README.md in dripdrop-0.0.3 vs README.md in dripdrop-0.1.0
- old
+ new
@@ -1,115 +1,34 @@
-# dripdrop
+# DripDrop
-0MQ Based App Event Monitoring / processing.
-A work in progress.
+**Note: For now, for PUSH/PULL support you'll need to build zmqmachine from my fork of zmqmachine, the official gem does not yet have this**
-# Why use dripdrop?
+DripDrop is ZeroMQ(using zmqmachine) + Event Machine simplified for the general use case + serialization helpers.
-You want to asynchronously process / monitor arbitrary messages from your app.
-dripdrop does this well for a few reasons.
+Here's an example of the kind of thing DripDrop makes easy, from [examples/pubsub.rb](http://github.com/andrewvc/dripdrop/blob/master/example/pubsub.rb)
+
+ DripDrop::Node.new do |node|
+ z_addr = 'tcp://127.0.0.1:2200'
-* It's fast. dripdrop doesn't slow down your app. 0MQ + Bert are fast. Sending a message never blocks.
-* It's flexible. By leveraging 0MQ pub/sub sockets you can have many different processors (collectors in dripdrop) that don't impact or even care about each other
-* It's easy. Check out the agent and collector examples below. You can be processing stuff in no time.
-
-## An example with a WebSocket UI:
-
-### You'll need to have the zmq dev libs on your machine. On OSX this means
-
-1. Download and build zeromq from [zeromq.org](http://www.zeromq.org/area:download)
-1. The agent just uses the plain zmq gem, which runs fine on ruby 1.8.7+, *EDIT!* If you use my fork of ffi-rzmq everything should work with 1.8.7! This isn't fully tested yet though *END EDIT* this is so you can use it in say your rails app. Everything else needs ruby 1.9.2 or jruby and uses Chuck Remes [ffi-rzmq](http://github.com/chuckremes/ffi-rzmq), and [zmqmachine](http://github.com/chuckremes/zmqmachine) gems which you must build yourself. I recommend using rvm to enable the use of multiple rubies on one machine.
-1. zmq_forwarder comes with zmq, use this to aggregate agent messages using the example config shown below
-
-### To run a simple example, feeding data to a websockets UI
-
-#### Aggregate agents with zmq_forwarder (comes with zmq)
- $ zmq_forwarder examples/forwarder.cfg
-
-#### Start up the drip drop publisher example
- $ drip-publisher
-
-#### Assuming you have mongodb running
- $ drip-mlogger
-
-#### Start up a webserver to host the HTML/JS for a sample websocket client
- $ cd DRIPDROPFOLDER/example/web/
- $ ruby server
-
-## Example Topology
-
-You can add as many listeners as you want, or reconfigure things any way you want. Heres how I plan on using it.
-
-
-
-## Sending Messages
-
-Sending messages is easy with the agent, an example:
-
- require 'rubygems'
- require 'dripdrop/agent'
-
- agent = DripDrop::Agent.new('tcp://127.0.0.1:2900')
-
- loop do
- #Test is the message name, this is the first part of the 0MQ message, used for filtering
- #at the 0MQ sub socket level, :head is always a hash, :body is freeform
- #EVERYTHING must be serializable to BERT
- agent.send_message('test', :body => 'hello', :head => {:key => 'value'})
- puts "SEND"
- sleep 1
- end
-
-## Writing a custom message processor
-
-Writing custom message processors is super easy, just create a new DripDrop::Collector
-and run it. DripDrop::Collector is based on Chuck Remes' awesome zmqmachine, an evented
-0MQ processor. Heres' the MongoDB logger as an example:
-
- require 'rubygems'
- require 'mongo'
- require 'dripdrop/collector'
-
- class DripDrop
- class MLoggerCollector < Collector
- attr_accessor :mongo_collection
-
- #Messages are a DripDrop::Message
- def on_recv(message)
- if @mongo_collection
- @mongo_collection.insert(message.to_hash)
- end
- end
+ pub = node.zmq_publish(z_addr,:bind)
+ sub = node.zmq_subscribe(z_addr,:connect).on_recv do |message|
+ puts "Receiver 1 #{message.inspect}"
end
+ sub = node.zmq_subscribe(z_addr, :connect).on_recv do |message|
+ puts "Receiver 2 #{message.inspect}"
+ end
- class MLogger
- attr_reader :sub_address, :sub_reactor, :mongo_host, :mongo_port, :mongo_db,
- :mongo_connection, :mongo_collection
-
- def initialize(sub_address='tcp://127.0.0.1:2901',mhost='127.0.0.1',mport=27017,mdb='dripdrop')
- @sub_address = URI.parse(sub_address)
- @sub_collector = MLoggerCollector.new('tcp://127.0.0.1:2901')
-
- @mongo_host, @mongo_port, @mongo_db = mhost, mport, mdb
- @mongo_connection = Mongo::Connection.new(@mongo_host,@mongo_port).db(@mongo_db)
- @mongo_collection = @mongo_connection.collection('raw')
- end
-
- def run
- @sub_collector.mongo_collection = @mongo_collection
- @sub_collector.run.join
- end
+ node.zm_reactor.periodical_timer(5) do
+ pub.send_message(DripDrop::Message.new('test', :body => 'Test Payload'))
end
end
+
+Want to see a longer example encapsulating both zmqmachine and eventmachine functionality? Check out [this file](http://github.com/andrewvc/dripdrop-webstats/blob/master/lib/dripdrop-webstats.rb), which encapsulates all the functionality of the diagram below:
-## Note on Patches/Pull Requests
-
-* Fork the project.
-* Make your feature addition or bug fix.
-* Commit, do not mess with rakefile, version, or history.
- (if you want to have your own version, that is fine but bump version in a commit by itself I can ignore when I pull)
-* Send me a pull request. Bonus points for topic branches.
+
-## Copyright
+#How It Works
+
+DripDrop encapsulates both zmqmachine, and eventmachine. It provides some sane default messaging choices, using [BERT](http://github.com/blog/531-introducing-bert-and-bert-rpc) (A binary, JSON, like serialization format) and JSON to automatically. zmqmachine and eventmachine have some good APIs, some convoluted ones, the goal here is to smooth over the bumps, and make writing highly concurrent programs both as terse and beautiful as possible.
Copyright (c) 2010 Andrew Cholakian. See LICENSE for details.