tabby
v0.3.0
Published
create tabbed web apps with progressive enhancement
Downloads
38
Readme
tabby
create web sites that span multiple tabs of content
- render plain html first, always
- if pushState is available, load additional section content dynamically over xhr (progressive enhancement)
- append
.json
to the url of any page to get the json data that was used to render it - subscribe to live updates
This module only handles setting up the glue to pipe together the data feed with the rendering logic at routes for presentation.
example
server code:
var http = require('http');
var fs = require('fs');
var ecstatic = require('ecstatic')(__dirname + '/static');
var trumpet = require('trumpet');
var duplexer = require('duplexer');
var sub = require('level-sublevel');
var db = sub(require('level')(__dirname + '/test.db', { encoding: 'json' }));
db.batch(require('./data.json'));
var tabby = require('tabby')(function (route, params) {
var tr = trumpet();
return duplexer(
tr.createWriteStream('#content'),
fs.createReadStream(__dirname + '/static/index.html').pipe(tr)
);
});
tabby.add('/', {
render: function () {
return fs.createReadStream(__dirname + '/static/home.html');
}
});
tabby.add('/cats', {
data: require('./data/cat.js')(db),
render: require('./render/cat.js')
});
tabby.add('/cats/:name', {
data: require('./data/cat_full.js')(db),
render: require('./render/cat_full.js')
});
tabby.add('/owners', {
data: require('./data/owner.js')(db),
render: require('./render/owner.js')
});
tabby.add('/owners/:name', {
data: require('./data/owner_full.js')(db),
render: require('./render/owner_full.js')
});
var server = http.createServer(function (req, res) {
if (tabby.test(req)) {
tabby.handle(req, res);
}
else ecstatic(req, res);
});
server.listen(5000);
browser code:
var tabby = require('tabby')('#content');
var sock = require('shoe')('/sock');
Each of the tabby routes has a render function, which returns a stream html. If a data function is provided, the stream it returns is piped into the render stream.
Here's what the simplest kind of data function might look like using leveldb:
module.exports = function (db) {
return db.createReadStream({ start: 'cat-', end: 'cat-~' });
};
and a render function for this data using hyperspace could be:
var hyperspace = require('hyperspace');
var fs = require('fs');
var html = fs.readFileSync(__dirname + '/cat.html', 'utf8');
module.exports = function () {
return hyperspace(html, function (row) {
return {
'.name': row.value.name,
'.owner': row.value.owner
};
});
};
In addition to building each /route
as full html, /route.json
is built with
just the data:
$ curl http://localhost:5000/owners.json
{"name":"Ryan Jenkins","location":"Brooklyn","link":"/owners/ryan-jenkins"}
{"name":"Sarah West","location":"Kansas","link":"/owners/sarah-west"}
{"name":"Sir Edmund Theodore Heathcliff
IV","location":"Newcastle","link":"/owners/sir-edmund-theodore-heathcliff-iv"}
and /route.html
is built with just the html fragments generated directly from
the data:
$ curl http://localhost:5000/owners.html
<div class="owner summary">
<h2><a class="name" href="/owners/ryan-jenkins">Ryan Jenkins</a></h2>
<div>
location: <span class="location">Brooklyn</span>
</div>
</div>
<div class="owner summary">
<h2><a class="name" href="/owners/sarah-west">Sarah West</a></h2>
<div>
location: <span class="location">Kansas</span>
</div>
</div>
<div class="owner summary">
<h2><a class="name" href="/owners/sir-edmund-theodore-heathcliff-iv">Sir Edmund Theodore Heathcliff IV</a></h2>
<div>
location: <span class="location">Newcastle</span>
</div>
</div>
If a stream hasn't been configured, tabby uses these routes to load content over xhr.
server methods
var tabby = require('tabby')
var t = tabby(outerFn)
Create a new tabby router.
If given, outerFn(route, params)
should return a stream that transforms the
render stream values into the final content to be delivered to the request. You
can also specify the outer function on a per-route basis by setting
route.outer
in t.add()
below.
t.add(pattern, route)
Add a route
for paths matching pattern
.
The pipeline for requests for each of route.data(params)
,
route.render(params)
, and route.outer(params)
is:
data.pipe(render).pipe(outer)
Where only route.render(params)
is required and the other 2 are optional.
You can put whatever other properties you want onto the route object for use in other places in the code.
t.test(url)
Return true if any of the patterns match for the pathname string url
.
t.handle(req, res)
Render all the pages that match for t.test(req.url)
.
browser methods
var tabby = require('tabby')
var t = tabby(target)
Create a new tabby instance t
bound to the element or query selector target
.
var r = t.add(pattern, route)
Add a route r
for a pattern
and route
like on the server, but r
emits
'render'
and 'update'
events when pattern
matches.
server events
t.on('error', function (err, req, res) {})
When there is an error in a data stream, this event fires. By default a listener will catch errors:
function (err, req, res) {
res.statusCode = 500;
res.end(err + '\n');
}
If you handle the error yourself you will need to send a response.
browser events
t.on('show', function (href) {})
This event fires when a new page has been requested and on the initial page load that was rendered server-side.
t.on('render', function (renderStream, elem, href) {})
r.on('render', function (renderStream, elem, href) {})
This event fires when a new render is created and on the initial page load for rendering patterns that match.
t.on('update', function (elem) {})
r.on('update', function (elem) {})
When the page has been completely rendered, this event fires.
todo
- route.live for live-streaming results
install
With npm do:
npm install tabby
license
MIT