github_archive_parser 0.1.0

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: dbbcc01faaf08baed69942f2c59ffe2fca262966
+  data.tar.gz: 4aaba6c19c5caa3e2a4b17fa331e2a0411cac792
+SHA512:
+  metadata.gz: f4e608379a73168cf4de92cd2ef11e62fb022f3f778afa43d57dd3508e26d9b188cc26114d86a5cac4b55c8638a856545a0e3ce2f525f17abdd19c633feb9cf3
+  data.tar.gz: 08d9faf62bdad2358c01d2f65440b8812fc91965ecc4c16f08af6fc98fb5dd4703914d43c9541b31d9b21ac7e4459b7b99ef1d7e7c2e4b2a9bf752eb3f0f9de9

data/bin/github_archive_parser

@@ -0,0 +1,9 @@
+#!/usr/bin/env ruby
+
+$:.unshift File.expand_path("../../lib", __FILE__)
+require 'github_archive_parser'
+
+github_archive_parser = GitHubArchiveParser::Processor.new
+ARGV.each do |url|
+  github_archive_parser.process(url)
+end

data/lib/github_archive_parser/XCust.rb

@@ -0,0 +1,10 @@
+module AwesomeApplication
+  class PrintCreateEvent
+    include GitHubArchiveParser::CreateEvent
+
+    def parse(event)
+      # The event is a Hashie::Mash object for easy (dot) access
+      puts "#{event.repository.owner}/#{event.repository.name}"
+    end
+  end
+end

data/lib/github_archive_parser/event_handler.rb

@@ -0,0 +1,13 @@
+module GitHubArchiveParser
+  module EventHandler
+    module ClassMethods
+      def descendants
+        @descendants ||= ObjectSpace.each_object(Class).select { |klass| klass < self }
+      end
+    end
+
+    def self.included(base)
+      base.extend(ClassMethods)
+    end
+  end
+end

data/lib/github_archive_parser/event_handlers/commit_comment_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module CommitCommentEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/create_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module CreateEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/delete_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module DeleteEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/deployment_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module DeploymentEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/deployment_status_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module DeploymentStatusEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/download_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module DownloadEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/follow_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module FollowEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/fork_apply_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module ForkApplyEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/fork_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module ForkEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/gist_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module GistEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/gollum_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module GollumEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/issue_comment_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module IssueCommentEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/issues_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module IssuesEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/member_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module MemberEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/public_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module PublicEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/pull_request_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module PullRequestEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/pull_request_review_comment_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module PullRequestReviewCommentEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/push_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module PushEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/release_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module ReleaseEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/status_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module StatusEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/team_add_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module TeamAddEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/event_handlers/watch_event.rb

@@ -0,0 +1,5 @@
+module GitHubArchiveParser
+  module WatchEvent
+    include EventHandler
+  end
+end

data/lib/github_archive_parser/initialize.rb

@@ -0,0 +1,9 @@
+require 'pathname'
+require 'ostruct'
+require 'optparse'
+require 'open-uri'
+require 'zlib'
+require 'yajl'
+require 'hashie'
+
+Dir.glob(File.dirname(__FILE__) + '/**/*.rb') { |file| require file }

data/lib/github_archive_parser/log.rb

@@ -0,0 +1,59 @@
+require 'logger'
+require 'singleton'
+
+module GitHubArchiveParser
+  class Log
+    include Singleton
+
+    attr_accessor :logger, :base_directory, :debugging
+
+    def initialize
+      @base_directory = File.expand_path("../..", __FILE__) + "/"
+      @debugging = false
+      @logger = Logger.new(STDOUT)
+      @logger.level = Logger::ERROR
+      @logger.formatter = proc do |sev, datetime, progname, msg|
+        "#{msg}\n"
+      end
+    end
+
+    def self.use_debug
+      instance.debugging = true
+      instance.logger.formatter = proc do |sev, datetime, progname, msg|
+        "#{sev} [#{progname}]: #{msg}\n"
+      end
+    end
+
+    # Determine the file, method, line number of the caller
+    def self.parse_caller(message)
+      if /^(?<file>.+?):(?<line>\d+)(?::in `(?<method>.*)')?/ =~ message
+        file = Regexp.last_match[:file]
+        line = Regexp.last_match[:line]
+        method = Regexp.last_match[:method]
+        "#{file.sub(instance.base_directory, "")}:#{line}"
+      end
+    end
+
+    def self.method_missing(method, *args, &blk)
+      if valid_method? method
+        instance.logger.progname = parse_caller(caller(1).first) if instance.debugging
+        instance.logger.send(method, *args, &blk)
+      else
+        super
+      end
+    end
+
+    def self.respond_to_missing?(method, include_all=false)
+      if valid_method? method
+        true
+      else
+        super
+      end
+    end
+
+    def self.valid_method?(method)
+      instance.logger.respond_to? method
+    end
+
+  end
+end

data/lib/github_archive_parser/version.rb

@@ -0,0 +1,3 @@
+module GitHubArchiveParser
+  VERSION = "0.1.0"
+end

data/lib/github_archive_parser.rb

@@ -0,0 +1,97 @@
+require 'github_archive_parser/initialize'
+
+module GitHubArchiveParser
+  class Processor
+    attr_reader :options
+
+    def initialize
+      @options = OpenStruct.new(
+        debug: false,
+      )
+      parse_options
+      determine_log_level
+
+      # Create the concrete handlers and store them for future use
+      create_event_handlers
+    end
+
+    def process(url)
+      Log.info "Processing #{url}"
+      if (!url.nil? && url.start_with?("http://data.githubarchive.org/"))
+        gz = open(url)
+        js = Zlib::GzipReader.new(gz).read
+
+        Yajl::Parser.parse(js) do |event|
+          event = Hashie::Mash.new(event)
+          event_class = class_from_string("GitHubArchiveParser::#{event.type}")
+          event_handler = @event_handlers[event_class]
+
+          event_handler.each { |handler|
+            if !handler.nil? && handler.respond_to?(:parse)
+              handler.parse(event)
+            end
+          }
+        end
+      else
+        Log.warn "URL[#{url}] does not belong to http://data.githubarchive.org/"
+      end
+    end
+
+    def parse_options
+      OptionParser.new do |opt|
+        opt.version = VERSION
+        opt.on "-d", "--debug", "Debug output (shows DEBUG level log statements)" do
+          options.debug = true
+        end
+        opt.on "-q", "--quite", "Hide all output (shows only UNKNOWN level log statements)" do
+          options.quite = true
+        end
+      end.parse!
+    end
+
+    private
+
+    def determine_log_level
+      if options.debug
+        Log.level = Logger::DEBUG
+        Log.use_debug
+      elsif options.quite
+        Log.level = Logger::UNKNOWN
+      else
+        Log.level = Logger::INFO
+      end
+    end
+
+    def class_from_string(string)
+      begin
+        string.split('::').inject(Object) do |mod, class_name|
+          mod.const_get(class_name)
+        end
+      rescue Exception
+        Log.warn "Event #{string} not found"
+        nil
+      end
+    end
+
+    def create_event_handlers
+      @event_handlers = {}
+
+      # Probably can do something to not hardcode this
+      # Iterate over the event types
+      [CommitCommentEvent, CreateEvent, DeleteEvent,
+      DeploymentEvent, DeploymentStatusEvent, DownloadEvent,
+      FollowEvent, ForkApplyEvent, ForkEvent,
+      GistEvent, GollumEvent, IssueCommentEvent,
+      IssueCommentEvent, IssuesEvent, MemberEvent,
+      PublicEvent, PullRequestEvent, PullRequestReviewCommentEvent,
+      PushEvent, ReleaseEvent, StatusEvent,
+      TeamAddEvent, WatchEvent].each do | event_type |
+
+        # Map list of concrete event handler to their event type
+        @event_handlers[event_type] = event_type.descendants.map { |handler|
+          handler.new
+        }
+      end
+    end
+  end
+end

data/spec/log_spec.rb

@@ -0,0 +1,54 @@
+require 'spec_helper'
+
+include GitHubArchiveParser
+
+describe Log do
+
+  let(:log) { Class.new(Log) }
+
+  context "initializes instance" do
+    it "should acts as singleton" do
+      log.instance.should == log.instance
+    end
+
+    it "should have a logger" do
+      log.instance.logger.should be_a Logger
+    end
+
+    it "should be a Log (FakeLog)" do
+      log.class.should eq Log.class
+    end
+
+    it "should react to Logger methods" do
+      Logger.public_instance_methods.each do |method|
+        log.valid_method?(method).should be_true
+      end
+    end
+  end
+
+  context "#use_debug" do
+    it "logger's progname before" do
+      log.progname.should be_nil
+    end
+
+    it "logger's progname after" do
+      log.use_debug
+      log.progname.should_not be_nil
+    end
+  end
+
+  context "#parse_caller" do
+    context "with nothing" do
+      it { log.parse_caller(nil).should be_nil }
+    end
+
+    context "with jumble (random text)" do
+      it { log.parse_caller("asdaacsdc").should be_nil }
+    end
+
+    context "with valid caller" do
+      it { log.parse_caller("github_archive_parser/lib/github_archive_parser.rb:45:in `respond_to_missing?'").should eq "github_archive_parser/lib/github_archive_parser.rb:45" }
+    end
+  end
+
+end

metadata

@@ -0,0 +1,75 @@
+--- !ruby/object:Gem::Specification
+name: github_archive_parser
+version: !ruby/object:Gem::Version
+  version: 0.1.0
+platform: ruby
+authors:
+- Kevin Jalbert
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2014-01-17 00:00:00.000000000 Z
+dependencies: []
+description: Gem which parses GitHub Archive data
+email:
+- kevin.j.jalbert@gmail.com
+executables:
+- github_archive_parser
+extensions: []
+extra_rdoc_files: []
+files:
+- bin/github_archive_parser
+- lib/github_archive_parser.rb
+- lib/github_archive_parser/XCust.rb
+- lib/github_archive_parser/event_handler.rb
+- lib/github_archive_parser/event_handlers/commit_comment_event.rb
+- lib/github_archive_parser/event_handlers/create_event.rb
+- lib/github_archive_parser/event_handlers/delete_event.rb
+- lib/github_archive_parser/event_handlers/deployment_event.rb
+- lib/github_archive_parser/event_handlers/deployment_status_event.rb
+- lib/github_archive_parser/event_handlers/download_event.rb
+- lib/github_archive_parser/event_handlers/follow_event.rb
+- lib/github_archive_parser/event_handlers/fork_apply_event.rb
+- lib/github_archive_parser/event_handlers/fork_event.rb
+- lib/github_archive_parser/event_handlers/gist_event.rb
+- lib/github_archive_parser/event_handlers/gollum_event.rb
+- lib/github_archive_parser/event_handlers/issue_comment_event.rb
+- lib/github_archive_parser/event_handlers/issues_event.rb
+- lib/github_archive_parser/event_handlers/member_event.rb
+- lib/github_archive_parser/event_handlers/public_event.rb
+- lib/github_archive_parser/event_handlers/pull_request_event.rb
+- lib/github_archive_parser/event_handlers/pull_request_review_comment_event.rb
+- lib/github_archive_parser/event_handlers/push_event.rb
+- lib/github_archive_parser/event_handlers/release_event.rb
+- lib/github_archive_parser/event_handlers/status_event.rb
+- lib/github_archive_parser/event_handlers/team_add_event.rb
+- lib/github_archive_parser/event_handlers/watch_event.rb
+- lib/github_archive_parser/initialize.rb
+- lib/github_archive_parser/log.rb
+- lib/github_archive_parser/version.rb
+- spec/log_spec.rb
+homepage: https://github.com/kevinjalbert/github_archive_parser
+licenses: []
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: 1.9.3
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 2.2.1
+signing_key: 
+specification_version: 4
+summary: Gem which parses GitHub Archive data
+test_files:
+- spec/log_spec.rb