This plugin adds set filtering using bloom filters to the Pacer graph and streaming data processing library.
This plugin is also meant to serve as an example of how easy it is to build a plugin for Pacer.
The bloomfilter method is added to the core route object in Pacer, which means that it will be available to all routes. The method takes 2 arguments and a block:
- false_positive_probability: between 0 and 1 with a lower number indicating a lower chance of different keys being considered equal
- expected_count: the maximum number of elements you think will be added to the bloom filter. The more accurate this number is the more accurate your false_positive_probability will be.
- block: The block should map the elements that will be iterated over to a value that will be used by the filter. You should return a string. This block does not affect the actual output of the route. If no block is given, to_s on the element itself will used for the filter.
Map the vertices to names and then filter by name:
graph.v[:name].bloomfilter(0.001, 10).except(['sam', 'bob'])
"steve" "gary"
Total: 2
=> #<GraphV -> Obj(name) -> Obj-Bloom>
Wrong! There is no way to map the vertices to the name, all vertices pass through:
graph.v.bloomfilter(0.001, 10).except(['sam', 'bob'])
#<V[0]> #<V[1]> #<V[2]> #<V[3]>
Total: 4
=> #<GraphV -> V-Bloom>
That's better. Here we tell the bloomfilter how to map the vertices to the name field (we switched it to #only though just to get more mileage out of the example):
graph.v.bloomfilter(0.001, 10) { |v| v[:name] }.only(['sam', 'bob'])
#<V[0]> #<V[3]>
Total: 2
=> #<GraphV -> V-Bloom>
And for completeness, the uniq method is pretty self explanitory I hope:
graph.v[:type].bloomfilter(0.001, 10).uniq
"band member"
Total: 1
=> #<GraphV -> V-Bloom>
I don't know. This plugin is currently only a proof of concept and has not been optimized, profiled or benchmarked! If you want to spend a couple of hours to make it blazing fast, I'll be quite impressed!
The plugin architecture of Pacer is very simple. You can define a module in one of 3 namespaces which correspond to the categories of functions that I've identified thus far:
- Pacer::Filter
- Pacer::Transform
- Pacer::SideEffect
Rather than try to explain those clearly, it is probably best to dig into the code and documenation of Pacer, Pipes and Gremlin.
The module that you define will be mixed in to a Pacer::Route instance at runtime when chain_route is called pointing to your module. Nearly every method that is called when a user is building a route is actually just a friendlier version of chain_route. The chain_route method takes a hash of arguments, a few of which are reserved, and the rest of which will be applied to your module via property setters. Note that in this plugin, :false_pos_prob, :expected_count, :bloomfilter, and :block all have corresponding attr_accessors in the BloomFilter module. In that way, settings are carried over from the route definition to the route instance.
Once the route is defined, it may be executed multiple times. Whenever it is executed, a new pipeline is built, of which your module is just one part. In almost all cases, all you will need to do is define the protected attach_pipe method. Inside that method you just need to create your pipe, call setStarts on it with the pipe that was given as a parameter, and return the pipe you created. Pacer will look after the rest of the pipe building process.
- Check out the latest master to make sure the feature hasn't been implemented or the bug hasn't been fixed yet
- Check out the issue tracker to make sure someone already hasn't requested it and/or contributed it
- Fork the project
- Start a feature/bugfix branch
- Commit and push until you are happy with your contribution
- Make sure to add tests for it. This is important so I don't break it in a future version unintentionally.
- Please try not to mess with the Rakefile, version, or history. If you want to have your own version, or is otherwise necessary, that is fine, but please isolate to its own commit so I can cherry-pick around it.
Copyright (c) 2011 Darrick Wiebe. See LICENSE.txt for further details.