Serve static files through streaming objects for Diet.js
$ npm install diet-static-stream
var diet = require("diet");
var dietStream = require("diet-static-stream");
var app = diet();
app.listen(80);
var static = dietStream({
path: "static"
});
app.footer(static);
##Documentation Diet-static-stream can take several options for it's usage
path [String] - Required
Path where diet-static-stream
will serve the files
index [String|Boolean] - Optional (Default: index)
Filename to look when there is no extension especified in the request
if value is Boolean, it will search for index file as default if true, otherwise if false, it will not do automatic search for an index file
#When value is boolean and it's true
GET: www.example.com/images -> Server/static/images/index.html
#When value is boolean and it's false
GET: www.example.com/images -> Server/static/images.html
#Otherwise if value is string, search for provided filename as index
#Value: main
GET: www.example.com/images -> Server/static/images/main.html
defaultExtension [String] - Optional (Default: html)
Extension to look when there is no extension on the request
GET: www.example.com/images.html -> Server/static/images.html
GET: www.example.com/images -> Server/static/images.html OR Server/static/images/index.html according to index property
cache [String|Function] - Optional (Default: max-age=3600)
Control-Cache value to set in header response, if function is provided then all requests are executed within this function passing $ signal, if funtion returns false then all necessary headers for cache will be excluded
stream({
cache: function($){
//Set cache only for index.html
if($.url.pathname == "index.html"){
return "max-age=36000";
}
//Exclude cache headers for all requests that are not for index.html
return false;
}
});
scriptName [Boolean] - Optional (Default: false)
Whether require extension file on the request, if true defaultExt
and index
options will be ignored
#Value: true
GET: www.example.com/images -> Server/static/images #404 not found
GET: www.example.com/images/user_placeholder.jpg -> Server/static/images/user_placeholder.jpg #200 Found
#Value: false
#Find according to index and defaultExt values
GET: www.example.com/images -> Server/static/images/index.html
hook [Object] - Optional
If there is a need for hooks on request, this is handy for setting custom headers or any other change necessary, this property is a map of events that's executed in certain process
-
request [Function] ($: Signal) - Optional Request hook is executed just when module starts, before the file is searched, the signal
$
is the request response passed from the middleware -
success [Function] ($: Signal, headers: Object, file: Object) - Optional Success hook is executed when the file is successfully opened, this function injects
$
signal along withheaders
as default headers set for the file found andfile
as object that contains the property used for the stream -
fail [Function] ($: Signal) - Optional Fail hook is executed when the requested file was not found, right before responding
404
, the signal$
is passed from the middleware -
cached [Function] ($: Signal) - Optional Cached hook is executed when the requested file was found but it was not modified since the last request, making the response to be
304
, the signal$
is passed from the middleware
As explained before, both diet-static
and diet-static-stream
achieves the same end for sending static data to the client, due to the time it takes to make the data available to the client, i have made 2 phases which are:
-
Resolve Time Resolve time is based on how long does the server take to make data available and start sending to client, this process was done by using
console.time()
right before the module starts loading the file,console.timeEnd()
is executed when the module starts sending. -
Response Time Response time is based on how long the client waits for the server to start recieving the requested data, this depends on latency and mostly
Resolve Time
on server, this process was achieved by using the comandcURL
The tests were done using fake files create from fallocate
on Linux
, the tests were done with this specs:
- OS: Linux/Debian Ubuntu 16.04 x64
- CPU: Intel Core I5 3.30GHz
- RAM: 4gb x 2 1600mhz
- HDD: 256gb 7200rpm
What is the difference between diet-static and this one?
The answer is simple, diet-static
uses fs.readFile
, and according to API docs of node.js
it reads the entire contents of a file, meaning that all the data from a file is temporary stored in the memory, this is nice if you only read few files with small sizes.
But what happens when you use it on a server that has hundreds of requests for media files, this could lead the server to overload and crash.
This package uses stream objects for sending chunks of data when they are being on read, this does not load data into the memory and we can avoid memory overload for hundreds of requests
##Test You can test the module by executing these commands:
$ npm test
if you have mocha globally installed:
$ mocha
##License MIT