Flexible resources for web applications.
Project description
hurry.resource
Introduction
Resources are files that are used as resources in the display of a web page, such as CSS files, Javascript files and images. Resources packaged together in a directory to be published as such are called a resource library.
When a resource is included in the head section of a HTML page, we call this a resource inclusion. An inclusion is of a particular resource in a particular library. There are two forms of this kind of inclusion in HTML: javascript is included using the script tag, and CSS (and KSS) are included using a link tag.
Inclusions may depend on other inclusions. A javascript resource may for instance be built on top of another javascript resource. This means both of them should be loaded when the page displays.
Page components may actually require a certain inclusion in order to be functional. A widget may for instance expect a particular Javascript library to loaded. We call this an inclusion requirement of the component.
hurry.resource provides a simple API to specify resource libraries, inclusion and inclusion requirements.
A resource library
We define a library foo. It takes two arguments, the name of the library as it should be published under in a URL and uniquely identify it, and a path to the root of the resources (rootpath) that this library publishes:
>>> from hurry.resource import Library >>> foo = Library('foo', 'dummy')
The full path to the directory with the resources is reconstructed from the package that the Library is defined in:
>>> foo.path #doctest: +ELLIPSIS '.../hurry.resource/src/hurry/resource/dummy'
Entry points
Libraries can be exposed for registration by whatever web framework that hurry.resource is integrated with. This web framework can then expose the library path on a URL somewhere. This is done using the hurry.resource.libraries entry point. To register Library instances foo and bar in your package as entry points include a section like this in your setup.py:
entry_points={ 'hurry.resource.libraries': [ 'foo = mypackage.foomodule:foo', 'bar = mypackage.barmodule:bar', ], }
There is an API to help you obtain all registered libraries:
>>> from hurry.resource import libraries
Nothing is registered however:
>>> list(libraries()) []
It would be nice to now have some tests that see whether entry points actually get picked up so, but that would require an involved test setup that we find difficult to construct.
Inclusion
We now create an inclusion of a particular resource in a library. This inclusion needs a.js from library and b.js as well:
>>> from hurry.resource import ResourceInclusion >>> x1 = ResourceInclusion(foo, 'a.js') >>> x2 = ResourceInclusion(foo, 'b.css')
Let’s now make an inclusion y1 that depends on x1 and x2:
>>> y1 = ResourceInclusion(foo, 'c.js', depends=[x1, x2])
Inclusion requirements
When rendering a web page we want to require the inclusion of a resource anywhere within the request handling process. We might for instance have a widget that takes care of rendering its own HTML but also needs a resource to be included in the page header.
We have a special object that represents the needed inclusions during a certain request cycle:
>>> from hurry.resource import NeededInclusions >>> needed = NeededInclusions()
We state that a resource is needed by calling the need method on this object:
>>> needed.need(y1)
Let’s now see what resources are needed by this inclusion:
>>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'b.css' in library 'foo'>, <ResourceInclusion 'a.js' in library 'foo'>, <ResourceInclusion 'c.js' in library 'foo'>]
As you can see, css resources are sorted before js resources.
Grouping resources
It is also possible to define a group that doesn’t get rendered itself, but groups other resources together that should be rendered:
>>> from hurry.resource import GroupInclusion >>> group = GroupInclusion([x1, x2])
When we need a group, we’ll get all inclusions referenced in it:
>>> needed = NeededInclusions() >>> needed.need(group) >>> group.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'a.js' in library 'foo'>, <ResourceInclusion 'b.css' in library 'foo'>]
A group can also be depended on; it won’t show up in the list of inclusions directly:
>>> more_stuff = ResourceInclusion(foo, 'more_stuff.js', depends=[group]) >>> more_stuff.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'a.js' in library 'foo'>, <ResourceInclusion 'b.css' in library 'foo'>, <ResourceInclusion 'more_stuff.js' in library 'foo'>]
A convenience spelling
When specifying that we want a resource inclusion to be rendered, we now need access to the current NeededInclusions object and the resource inclusion itself.
Let’s introduce a more convenient spelling of needs now:
y1.need()
We can require a resource without reference to the needed inclusions object directly as there is typically only a single set of needed inclusions that is generated during the rendering of a page.
So let’s try out this spelling to see it fail:
>>> y1.need() Traceback (most recent call last): ... NotImplementedError: need to implement plugin.get_current_needed_inclusions()
We get an error because we haven’t configured the framework yet. The system says we need to implement plugin.get_current_needed_inclusions() first. This is a method that we need to implement so we can tell the system how to obtain the current NeededInclusions object.
This needed inclusions should be maintained on an object that is going to be present throughout the request/response cycle that generates the web page that has the inclusions on them. One place where we can maintain the needed inclusions is the request object itself, if we indeed have global access to it. Alternatively you could store the currently needed inclusions in a thread local variable.
Let’s introduce a simple request object (your mileage may vary in your own web framework):
>>> class Request(object): ... def __init__(self): ... self.needed = NeededInclusions()
We now make a request, imitating what happens during a typical request/response cycle in a web framework:
>>> request = Request()
We now define a plugin class that implements the get_current_needed_inclusions() method by obtaining it from the request:
>>> class Plugin(object): ... def get_current_needed_inclusions(self): ... return request.needed
We now need to register this plugin with the framework:
>>> from hurry.resource import register_plugin >>> register_plugin(Plugin())
There is an API to retrieve the current needed inclusions, so let’s check which resources our request needs currently:
>>> from hurry.resource import get_current_needed_inclusions >>> get_current_needed_inclusions().inclusions() []
Nothing yet. We now make y1 needed using our simplified spelling:
>>> y1.need()
The resource inclusion will now indeed be needed:
>>> get_current_needed_inclusions().inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'b.css' in library 'foo'>, <ResourceInclusion 'a.js' in library 'foo'>, <ResourceInclusion 'c.js' in library 'foo'>]
Let’s go back to the original spelling of needed.need(y) now. While this is a bit more cumbersome to use in application code, it is easier to read for the purposes of this document.
A note on optimization
There are various optimizations for resource inclusion that hurry.resource supports. Because some optimizations can make debugging more difficult, the optimizations are disabled by default.
We will summarize the optimization features here and tell you how to enable them. Later sections below go into more details.
minified resources. Resources can specify minified versions using the mode system. You can use hurry.resource.mode('minified') somewhere in the request handling of your application. This will make sure that resources included on your page are supplied as minified versions, if these are available.
rolling up of resources. Resource libraries can specify rollup resources that combine multiple resources into one. This reduces the amount of server requests to be made by the web browser, and can help with caching. To enable rolling up, call hurry.resource.rollup somewhere in your request handling.
javascript inclusions at the bottom of the web page. If your framework integration uses the special render_topbottom method, you can enable the inclusion of javascript files at the bottom by calling hurry.resource.bottom(). This will only include resources at the bottom that have explicitly declared themselves to be bottom-safe. You can declare a resource bottom safe by passing bottom=True when constructing a ResourceInclusion. If you want to force all javascript to be including at the bottom of the page by default, you can call hurry.resource.bottom(force=True).
To find out more about these and other optimizations, please read this best practices article that describes some common optimizations to speed up page load times.
Multiple requirements
In this section, we will show what happens in various scenarios where we requiring multiple ResourceInclusion objects.
We create a new set of needed inclusions:
>>> needed = NeededInclusions() >>> needed.inclusions() []
We need y1 again:
>>> needed.need(y1) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'b.css' in library 'foo'>, <ResourceInclusion 'a.js' in library 'foo'>, <ResourceInclusion 'c.js' in library 'foo'>]
Needing the same inclusion twice won’t make any difference for the resources needed. So when we need y1 again, we see no difference in the needed resources:
>>> needed.need(y1) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'b.css' in library 'foo'>, <ResourceInclusion 'a.js' in library 'foo'>, <ResourceInclusion 'c.js' in library 'foo'>]
Needing x1 or x2 won’t make any difference either, as y1 already required x1 and x2:
>>> needed.need(x1) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'b.css' in library 'foo'>, <ResourceInclusion 'a.js' in library 'foo'>, <ResourceInclusion 'c.js' in library 'foo'>] >>> needed.need(x2) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'b.css' in library 'foo'>, <ResourceInclusion 'a.js' in library 'foo'>, <ResourceInclusion 'c.js' in library 'foo'>]
Let’s do it in reverse, and require the x1 and x2 resources before we need those in y1. Again this makes no difference:
>>> needed = NeededInclusions() >>> needed.need(x1) >>> needed.need(x2) >>> needed.need(y1) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'b.css' in library 'foo'>, <ResourceInclusion 'a.js' in library 'foo'>, <ResourceInclusion 'c.js' in library 'foo'>]
Let’s try it with more complicated dependency structures now:
>>> needed = NeededInclusions() >>> a1 = ResourceInclusion(foo, 'a1.js') >>> a2 = ResourceInclusion(foo, 'a2.js', depends=[a1]) >>> a3 = ResourceInclusion(foo, 'a3.js', depends=[a2]) >>> a4 = ResourceInclusion(foo, 'a4.js', depends=[a1]) >>> needed.need(a3) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'a1.js' in library 'foo'>, <ResourceInclusion 'a2.js' in library 'foo'>, <ResourceInclusion 'a3.js' in library 'foo'>] >>> needed.need(a4) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'a1.js' in library 'foo'>, <ResourceInclusion 'a2.js' in library 'foo'>, <ResourceInclusion 'a3.js' in library 'foo'>, <ResourceInclusion 'a4.js' in library 'foo'>]
If we reverse the requirements for a4 and a3, we get the following inclusion structure, based on the order in which need was expressed:
>>> needed = NeededInclusions() >>> needed.need(a4) >>> needed.need(a3) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'a1.js' in library 'foo'>, <ResourceInclusion 'a4.js' in library 'foo'>, <ResourceInclusion 'a2.js' in library 'foo'>, <ResourceInclusion 'a3.js' in library 'foo'>]
Let’s look at the order in which resources are listed when we need something that ends up depending on everything:
>>> a5 = ResourceInclusion(foo, 'a5.js', depends=[a4, a3]) >>> needed = NeededInclusions() >>> needed.need(a5) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'a1.js' in library 'foo'>, <ResourceInclusion 'a4.js' in library 'foo'>, <ResourceInclusion 'a2.js' in library 'foo'>, <ResourceInclusion 'a3.js' in library 'foo'>, <ResourceInclusion 'a5.js' in library 'foo'>]
When we introduce the extra inclusion of a3 earlier on, we still get a valid list of inclusions given the dependency structure, even though the sorting order is different:
>>> needed = NeededInclusions() >>> needed.need(a3) >>> needed.need(a5) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'a1.js' in library 'foo'>, <ResourceInclusion 'a2.js' in library 'foo'>, <ResourceInclusion 'a3.js' in library 'foo'>, <ResourceInclusion 'a4.js' in library 'foo'>, <ResourceInclusion 'a5.js' in library 'foo'>]
Modes
A resource can optionally exist in several modes, such as for instance a minified and a debug version. Let’s define a resource that exists in two modes (a main one and a debug alternative):
>>> k1 = ResourceInclusion(foo, 'k.js', debug='k-debug.js')
Let’s need this resource:
>>> needed = NeededInclusions() >>> needed.need(k1)
By default, we get k.js:
>>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'k.js' in library 'foo'>]
We can however also get the resource for mode debug and get k-debug.js:
>>> needed.mode('debug') >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'k-debug.js' in library 'foo'>]
Modes can also be specified fully with a resource inclusion, which allows you to specify a different library argumnent:
>>> k2 = ResourceInclusion(foo, 'k2.js', ... debug=ResourceInclusion(foo, 'k2-debug.js')) >>> needed = NeededInclusions() >>> needed.need(k2)
By default we get k2.js:
>>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'k2.js' in library 'foo'>]
We can however also get the resource for mode debug and get k2-debug.js:
>>> needed.mode('debug') >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'k2-debug.js' in library 'foo'>]
Note that modes are assumed to be identical in dependency structure; they functionally should do the same.
If you request a mode and a resource doesn’t support it, it will return its default resource instead:
>>> needed = NeededInclusions() >>> needed.mode('minified') >>> needed.need(k1) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'k.js' in library 'foo'>]
hurry.resource suggests resource libraries follow the following conventions for modes:
default - the original source text, non-minified, and without any special extra debugging functionality.
debug - an optional version of the source text that offers more debugging support, such as logging.
minified - an optional minified (compressed) form of the resource.
In the case of rollups, several resources can be consolidated into one larger one for optimization purposes. A library might only offer a minified version of a rollup resource; if the developer wants to debug, it is expected he uses the resources in non-rolledup format. In this case you should make a resource inclusion where the default mode is equal to the minified mode, like this:
>>> example = ResourceInclusion(foo, 'k.js', minified='k.js')
If the developer wants to debug, he will need to disable rolling up (by calling hurry.resource.rollup(disable=True), or by simply never calling hurry.resource.rollup() in the request cycle).
Mode convenience
Like for need, there is also a convenience spelling for mode. It uses ICurrentNeededInclusions, which we’ve already set up to look at the request.needed variable. Let’s set up a new request:
>>> request = Request()
Let’s set up a resource and need it:
>>> l1 = ResourceInclusion(foo, 'l1.js', debug='l1-debug.js') >>> l1.need()
Let’s look at the resources needed by default:
>>> c = get_current_needed_inclusions() >>> c.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'l1.js' in library 'foo'>]
Let’s now change the mode using the convenience hurry.resource.mode spelling:
>>> from hurry.resource import mode >>> mode('debug')
When we request the resources now, we get them in the debug mode:
>>> c.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'l1-debug.js' in library 'foo'>]
“Rollups”
For performance reasons it’s often useful to consolidate multiple resources into a single, larger resource, a so-called “rollup”. Multiple javascript files could for instance be offered in a single, larger one. These consolidations can be specified as a resource:
>>> b1 = ResourceInclusion(foo, 'b1.js') >>> b2 = ResourceInclusion(foo, 'b2.js') >>> giant = ResourceInclusion(foo, 'giant.js', supersedes=[b1, b2])
Rolling up of resources is not enabled by default, as sometimes a library only offers these rollups in minified form, and automatically rolling up would not be nice during debugging. It’s therefore a performance feature you can enable.
Without rollups enabled nothing special happens:
>>> needed = NeededInclusions() >>> needed.need(b1) >>> needed.need(b2) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'b1.js' in library 'foo'>, <ResourceInclusion 'b2.js' in library 'foo'>]
Let’s enable rollups:
>>> needed = NeededInclusions() >>> needed.rollup()
The convenience spelling to enable rollups during request handling looks like this:
hurry.resource.rollup()
If we now find multiple resources that are also part of a consolidation, the system automatically collapses them:
>>> needed.need(b1) >>> needed.need(b2) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'giant.js' in library 'foo'>]
The system will by default only consolidate exactly. That is, if only a single resource out of two is present, the consolidation will not be triggered:
>>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(b1) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'b1.js' in library 'foo'>]
Let’s look at this with a larger consolidation of 3 resources:
>>> c1 = ResourceInclusion(foo, 'c1.css') >>> c2 = ResourceInclusion(foo, 'c2.css') >>> c3 = ResourceInclusion(foo, 'c3.css') >>> giantc = ResourceInclusion(foo, 'giantc.css', supersedes=[c1, c2, c3])
It will not roll up one resource:
>>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(c1) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'c1.css' in library 'foo'>]
Neither will it roll up two resources:
>>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(c1) >>> needed.need(c2) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'c1.css' in library 'foo'>, <ResourceInclusion 'c2.css' in library 'foo'>]
It will however roll up three resources:
>>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(c1) >>> needed.need(c2) >>> needed.need(c3) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'giantc.css' in library 'foo'>]
The default behavior is to play it safe: we cannot be certain that we do not include too much if we were to include giantc.css if only c1 and c2 are required. This is especially important with CSS libraries: if only c1.css and c2.css are to be included in a page, including giantc.css is not appropriate as that also includes the content of c3.css, which might override and extend the behavior of c1.css and c2.css.
The situation is sometimes different with Javascript libraries, which can be written in such a way that a larger rollup will just include more functions, but will not actually affect page behavior. If we have a rollup resource that we don’t mind kicking in even if part of the requirements have been met, we can indicate this:
>>> d1 = ResourceInclusion(foo, 'd1.js') >>> d2 = ResourceInclusion(foo, 'd2.js') >>> d3 = ResourceInclusion(foo, 'd3.js') >>> giantd = ResourceInclusion(foo, 'giantd.js', supersedes=[d1, d2, d3], ... eager_superseder=True)
We will see giantd.js kick in even if we only require d1 and d2:
>>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(d1) >>> needed.need(d2) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'giantd.js' in library 'foo'>]
In fact even if we only need a single resource the eager superseder will show up instead:
>>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(d1) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'giantd.js' in library 'foo'>]
If there are two potential eager superseders, the biggest one will be taken:
>>> d4 = ResourceInclusion(foo, 'd4.js') >>> giantd_bigger = ResourceInclusion(foo, 'giantd-bigger.js', ... supersedes=[d1, d2, d3, d4], eager_superseder=True) >>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(d1) >>> needed.need(d2) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'giantd-bigger.js' in library 'foo'>]
If there is a potential non-eager superseder and an eager one, the eager one will be taken:
>>> giantd_noneager = ResourceInclusion(foo, 'giantd-noneager.js', ... supersedes=[d1, d2, d3, d4]) >>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(d1) >>> needed.need(d2) >>> needed.need(d3) >>> needed.need(d4) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'giantd-bigger.js' in library 'foo'>]
A resource can be part of multiple rollups. In this case the rollup that rolls up the most resources is used. So, if there are two potential non-eager superseders, the one that rolls up the most resources will be used:
>>> e1 = ResourceInclusion(foo, 'e1.js') >>> e2 = ResourceInclusion(foo, 'e2.js') >>> e3 = ResourceInclusion(foo, 'e3.js') >>> giante_two = ResourceInclusion(foo, 'giante-two.js', ... supersedes=[e1, e2]) >>> giante_three = ResourceInclusion(foo, 'giante-three.js', ... supersedes=[e1, e2, e3]) >>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(e1) >>> needed.need(e2) >>> needed.need(e3) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'giante-three.js' in library 'foo'>]
Consolidation also works with modes:
>>> f1 = ResourceInclusion(foo, 'f1.js', debug='f1-debug.js') >>> f2 = ResourceInclusion(foo, 'f2.js', debug='f2-debug.js') >>> giantf = ResourceInclusion(foo, 'giantf.js', supersedes=[f1, f2], ... debug='giantf-debug.js') >>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(f1) >>> needed.need(f2) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'giantf.js' in library 'foo'>] >>> needed.mode('debug') >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'giantf-debug.js' in library 'foo'>]
What if the rolled up resources have no mode but the superseding resource does? In this case the superseding resource’s mode has no meaning, so modes have no effect:
>>> g1 = ResourceInclusion(foo, 'g1.js') >>> g2 = ResourceInclusion(foo, 'g2.js') >>> giantg = ResourceInclusion(foo, 'giantg.js', supersedes=[g1, g2], ... debug='giantg-debug.js') >>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(g1) >>> needed.need(g2) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'giantg.js' in library 'foo'>] >>> needed.mode('debug') >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'giantg.js' in library 'foo'>]
What if the rolled up resources have a mode but the superseding resource does not? Let’s look at that scenario:
>>> h1 = ResourceInclusion(foo, 'h1.js', debug='h1-debug.js') >>> h2 = ResourceInclusion(foo, 'h2.js', debug='h2-debug.js') >>> gianth = ResourceInclusion(foo, 'gianth.js', supersedes=[h1, h2]) >>> needed = NeededInclusions() >>> needed.rollup() >>> needed.need(h1) >>> needed.need(h2) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'gianth.js' in library 'foo'>]
Since there is no superseder for the debug mode, we will get the two resources, not rolled up:
>>> needed.mode('debug') >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'h1-debug.js' in library 'foo'>, <ResourceInclusion 'h2-debug.js' in library 'foo'>]
Rendering resources
Let’s define some needed resource inclusions:
>>> needed = NeededInclusions() >>> needed.need(y1) >>> needed.inclusions() #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'b.css' in library 'foo'>, <ResourceInclusion 'a.js' in library 'foo'>, <ResourceInclusion 'c.js' in library 'foo'>]
Now let’s try to render these inclusions:
>>> print needed.render() Traceback (most recent call last): ... AttributeError: 'Plugin' object has no attribute 'get_library_url'
That didn’t work. In order to render an inclusion, we need to tell hurry.resource how to get the URL for a resource inclusion. We already know the relative URL, so we need to specify how to get a URL to the library itself that the relative URL can be added to.
We’ll extend the existing plugin that already knows how to obtain the current needed inclusions. For the purposes of this document, we define a function that renders resources as some static URL on localhost:
>>> class NewPlugin(Plugin): ... def get_library_url(self, library): ... return 'http://localhost/static/%s' % library.name
Let’s register the plugin:
>>> register_plugin(NewPlugin())
Rendering the inclusions now will result in the HTML fragments we need to include on the top of our page (just under the <head> tag for instance):
>>> print needed.render() <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script>
Let’s set this a currently needed inclusions:
>>> request.needed = needed
There is a function available as well for rendering the resources for the currently needed inclusion:
>>> from hurry import resource >>> print resource.render() <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script>
Inserting resources in HTML
When you have the HTML it can be convenient to have a way to insert resources directly into some HTML.
The insertion system assumes a HTML text that has a <head> tag in it:
>>> html = "<html><head>something more</head></html>"
To insert the resources directly in HTML we can use render_into_html on needed:
>>> print needed.render_into_html(html) <html><head> <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script> something more</head></html>
The top-level convenience function does this for the currently needed resources:
>>> print resource.render_into_html(html) <html><head> <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script> something more</head></html>
See below for a way to insert into HTML when bottom fragments are involved.
Top and bottom fragments
It’s also possible to render the resource inclusions into two fragments, some to be included just after the <head> tag, but some to be included at the very bottom of the HTML page, just before the </body> tag. This is useful as it can speed up page load times.
Let’s look at the same resources, now rendered separately into top and bottom fragments:
>>> top, bottom = needed.render_topbottom() >>> print top <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script> >>> print bottom <BLANKLINE>
There is effectively no change; all the resources are still on the top. We can enable bottom rendering by calling the bottom method before we render:
>>> needed.bottom()
Since none of the resources indicated it was safe to render them at the bottom, even this explicit call will not result in any changes:
>>> top, bottom = needed.render_topbottom() >>> print top <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script> >>> print bottom <BLANKLINE>
bottom(force=True) will however force all javascript inclusions to be rendered in the bottom fragment:
>>> needed.bottom(force=True) >>> top, bottom = needed.render_topbottom() >>> print top <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> >>> print bottom <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script>
Let’s now introduce a javascript resource that says it is safe to be included on the bottom:
>>> y2 = ResourceInclusion(foo, 'y2.js', bottom=True)
When we start over without bottom enabled, we get this resource show up in the top fragment after all:
>>> needed = NeededInclusions() >>> needed.need(y1) >>> needed.need(y2) >>> top, bottom = needed.render_topbottom() >>> print top <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script> <script type="text/javascript" src="http://localhost/static/foo/y2.js"></script> >>> print bottom <BLANKLINE>
We now tell the system that it’s safe to render inclusions at the bottom:
>>> needed.bottom()
We now see the resource y2 show up in the bottom fragment:
>>> top, bottom = needed.render_topbottom() >>> print top <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script> >>> print bottom <script type="text/javascript" src="http://localhost/static/foo/y2.js"></script>
There’s also a convenience function for the currently needed inclusion:
>>> request.needed = needed >>> top, bottom = resource.render_topbottom() >>> print top <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script> >>> print bottom <script type="text/javascript" src="http://localhost/static/foo/y2.js"></script>
When we force bottom rendering of Javascript, there is no effect of making a resource bottom-safe: all .js resources will be rendered at the bottom anyway:
>>> needed.bottom(force=True) >>> top, bottom = needed.render_topbottom() >>> print top <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> >>> print bottom <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script> <script type="text/javascript" src="http://localhost/static/foo/y2.js"></script>
Note that if bottom is enabled, it makes no sense to have a resource inclusion b that depends on a resource inclusion a where a is bottom-safe and b, that depends on it, is not bottom-safe. In this case a would be included on the page at the bottom after b in the <head> section, and this might lead to ordering problems. Likewise a rollup resource shouldn’t combine resources where some are bottom-safe and others aren’t.
The system makes no sanity checks for misconfiguration of bottom-safety however; it could be the user simply never enables bottom mode at all and doesn’t care about this issue. In this case the user will want to write Javascript code that isn’t safe to be included at the bottom of the page and still be able to depend on Javascript code that is.
Inserting top and bottom resources in HTML
You can also insert top and bottom fragments into HTML. This assumes a HTML text that has a <head> tag in it as well as a </body> tag:
>>> html = "<html><head>rest of head</head><body>rest of body</body></html>"
To insert the resources directly in HTML we can use render_topbottom_into_html on needed:
>>> print needed.render_topbottom_into_html(html) <html><head> <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> rest of head</head><body>rest of body<script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script> <script type="text/javascript" src="http://localhost/static/foo/y2.js"></script></body></html>
There’s also a function available to do this for the currently needed resources:
>>> print resource.render_topbottom_into_html(html) <html><head> <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> rest of head</head><body>rest of body<script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script> <script type="text/javascript" src="http://localhost/static/foo/y2.js"></script></body></html>
Using WSGI middleware to insert into HTML
There is also a WSGI middleware available to insert the top (and bottom) into the HTML. We are using WebOb to create a response object that will serve as our WSGI application.
We create a simple WSGI application. In our application we declare that we need a resource (y1) and put that in the WSGI environ under the key hurry.resource.needed:
>>> def app(environ, start_response): ... start_response('200 OK', []) ... needed = environ['hurry.resource.needed'] = NeededInclusions() ... needed.need(y1) ... return ['<html><head></head><body</body></html>']
We now wrap this in our middleware, so that the middleware is activated:
>>> from hurry.resource.wsgi import Middleware >>> wrapped_app = Middleware(app)
Now we make a request (using webob for convenience):
>>> import webob >>> req = webob.Request.blank('/') >>> res = req.get_response(wrapped_app)
We can now see that the resources are added to the HTML by the middleware:
>>> print res.body <html><head> <link rel="stylesheet" type="text/css" href="http://localhost/static/foo/b.css" /> <script type="text/javascript" src="http://localhost/static/foo/a.js"></script> <script type="text/javascript" src="http://localhost/static/foo/c.js"></script> </head><body</body></html>
When we set the response Content-Type to non-HTML, the middleware won’t be active even if we need things and the body appears to contain HTML:
>>> def app(environ, start_response): ... start_response('200 OK', [('Content-Type', 'text/plain')]) ... needed = environ['hurry.resource.needed'] = NeededInclusions() ... needed.need(y1) ... return ['<html><head></head><body</body></html>'] >>> wrapped_app = Middleware(app) >>> req = webob.Request.blank('/') >>> res = req.get_response(wrapped_app) >>> res.body '<html><head></head><body</body></html>'
bottom convenience
Like for need and mode, there is also a convenience spelling for bottom:
>>> request = Request() >>> l1 = ResourceInclusion(foo, 'l1.js', bottom=True) >>> l1.need()
Let’s look at the resources needed by default:
>>> c = get_current_needed_inclusions() >>> top, bottom = c.render_topbottom() >>> print top <script type="text/javascript" src="http://localhost/static/foo/l1.js"></script> >>> print bottom <BLANKLINE>
Let’s now change the bottom mode using the convenience hurry.resource.bottom spelling:
>>> from hurry.resource import bottom >>> bottom()
Re-rendering will show it’s honoring the bottom setting:
>>> top, bottom = c.render_topbottom() >>> print top <BLANKLINE> >>> print bottom <script type="text/javascript" src="http://localhost/static/foo/l1.js"></script>
Generating resource code
Sometimes it is useful to generate code that expresses a complex resource dependency structure. One example of that is in hurry.yui. We can use the generate_code function to render resource inclusions:
>>> i1 = ResourceInclusion(foo, 'i1.js') >>> i2 = ResourceInclusion(foo, 'i2.js', depends=[i1]) >>> i3 = ResourceInclusion(foo, 'i3.js', depends=[i2]) >>> i4 = ResourceInclusion(foo, 'i4.js', depends=[i1]) >>> i5 = ResourceInclusion(foo, 'i5.js', depends=[i4, i3]) >>> from hurry.resource import generate_code >>> print generate_code(i1=i1, i2=i2, i3=i3, i4=i4, i5=i5) from hurry.resource import Library, ResourceInclusion <BLANKLINE> foo = Library('foo', 'dummy') <BLANKLINE> i1 = ResourceInclusion(foo, 'i1.js') i2 = ResourceInclusion(foo, 'i2.js', depends=[i1]) i3 = ResourceInclusion(foo, 'i3.js', depends=[i2]) i4 = ResourceInclusion(foo, 'i4.js', depends=[i1]) i5 = ResourceInclusion(foo, 'i5.js', depends=[i4, i3])
Let’s look at a more complicated example with modes and superseders:
>>> j1 = ResourceInclusion(foo, 'j1.js', debug='j1-debug.js') >>> j2 = ResourceInclusion(foo, 'j2.js', debug='j2-debug.js') >>> giantj = ResourceInclusion(foo, 'giantj.js', supersedes=[j1, j2], ... debug='giantj-debug.js') >>> print generate_code(j1=j1, j2=j2, giantj=giantj) from hurry.resource import Library, ResourceInclusion <BLANKLINE> foo = Library('foo', 'dummy') <BLANKLINE> j1 = ResourceInclusion(foo, 'j1.js', debug='j1-debug.js') j2 = ResourceInclusion(foo, 'j2.js', debug='j2-debug.js') giantj = ResourceInclusion(foo, 'giantj.js', supersedes=[j1, j2], debug='giantj-debug.js')
We can control the name the inclusion will get in the source code by using keyword parameters:
>>> print generate_code(hoi=i1) from hurry.resource import Library, ResourceInclusion <BLANKLINE> foo = Library('foo', 'dummy') <BLANKLINE> hoi = ResourceInclusion(foo, 'i1.js') >>> print generate_code(hoi=i1, i2=i2) from hurry.resource import Library, ResourceInclusion <BLANKLINE> foo = Library('foo', 'dummy') <BLANKLINE> hoi = ResourceInclusion(foo, 'i1.js') i2 = ResourceInclusion(foo, 'i2.js', depends=[hoi])
Sorting inclusions by dependency
This is more a footnote than something that you should be concerned about. In case assumptions in this library are wrong or there are other reasons you would like to sort resource inclusions that come in some arbitrary order into one where the dependency relation makes sense, you can use sort_inclusions_topological:
>>> from hurry.resource import sort_inclusions_topological
Let’s make a list of resource inclusions not sorted by dependency:
>>> i = [a5, a3, a1, a2, a4] >>> sort_inclusions_topological(i) #doctest: +NORMALIZE_WHITESPACE [<ResourceInclusion 'a1.js' in library 'foo'>, <ResourceInclusion 'a4.js' in library 'foo'>, <ResourceInclusion 'a2.js' in library 'foo'>, <ResourceInclusion 'a3.js' in library 'foo'>, <ResourceInclusion 'a5.js' in library 'foo'>]
Inclusion renderers
The HTML fragments for inclusions are rendered by inclusion renderers that are simple functions registered per extension.
Renderers are registered in the inclusion_renderers dictionary:
>>> from hurry.resource.core import inclusion_renderers >>> sorted(inclusion_renderers) ['.css', '.js', '.kss']
Renderers render HTML fragments using given resource URL:
>>> inclusion_renderers['.js']('http://localhost/script.js') '<script type="text/javascript" src="http://localhost/script.js"></script>'
Let’s create an inclusion of unknown resource:
>>> a6 = ResourceInclusion(foo, 'nothing.unknown')>>> from hurry.resource.core import render_inclusions >>> render_inclusions([a6]) #doctest: +NORMALIZE_WHITESPACE +ELLIPSIS Traceback (most recent call last): ... UnknownResourceExtension: Unknown resource extension .unknown for resource inclusion: <ResourceInclusion 'nothing.unknown' in library 'foo'>
Now let’s add a renderer for our “.unknown” extension and try again:
>>> def render_unknown(url): ... return '<link rel="unknown" href="%s" />' % url >>> inclusion_renderers['.unknown'] = render_unknown>>> render_inclusions([a6]) '<link rel="unknown" href="http://localhost/static/foo/nothing.unknown" />'
CHANGES
0.10 (2010-07-24)
The WSGI Middleware depends on WebOb, which is an optional dependency. Don’t expose it into the __init__.py therefore, as that needs to be independent of WebOb. To use the middleware import it directly from hurry.resource.wsgi.
The Library factory now has to be called with a rootpath argument as the second argument. This is a path to the actual resource directory that the library represents, relative to the directory of the package. This is needed to make `` hurry.resource`` useful for frameworks not based on zope.configuration.
This breaks backwards compatibility with previous versions of hurry.resource; all code that uses Library should be adjusted. Please also update their setup.py to depend on hurry.resource >= 0.10.
Packages should register their Library instances with hurry.resource using the hurry.resource.libraries entry point.
0.9 (2010-07-13)
WSGI support: hurry.resource.Middleware can be used to wrap WSGI applications. If the application supplies a NeededInclusions object in environ with the key hurry.resource.needed, the middleware will pick up on this and insert the needed inclusions.
The WebOb library is needed to make this work and depending on hurry.resource [wsgi] will pull in the required dependency.
Fixed some typos in README.txt.
0.4.1 (2009-12-16)
Fix restructured text.
0.4 (2009-12-16)
Expose a GroupInclusion that only exists to depend on other inclusions. This way it becomes possible to expose a set of different resources together without there being a central resource that depends on them. The GroupInclusion itself is therefore not rendered.
Create render_into_html and render_topbottom_into_html functionality, which insert the rendered inclusions into HTML.
Expose render, render_into_html, render_topbottom and render_topbottom_into_html as top-level functions in hurry.resource.
Fix NameError when trying to render an inclusion with unknown extension.
Document inclusion renderers mechanism in the README.txt.
0.3 (2008-10-15)
Features added
Consolidating resources into rollups is now disabled by default. This is to help developers with debugging; rollups are typically minified for optimization purposes, and libraries such as YUI do not offer rollups in non-minified form. Since rollups must now be explicitly enabled by the application developer (hurry.resource.rollup()), an application developer who needs to debug can choose not to call it (or call hurry.resource.rollup(disable=True).
Added an optimization section to the documentation.
Added some more details about how modes are expected to be used to the documentation.
Bug fixes
hurry.resource.bottom now takes a second optional disable parameter.
0.2 (2008-10-13)
Changed the API to set the mode. Instead of passing it to render and inclusions, the requested mode can be set with the mode method on INeededInclusions. For convenience there is a mode function as well that can be imported directly from hurry.resource that sets the mode for the current needed inclusions.
Added support for rendering resources into two fragments, one to be included at the top of the HTML page in the <head> section, the other to be included just before the </body> section. In some circumstances doing this can speed up page load time.
0.1 (2008-10-07)
Initial public release.
Download
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.