gh_archive_challenge.md

GitHub Archive leaderboard challenge

GitHub Archive provides stats for projects on GitHub via a RESTful HTTP service.

Create a command-line program that lists the most active repositories for a given time range. It should support the following interface:

  gh_repo_stats [--after DATETIME] [--before DATETIME] [--event EVENT_NAME] [-n COUNT]

gh_repo_stats --after 2012-11-01T13:00:00Z --before 2012-11-02T03:12:14-03:00 --event PushEvent --count 42

sakai-mirror/melete - 168 events
runningforworldpeace/feeds - 103 events
chapuni/llvm-project-submodule - 98 events
chapuni/llvm-project - 98 events
Frameset91/untitled0815 - 94 events
josmera01/juanrueda-internation - 78 events
artmig/artmig.github.com - 76 events
mozilla/mozilla-central - 69 events
bcomdlc/bcom-homepage-archive - 68 events
sakai-mirror/ambrosia - 68 events
sakai-mirror/test-center - 65 events
sakai-mirror/mneme - 65 events
klange/tales-of-darwinia - 63 events
esc/bottlepaste - 62 events
ehsan/mozilla-history-tools - 58 events
all4senses/Gv - 54 events
herry13/nuri - 51 events
SeqAlignViz/seqalignviz.github.com - 44 events
illcreative/Get-Home-Safe-NYC - 44 events
incxnt/incxnt.github.com - 41 events
daknok/Filmpje - 40 events
ChildOfWar/MIU - 40 events
aleontiev/Turntable.FM-Squared - 40 events
DigiZeit/dwa - 37 events
DigiZeit/dwa-debug - 37 events
danielcooper/radarsite - 37 events
navanjr/kts - 36 events
githubtrainer/poems - 35 events
wikimedia/mediawiki-extensions - 35 events
DigiZeit/dwa-pro - 35 events
gerritwm/MediaWiki - 34 events
cloudweekhec/stack - 34 events
andyisimprovised/goatmachine - 34 events
GreenplumChorus/chorus - 33 events
ros-gbp/nodelet_core-release - 33 events
honovation/veil - 33 events
McGill-CSB/PHYLO - 32 events
ceph/ceph - 31 events
aaronlbloom/dwa - 31 events
dotCMS/dotCMS - 31 events
RC5Group6/research-camp-5 - 30 events
ShreyaPandita/oftest - 30 events

Going further

• There are 18 published Event Types. How would you manage them? What would you do if GitHub added more Event Types?
• What factors impact performance? What would you do to improve them?
• The example shows one type of output report. How would you add additional reporting formats?
• If you had to implement this using only one gem, which would it be? Why?

github_archive.rb

class Archiver

  require 'google/api_client'
  require 'google/api_client/client_secrets'
  require 'google/api_client/auth/installed_app'
  require 'bigquery'
  require 'json'
  
  def initialize
    print_line
    puts "Welcome to the Github Archive Leaderboard"
    puts "Find the most active repositories for a given time range"
    puts "Type in the following (example):" 
    puts "2014-03-16 13:00:00,2014-03-18 15:12:14,PushEvent,10"
    print_line
    parse_input
  end

  def parse_input
    print_prompt
    input = gets.chomp.split(',')
    @after = input[0]
    @before = input[1]
    @event_name = input[2]
    @count = input[3]
    get_data
  end

  def get_data
    opts = {}
    opts['client_id']     = '992331757661-9t3f55b02btia14okppskbi35bgn91eb.apps.googleusercontent.com'
    opts['service_email'] = '992331757661-9t3f55b02btia14okppskbi35bgn91eb@developer.gserviceaccount.com'
    opts['key']           = '39ebfef9a32360852aa95a4b900bf9fc59303ebb-privatekey.p12'
    opts['project_id']    = '992331757661'
    bq = BigQuery.new(opts)
    @hash = bq.query(
      "SELECT repository_name, count(repository_name) as pushes, repository_description, repository_url
      FROM [githubarchive:github.timeline]
      WHERE type='#{@event_name}'
      AND repository_language='Ruby'
      AND PARSE_UTC_USEC(created_at) >= PARSE_UTC_USEC('#{@after}') 
      AND PARSE_UTC_USEC(created_at) < PARSE_UTC_USEC('#{@before}')
      GROUP BY repository_name, repository_description, repository_url
      ORDER BY pushes DESC
      LIMIT #{@count}"
      )
    output
  end

  def output
    repos = @hash['rows']
    repos.each do |repo|
      puts "#{repo['f'][3]['v'][19..-1]} - #{repo['f'][1]['v']} events"
    end
  end

  def print_prompt
    puts "Type here:"
  end

  def print_line
    puts "----------------------------------------"
  end
end

Archiver.new

sample_output.md

Run:

$ ruby github_archive.rb 

Prompt:

----------------------------------------
Welcome to the Github Archive Leaderboard
Find the most active repositories for a given time range
Type in the following (example):
2014-03-16 13:00:00,2014-03-18 15:12:14,PushEvent,10
----------------------------------------
Type here:

User types in:

2014-03-16 13:00:00,2014-03-18 15:12:14,PushEvent,10

Output:

W, [2014-03-18T23:33:37.405502 #20264]  WARN -- : Google::APIClient - Please provide :application_name and :application_version when initializing the client
lingohub/github-hooks-test - 119 events
Ortask/mutator - 113 events
CocoaPods/Specs - 109 events
Homebrew/homebrew - 85 events
BrewTestBot/homebrew - 61 events
YusukeAoki/webap_materials - 56 events
zunda/emoticommits - 45 events
moneyadviceservice/frontend - 43 events
avondohren/project-git - 41 events
opf/openproject - 41 events