node-stream-parser
Generic interruptible "parser" mixin for Transform & Writable streams
This module offers the stream-parser
mixin, which provides an easy-to-use API
for parsing bytes from Writable
and/or Transform
stream instances. This module
is great for implementing streaming parsers for standardized file formats.
For Writable
streams, the parser takes control over the _write
callback
function. For Transform
streams, the parser controls the _transform
callback
function.
Installation
$ npm install stream-parser
Example
Let's create a quick Transform
stream subclass that utilizes the parser's
_bytes()
and _passthrough()
functions to parse a theoretical file format that
has an 8-byte header we want to parse, and then pass through the rest of the data.
var Parser = require('stream-parser');
var inherits = require('util').inherits;
var Transform = require('stream').Transform;
function MyParser () {
Transform.call(this);
this._bytes(8, this.onheader);
}
inherits(MyParser, Transform);
Parser(MyParser.prototype);
MyParser.prototype.onheader = function (buffer, output) {
var header = {};
header.type = buffer.readUInt32LE(0);
header.name = buffer.toString('utf8', 4);
this.emit('header', header);
this._passthrough(Infinity);
};
var parser = new MyParser();
parser.on('header', function (header) {
console.error('got "header"', header);
});
process.stdin.pipe(parser).pipe(process.stdout);
Here's an example of manually creating a Transform
stream and turning it into a
"pass through" stream equivalent to the one built into node core:
var Parser = require('stream-parser');
var Transform = require('stream').Transform;
var p = new Transform();
Parser(p);
p._passthrough(Infinity);
process.stdin.pipe(p).pipe(process.stdout);
See the test
directory for some more example code in the test cases.
A list of known concrete implementations is here (send pull requests for more!):
API
Parser()
The Parser
stream mixin works with either Writable
or Transform
stream
instances/subclasses. Provides a convenient generic "parsing" API:
_bytes(n, cb) - buffers "n" bytes and then calls "cb" with the "chunk"
_skipBytes(n, cb) - skips "n" bytes and then calls "cb" when done
If you extend a Transform
stream, then the _passthrough()
function is also
added:
_passthrough(n, cb) - passes through "n" bytes untouched and then calls "cb"
._bytes(n, cb)
Buffers n
bytes and then invokes cb
once that amount has been collected.
._skipBytes(n, cb)
Skips over the next n
bytes and then invokes cb
once that amount has been
discarded.
._passthrough(n, cb)
Passes through n
bytes to the readable side of this stream untouched,
then invokes cb
once that amount has been passed through. This function is only defined
when stream-parser is extending a Transform
stream.