hello World</h1>
{{include 'page.html'}}

The extended layout file must contain an {{include}} directive, something like:

<html>
  <head>
    <title>Page Title</title>
  </head>
  <body>
    {{include}}
  </body>
</html>

When the view is called, the extended (layout) view is loaded, and the calling view replaces the {{include}} directive inside the layout. Processing continues recursively until all extend and include directives have been processed. The resulting template is then translated into Python code. Note, when an application is bytecode compiled, it is this Python code that is compiled, not the original view files themselves. So, the bytecode compiled version of a given view is a single .pyc file that includes the Python code not just for the original view file, but for its entire tree of extended and included views.

extend, include, block and super are special template directives, not Python commands.

Any content or code that precedes the {{extend ...}} directive will be inserted (and therefore executed) before the beginning of the extended view's content/code. Although this is not typically used to insert actual HTML content before the extended view's content, it can be useful as a means to define variables or functions that you want to make available to the extended view. For example, consider a view "index.html":

{{sidebar_enabled=True}}
{{extend 'layout.html'}}
<h1>Home Page</h1>

and an excerpt from "layout.html":

{{if sidebar_enabled:}}
    <div id="sidebar">
        Sidebar Content
    </div>
{{pass}}

Because the sidebar_enabled assignment in "index.html" comes before the extend, that line gets inserted before the beginning of "layout.html", making sidebar_enabled available anywhere within the "layout.html" code (a somewhat more sophisticated version of this is used in the welcome app).

It is also worth pointing out that the variables returned by the controller function are available not only in the function's main view, but in all of its extended and included views as well.

The argument of an extend or include (i.e., the extended or included view name) can be a python variable (though not a python expression). However, this imposes a limitation -- views that use variables in extend or include statements cannot be bytecode compiled. As noted above, bytecode compiled views include the entire tree of extended and included views, so the specific extended and included views must be known at compile time, which is not possible if the view names are variables (whose values are not determined until run time). Because bytecode compiling views can provide a significant speed boost, using variables in extend and include should generally be avoided if possible.

In some cases, an alternative to using a variable in an include is simply to place regular {{include ...}} directives inside an if...else block.

{{if some_condition:}}
{{include 'this_view.html'}}
{{else:}}
{{include 'that_view.html'}}
{{pass}}

The above code does not present any problem for bytecode compilation because no variables are involved. Note, however, that the bytecode compiled view will actually include the Python code for both "this_view.html" and "that_view.html", though only the code for one of those views will be executed, depending on the value of some_condition.

Keep in mind, this only works for include -- you cannot place {{extend ...}} directives inside if...else blocks.

response.menu
menu
response.meta
meta

Layouts are used to encapsulate page commonality (headers, footers, menus), and though they are not mandatory, they will make your application easier to write and maintain. In particular, we suggest writing layouts that take advantage of the following variables that can be set in the controller. Using these well known variables will help make your layouts interchangeable:

response.title
response.subtitle
response.meta.author
response.meta.keywords
response.meta.description
response.flash
response.menu
response.files

Except for menu and files, these are all strings and their meaning should be obvious.

response.menu menu is a list of 3-tuples or 4-tuples. The three elements are: the link name, a boolean representing whether the link is active (is the current link), and the URL of the linked page. For example:

response.menu = [('Google', False, 'http://www.google.com',[]),
                 ('Index',  True,  URL('index'), [])]
sub-menu

The fourth tuple element is an optional sub-menu.

response.files is a list of CSS and JS files that are needed by your page.

We also recommend that you use:

{{include 'web2py_ajax.html'}}

in the HTML head, since this will include the jQuery libraries and define some backward-compatible JavaScript functions for special effects and Ajax. "web2py_ajax.html" includes the response.meta tags in the view, jQuery base, the calendar datepicker, and all required CSS and JS response.files.

Default page layout

superfish
ez.css

Below is the "views/layout.html" that ships with the web2py scaffolding application welcome (stripped down of some optional parts). Any new application will have a similar default layout:

<!DOCTYPE html>
<head>
  <meta charset="utf-8" />
  <title>{{=response.title or request.application}}</title>

  <!-- http://dev.w3.org/html5/markup/meta.name.html -->
  <meta name="application-name" content="{{=request.application}}" />

  <script src="{{=URL('static','js/modernizr.custom.js')}}"></script>

  <!-- include stylesheets -->
  {{
  response.files.append(URL('static','css/skeleton.css'))
  response.files.append(URL('static','css/web2py.css'))
  response.files.append(URL('static','css/superfish.css'))
  response.files.append(URL('static','js/superfish.js'))
  }}

  {{include 'web2py_ajax.html'}}

  <script type="text/javascript">
    jQuery(function(){ jQuery('ul.sf-menu').supersubs({minWidth:12,maxWidth:30,extraWidth:3}).superfish(); });
  </script>

  {{
  # using sidebars need to know what sidebar you want to use
  left_sidebar_enabled = globals().get('left_sidebar_enabled',False)
  right_sidebar_enabled = globals().get('right_sidebar_enabled',False)
  middle_columns = {0:'sixteen',1:'twelve',2:'eight'}[
     (left_sidebar_enabled and 1 or 0)+(right_sidebar_enabled and 1 or 0)]
  }}

</head>
<body>
  <div class="wrapper"><!-- for sticky footer -->

    <div class="topbar">
      <div class="container">
        <div class="sixteen columns">
          <div id="navbar">
            {{='auth' in globals() and auth.navbar(separators=(' ',' | ',''))}}
          </div>
          <div id="menu">
            {{=MENU(response.menu, 
                    _class='mobile-menu' if is_mobile else 'sf-menu',
                    mobile=request.user_agent().is_mobile)}}
          </div>
        </div>
      </div>
    </div><!-- topbar -->

    <div class="flash">{{=response.flash or ''}}</div>

    <div class="header">
      <div class="container">
        <div class="sixteen columns">
          <h1 class="remove-bottom" style="margin-top: .5em;">
          {{=response.title or request.application}}
          </h1>
          <h5>{{=response.subtitle or ''}}</h5>
        </div>

        <div class="sixteen columns">
          <div class="statusbar">
            {{block statusbar}}
            <span class="breadcrumbs">{{=request.function}}</span>
            {{end}}
          </div>
        </div>
      </div>
    </div>

    <div class="main">
      <div class="container">
        {{if left_sidebar_enabled:}}
        <div class="four columns left-sidebar">
          {{block left_sidebar}}
          <h3>Left Sidebar</h3>
          <p></p>
          {{end}}
        </div>
        {{pass}}

        <div class="{{=middle_columns}} columns center">
          {{block center}}
          {{include}}
          {{end}}
        </div>

        {{if right_sidebar_enabled:}}
        <div class="four columns">
          {{block right_sidebar}}
          <h3>Right Sidebar</h3>
          <p></p>
          {{end}}
        </div>
        {{pass}}

      </div><!-- container -->
    </div><!-- main -->

    <div class="push"></div>
  </div><!-- wrapper -->

  <div class="footer">
    <div class="container header">
      <div class="sixteen columns">
        {{block footer}} <!-- this is default footer -->
        <div class="footer-content" >
          {{=T('Copyright')}} &#169; 2011
          <div style="float: right;">
            <a href="http://www.web2py.com/">
            <img style="padding-bottom: 0;" 
                 src="{{=URL('static','images/poweredby.png')}}"/>
            </a>
          </div>
        </div>
        {{end}}
      </div>
    </div><!-- container -->
  </div><!-- footer -->

</body>
</html>

There are a few features of this default layout that make it very easy to use and customize:

  • It is written in HTML5 and uses the "modernizr" [modernizr] library for backward compatibility. The actual layout include some extra conditional statements required by IE and they are omitted for brevity.
  • It displays both response.title and response.subtitle which can be set in a model. If they are not set, it adopts the application name as title
  • It includes the web2py_ajax.html file in the header which generated all the link and script import statements.
  • It uses a modified version of "skeleton" [skeleton] a library for flexible layouts which works on mobile devices and re-arranges columns to fit small screened.
  • It uses "superfish.js" for dynamic cascading menus. There is an explicit script to activate the superfish cascading menu and it can be removed if not necessary.
  • The {{=auth.navbar(...)}} displays a welcome to the current user and links to auth functions like login, logout, register, change password, etc. depending on context. It is a heler factory and its output can be manipulated as any other helper. It is placed in a {{try:}}...{{except:pass}} in case auth is undefined.
  • The {{=MENU(response.menu) displays the menu structure as <ul>...</ul>.
  • {{include}} is replaced by the content of the extending view when the page is rendered.
  • By default it uses a conditional three column (the left and right sidebars can be turned off by the extending views)
  • It uses the following classes: header, main, footer
  • It contains the following blocks: statusbar, left_sidebar, center, right_sidebar, footer.

Views can turn on and fill sidebars as follows:

{{left_sidebar_enable=True}}
{{extend 'layout.html'}}

This text goes in center

{{block left_sidebar}}
This text goes in sidebar
{{end}}

Customizing the default layout

CSS

Customizing the default layout without editing is very easy because the CSS files are documented:

  • "skeleton.css" contains reset, grid layout and form styles
  • "web2py.css" contains web2py specific styles
  • "superfish.css" contains menu styles

To change colors and background images, simply append the following code to "web2py.css":

body { background: url('images/background.png') repeat-x #3A3A3A; }
a { color: #349C01; }
.header h1 { color: #349C01; }
.header h2 { color: white; font-style: italic; font-size: 14px;}
.statusbar { background: #333333; border-bottom: 5px #349C01 solid; }
.statusbar a { color: white; }
.footer { border-top: 5px #349C01 solid; }

The menu is built in a color-neutral way but you can change that too.

Of course you can also completely replace the "layout.html" and "web2py.css" files with your own.

Mobile development

The default layout.html is designed to be friendly to mobile devices but that is not enought. One may need to use different views when a page is visited by a mobile device.

To make developing for desktop and mobile devices easier, web2py includes the @mobilize decorator. This decorator is applied to actions that should have a normal view and a mobile view. This is demonstrated here:

from gluon.contrib.user_agent_parser import mobilize 
@mobilize
def index():
   return dict()

Notice that the decorator must be important once before using it in a controller. When the "index" function is called from a regular browser (desktop computer), web2py will render the returned dictionary using the view "[controller]/index.html". However, when it is called by a mobile device, the dictionary will be rendered by "[controller]/index.mobile.html". Notice that mobile views have the "mobile.html" extension.

Alternatively you can apply the following logic to make all views mobile friendly:

if request.user_agent().is_mobile:
    response.view.replace('.html','.mobile.html')

The task of creating the "*.mobile.html" views is left to the developer but we strongly suggest using the "jQuery Mobile" plugin which makes the task very easy.

Functions in views

Consider this "layout.html":

<html>
  <body>
    {{include}}
    <div class="sidebar">
      {{if 'mysidebar' in globals():}}{{mysidebar()}}{{else:}}
        my default sidebar
      {{pass}}
    </div>
  </body>
</html>

and this extending view

{{def mysidebar():}}
my new sidebar!!!
{{return}}
{{extend 'layout.html'}}
Hello World!!!

Notice the function is defined before the {{extend...}} statement -- this results in the function being created before the "layout.html" code is executed, so the function can be called anywhere within "layout.html", even before the {{include}}. Also notice the function is included in the extended view without the = prefix.

The code generates the following output:

<html>
  <body>
    Hello World!!!
    <div class="sidebar">
        my new sidebar!!!
    </div>
  </body>
</html>

Notice that the function is defined in HTML (although it could also contain Python code) so that response.write is used to write its content (the function does not return the content). This is why the layout calls the view function using {{mysidebar()}} rather than {{=mysidebar()}}. Functions defined in this way can take arguments.

Blocks in views

block

Another way to make a view more modular is by using {{block...}}s and this mechanism is an alternative to the mechanism discussed in the previous section.

Consider this "layout.html":

<html>
  <body>
    {{include}}
    <div class="sidebar">
      {{block mysidebar}}
        my default sidebar
      {{end}}
    </div>
  </body>
</html>

and this extending view

{{extend 'layout.html'}}
Hello World!!!
{{block mysidebar}}
my new sidebar!!!
{{end}}

It generates the following output:

<html>
  <body>
    Hello World!!!
    <div class="sidebar">
        my new sidebar!!!
    </div>
  </body>
</html>

You can have many blocks, and if a block is present in the extended view but not in the extending view, the content of the extended view is used. Also, notice that unlike with functions, it is not necessary to define blocks before the {{extend ...}} -- even if defined after the extend, they can be used to make substitutions anywhere in the extended view.

super

Inside a block, you can use the expression {{super}} to include the content of the parent. For example, if we replace the above extending view with:

{{extend 'layout.html'}}
Hello World!!!
{{block mysidebar}}
{{super}}
my new sidebar!!!
{{end}}

we get:

<html>
  <body>
    Hello World!!!
    <div class="sidebar">
        my default sidebar
        my new sidebar!!!
    </div>
  </body>
</html>
 top