README.md

Spend 5 literal minutes thinking:

• How much time do you have to devote to this exercise?
• How can you verify that your solution is correct?

Then:

1. Fork this gist and clone it locally.
2. Send tsnowhite@taximagic.com and aharris@taximagic.com a link to your forked gist with "[gh_repo_stats]" in the subject. Provide an estimate of how long you'll be working on your solution in man-hours and an estimate of when you will be finished. Feel free to ask any questions about the assignment, as well.
3. Complete the exercise, pushing your changes to the gist repo as you go.
4. Send tsnowhite@taximagic.com and aharris@taximagic.com a link to your forked gist, along with the actual amount of time you spent on the problem, and a discussion of your solution. If you choose to use external dependencies besides ruby, please provide a Rakefile which allows one to run rake install to install the dependencies.

Gemfile

source :rubygems

gem 'yajl-ruby'

Gemfile.lock

GEM
  remote: http://rubygems.org/
  specs:
    yajl-ruby (1.1.0)

PLATFORMS
  ruby

DEPENDENCIES
  yajl-ruby

gh_repo_stats

#!/usr/bin/env ruby

require 'bundler/setup'
require 'yajl'
require 'zlib'
require 'uri'
require 'time'
require 'open-uri'
require 'optparse'

class DateRange
  include Enumerable

  attr_reader :from, :to
  def initialize(from, to)
    @from, @to = from, to
  end

  def hour_blocks
    return [hour_block(from)] if seconds_between == 0
    @hour_blocks ||= (seconds_between/3600).ceil.times.map do |i|
      hour_block(from + (i * 3600))
    end
  end

  def seconds_between
    to - from
  end

  def hour_block(time)
    time.strftime("%Y-%m-%d-") + time.hour.to_s
  end

  def each
    hour_blocks.each {|block| yield block }
  end

end

module Github
  module Events
    class Event
      attr_reader :attrs
      def initialize(attrs)
        @attrs = attrs
      end

      def event_name
        attrs['type']
      end

      def created_at
        Time.parse(attrs['created_at'])
      end
    end

    module Repository
      def repository
        attrs['repository']
      end

      def owner_name
        repository['owner']
      end

      def repo_name
        repository['name']
      end

      def key
        "#{owner_name}/#{repo_name}"
      end
    end

    class PushEvent < Event
      include Repository
    end

    class PullRequestEvent < Event
      include Repository
    end

    # Had I time, or were so inclined, I would define an Event subclass for each
    # type at http://developer.github.com/v3/activity/events/types that responds
    # to #key.
    # Why are we doing this? Because, regardless of what Github claims, the JSON
    # schema for each event type is not all that similar. Since we only ever filter
    # by one event type, it's reasonable to expect different result types.

    class Index < Hash
      def <<(event)
        if self[event.key]
          self[event.key][:count] += 1
          self[event.key][:events] << event
        else
          self[event.key] = {
            :count => 1,
            :events => [event]
          }
        end
      end

      def sort
        to_a.sort { |a,b| a[1][:count] <=> b[1][:count] }
      end
    end
  end
end

class Github::Stats
  # WAT - http://data.githubarchive.org expects keys in Mountain time zone,
  # and returns events in it as well.
  GITHUB_ARCHIVE_TOTALLY_ARBITRARY_TIMEZONE_WAT = "-07:00"

  attr_reader :options, :errors
  def initialize(options)
    @options = options
    @errors = []
  end

  def to
    Time.parse(options[:to]).localtime(GITHUB_ARCHIVE_TOTALLY_ARBITRARY_TIMEZONE_WAT)
  end

  def from
    Time.parse(options[:from]).localtime(GITHUB_ARCHIVE_TOTALLY_ARBITRARY_TIMEZONE_WAT)
  end

  def limit
    options[:limit].to_i
  end

  def base_uri
    URI.parse("http://data.githubarchive.org/")
  end

  def date_range
    DateRange.new(from, to)
  end

  def event_index
    @event_index ||= Github::Events::Index.new
  end

  def gather
    date_range.each do |block|
      uri = base_uri
      uri.path = "/#{block}.json.gz"
      gzip = Zlib::GzipReader.new(open(uri)).read
      Yajl::Parser.parse(gzip) do |hash|
        if hash['type'] == options[:event_type]
          event = event_class.new(hash)
          event_index << event if (from <= event.created_at && event.created_at < to)
        end
      end
    end
  end

  def report
    event_index.sort.reverse.take(limit).each do |event|
      reporter.puts "#{event[0].to_s} - #{event[1][:count]} events"
    end
  end

  attr_writer :reporter
  def reporter
    @reporter ||= STDOUT
  end

  # Let's just say every input is valid, and if the program blows up, it's user error.
  def valid?
    true
  end

  def event_class
    @event_class ||= Github::Events::const_get(options[:event_type])
  end
end

options = {}
OptionParser.new do |opts|
  opts.banner = "Usage gh_repo_stats --after 2012-11-01T13:00:00Z --before 2012-11-02T03:12:14-03:00 --event PushEvent --count 42"
  opts.on("--after DATE", "Query dates on or after DATE, an ISO-8601 formatted date string") do |a|
    options[:from] = a
  end

  opts.on("--before DATE", "Query dates before DATE, an ISO-8601 formatted date string") do |b|
    options[:to] = b
  end

  opts.on("--event TYPE", "Filter event type; see http://developer.github.com/v3/activity/events/types/#gistevent for details") do |e|
    options[:event_type] = e
  end

  opts.on("--count N", "Report on the top (n) results") do |c|
    options[:limit] = c
  end
end.parse!

stats = Github::Stats.new(options)

if stats.valid?
  stats.gather
  stats.report
else
  stats.errors.each do |error|
    puts error
  end
end

problem.md

Create a ruby script utilizing http://www.githubarchive.org/ which has the following options and output:

   > ./gh_repo_stats -h
   gh_repo_stats [--after DATETIME] [--before DATETIME] [--event EVENT_NAME] [-n COUNT]

Here's some example interactions:

   > ./gh_repo_stats --after 2012-10-12T10:00:00-08:00 --before 2012-10-12T11:00:00-08:00 \
                     --event PushEvent -n 20
   hughht5/Guido - 30 events
   gordonbrander/French-Toast-Assets - 19 events
   iblanky/iblanky.github.com - 18 events
   honielui/hackathon_hnb - 17 events
   MilkZoft/codejobs - 16 events
   danberindei/infinispan - 12 events
   gilgomesp/site - 11 events
   HKCodeCamp/bartr - 11 events
   josephwilk/tlearn-rb - 10 events
   soapboxCommunications/Young-Final - 10 events
   sakai-mirror/melete - 10 events
   fabiantheblind/auto-typo-adbe-id - 9 events
   demobox/jclouds-maven-site-1.5.2 - 9 events
   davecoa/opendataday_workshop - 9 events
   freebsd/freebsd-ports - 9 events
   joonasrouhiainen/studio4-election - 9 events
   bunuelo/funk2 - 9 events
   Certainist/To_aru_Library - 9 events
   lanticezdd/uni - 8 events
   pfinette/finettedotcom - 8 events
   >
   > ./gh_repo_stats --after 2013-01-12T10:00:01Z --before 2013-01-12T11:00:00Z \
                     --event WatchEvent -n 2
   airbnb/javascript - 19 events
   piranha/gostatic - 4 events