github_repo_statistics 2.2.14 → 2.2.16

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: dd2582ff1a032bad4e27ef0529ed83df931ffa781a9e9c2ca59b935ec8620deb
-  data.tar.gz: 931b612913c053acd5207092a0955c4532b301cfa62391f4314a046114e953c2
+  metadata.gz: b4d3e1e79a329a849e3dc178cd55be3d0b420738ca81f04e8ac54775f1c92e4c
+  data.tar.gz: e609fbeea3149601a2a14d823d4e5fd47b4f09626b651bab2e419261482018f7
 SHA512:
-  metadata.gz: f434060566ac717c0b8542a7825d240e85151fccd5a711cce0050a32764435247b147ec6d80fdb69ecc5fee43dab4a99eba1766557197faf8e4d5d933fe7aa10
-  data.tar.gz: 1ceeafecbbd2146816d8669e75c2d4fd645322b8c88259937520aeeb9ca11ccee3d9388e4eecf65d36777b95481b5c7ecf548d7499c33b169170245cdb274349
+  metadata.gz: ba6a269abb17d017b184baf8f68df4f7ed749d9dca1d7a361baa2a1bba3e17de6255cd10379177089e3ac2f2d660ad01ce7862b37f2c67b02a268cac17e8625f
+  data.tar.gz: 5ab0ae384a28af25dde05d51853969213fa941f7f4add7edbe7718d71d4c03f08955c316c8ebe753be77c72756aaec6775c67187724eb72d4055664575ba1bf1

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    github_repo_statistics (2.2.13)
+    github_repo_statistics (2.2.15)
       date
       faraday-retry
       google-cloud-bigquery

data/bin/release_merge_report

@@ -0,0 +1,57 @@
+#!/usr/bin/env ruby
+# frozen_string_literal: true
+
+require 'date'
+require 'optparse'
+require 'pry'
+require_relative '../lib/github_repo_statistics/version'
+require_relative '../lib/github_repo_statistics/release_merge_report'
+
+
+options = {}
+OptionParser.new do |opts|
+  opts.banner = 'Usage: github_repo_statistics [options]'
+
+  opts.on('--duration-in-days STRING',
+          'Number of days to aggregate the changes for [default: 90]') do |duration_in_days|
+    options[:duration_in_days] = duration_in_days.to_i
+  end
+
+  opts.on('--default-branch STRING',
+          'The default branch to pull and run metrics for [default: master]') do |default_branch|
+    options[:default_branch] = default_branch
+  end
+
+  opts.on('--github-token STRING',
+          'GitHub API token [default: ""]') do |github_token|
+    options[:github_token] = github_token
+  end
+
+  opts.on('--github-repo STRING',
+          'GitHub repository name and owner (e.g. octokit/octokit) [default: ""]') do |github_repo|
+    options[:github_repo] = github_repo
+  end
+
+  opts.on('-v', '--version', 'Display the version of the gem') do
+    puts "github_repo_statistics version #{GithubRepoStatistics::VERSION}"
+    exit
+  end
+
+  opts.on('-h', '--help', 'Display this help message') do
+    puts opts
+    puts <<~EXAMPLES
+
+      Examples:
+        review_report --github-token ghp_bKasj29ndkfdksf3rjt2ngi43j --github-repo octokit/octokit
+    EXAMPLES
+    exit
+  end
+end.parse!
+
+DEFAULT_BRANCH = options[:default_branch] || 'master'
+
+raise 'Please provide GitHub token using --github-token flag' if options[:github_token].nil?
+raise 'Please provide GitHub repo name using --github-repo flag' if options[:github_repo].nil?
+raise 'Please provide default GitHub branch using --default-branch flag' if DEFAULT_BRANCH.nil?
+
+ReleaseMergeReport.new(token: options[:github_token], repo: options[:github_repo], branch_prefix: "release/").report

data/lib/github_repo_statistics/force_merge_report.rb

@@ -30,8 +30,6 @@ class ForceMergeReport
     weekly_summaries = Hash.new { |hash, key| hash[key] = { total: 0, failed: 0, workflows: Hash.new(0) } }
 
     weeks = @duration_in_days / 7
-    require 'pry'
-    binding.pry
 
     (weekly_pull_requests.keys.sort[-weeks..] || []).each do |month|
       pull_requests_for_month = weekly_pull_requests[month]
@@ -93,6 +91,8 @@ class ForceMergeReport
         puts "  - #{workflow}: #{count}"
       end
 
+      # ENV['BQ_CREDENTIALS'] = `cat /Users/serghei.moret/.config/gcloud/application_default_credentials.json`
+
       if ENV['BQ_CREDENTIALS']
         require "google/cloud/bigquery"
         require "json"

data/lib/github_repo_statistics/github_repo_statistics.rb

@@ -2,6 +2,7 @@
 
 require 'pry'
 require 'date'
+require 'csv'
 
 class GithubRepoStatistics
   def initialize(directory_path:, duration_in_days:, begin_time:, debug: nil, steps: 1)
@@ -145,6 +146,10 @@ class GithubRepoStatistics
     count
   end
 
+  def count_lines_of_code(file)
+    File.foreach(file).reject { |line| line.match(%r{^\s*(//|/\*.*\*/|\s*$)}) }.count
+  end
+
   def filter_existing_code_files(files)
     files.select do |f|
       next unless File.exist?(f)
@@ -242,7 +247,9 @@ class GithubRepoStatistics
     files_with_single_contributor_percentage = (100 - ((files_changed_by_many_teams.to_f / file_count) * 100)).round(2)
     hotspot_lines = count_hotspot_lines(filtered_files.keys)
     big_files_count = count_big_files(@directory_path)
-    
+
+    # ENV['BQ_CREDENTIALS'] = `cat /Users/serghei.moret/.config/gcloud/application_default_credentials.json`
+
     if ENV['BQ_CREDENTIALS']
       require "google/cloud/bigquery"
       require "json"
@@ -251,38 +258,39 @@ class GithubRepoStatistics
         project_id: "hellofresh-android",
         credentials: creds
       )
-      dataset = bigquery.dataset "modularization"
+      dataset = bigquery.dataset "github_data"
 
       files_with_multiple_contributor = file_team_map.count
       big_files_with_multiple_contributors = filtered_top_touched_files.count
       total_files_changed = uniq_code_files_with_changes.count
 
+      platform = if @directory_path == 'HelloFresh/HelloFresh/'
+                   'ios'
+                 elsif @directory_path == 'features/legacy/'
+                   'android'
+                 end
+
       query = <<~SQL
-  MERGE INTO modularization AS target
-  USING (SELECT '#{@begin_time}' AS date, '#{@directory_path}' AS platform) AS source
-  ON target.date = source.date AND target.platform = source.platform
-  WHEN MATCHED THEN
-    UPDATE SET
-      target.platform = '#{@directory_path}',
-      target.single_contributor_percentage = #{files_with_single_contributor_percentage},
-      target.files_changed_by_many_teams = #{files_changed_by_many_teams},
-      target.file_count = #{file_count},
-      target.cross_teams_count = #{cross_teams_count},
-      target.single_ownership_teams_count = #{single_ownership_teams_count},
-      target.hotspot_changes_percentage = #{hotspot_changes_percentage},
-      target.churn_count = #{churn_count},
-      target.total_changes = #{total_changes},
-      target.files_with_multiple_contributor = #{files_with_multiple_contributor},
-      target.big_files_with_multiple_contributors = #{big_files_with_multiple_contributors},
-      target.total_files_changed = #{total_files_changed},
-      target.hotspot_lines = #{hotspot_lines},
-      target.big_files_count = #{big_files_count}
-  WHEN NOT MATCHED THEN
-    INSERT (date,platform,single_contributor_percentage,files_changed_by_many_teams,file_count,cross_teams_count,single_ownership_teams_count,hotspot_changes_percentage,churn_count,total_changes,files_with_multiple_contributor,big_files_with_multiple_contributors,total_files_changed,hotspot_lines,big_files_count)
-    VALUES (#{@begin_time}, #{@directory_path}, #{files_with_single_contributor_percentage}, #{files_changed_by_many_teams}, #{file_count}, ##{cross_teams_count}, #{single_ownership_teams_count}, #{hotspot_changes_percentage}, #{churn_count}, #{total_changes}, #{files_with_multiple_contributor}, #{big_files_with_multiple_contributors}, #{total_files_changed}, #{hotspot_lines}, #{big_files_count});
+    INSERT INTO modularization (date, platform, single_contributor_percentage, files_changed_by_many_teams, file_count, cross_teams_count, single_ownership_teams_count, hotspot_changes_percentage, churn_count, total_changes, files_with_multiple_contributor, big_files_with_multiple_contributors, total_files_changed, hotspot_lines, big_files_count)
+    VALUES ('#{@begin_time}', '#{platform}', #{files_with_single_contributor_percentage}, #{files_changed_by_many_teams}, #{file_count}, #{cross_teams_count}, #{single_ownership_teams_count}, #{hotspot_changes_percentage}, #{churn_count}, #{total_changes}, #{files_with_multiple_contributor}, #{big_files_with_multiple_contributors}, #{total_files_changed}, #{hotspot_lines}, #{big_files_count});
       SQL
 
       dataset.query(query)
+
+#       delete_query = <<~SQL
+# DELETE FROM modularization
+# WHERE CONCAT(DATE(date), ' ', TIME(date)) NOT IN (
+#     SELECT CONCAT(DATE(date), ' ', TIME(date))
+#     FROM modularization AS m1
+#     WHERE TIME(date) = (
+#         SELECT MAX(TIME(date))
+#         FROM modularization AS m2
+#         WHERE DATE(m1.date) = DATE(m2.date)
+#     )
+# );
+#       SQL
+
+      # dataset.query(delete_query)
     end
 
     puts ''
@@ -306,23 +314,30 @@ class GithubRepoStatistics
     puts "     *Contributors:* #{contributors}"
 
     if HOTSPOT
-      hotspot_output = "\n   *Hotspot files(#{filtered_top_touched_files.count}):*\n"
+      hotspot_output = []
 
       filtered_top_touched_files.each do |line|
-        hotspot_output += "\n"
         file = line.first
         contributors = line.last.first
+        lines_of_code = count_lines_of_code(file)
         commits = line.last.last
-        hotspot_output += "     #{file.gsub(@directory_path,
-                                            '')} Contributors: #{contributors} Commits: #{commits} Owner: #{find_owner(file:)}\n"
+        owner = find_owner(file:)
+
+        hotspot_output << [file.gsub(@directory_path, ''), contributors, lines_of_code, commits, owner]
       end
 
       if FILE_OUTPUT
-        File.open('hotspot.txt', 'w') do |f|
-          f.puts hotspot_output
+        CSV.open('hotspot.csv', 'w') do |csv|
+          csv << ['File', 'Contributors', 'Lines', 'Commits', 'Owner']
+          hotspot_output.each do |row|
+            csv << row
+          end
         end
       else
-        puts hotspot_output
+        puts "\n   *Hotspot files(#{filtered_top_touched_files.count}):*\n"
+        hotspot_output.each do |row|
+          puts "     #{row[0]} Contributors: #{row[1]} Lines: #{row[2]} Commits: #{row[3]} Owner: #{row[4]}"
+        end
       end
     end
 

data/lib/github_repo_statistics/release_merge_report.rb

@@ -0,0 +1,112 @@
+require 'octokit'
+require 'json'
+require 'google/cloud/bigquery'
+
+class ReleaseMergeReport
+  def initialize(token:, repo:, branch_prefix:)
+    @token = token
+    @repo = repo
+    @branch_prefix = branch_prefix
+  end
+
+  def report
+    branch_counts = count_merged_pull_requests_per_branch
+    grouped_branch_counts = group_branch_counts(branch_counts)
+    require 'pry'
+    binding.pry
+
+    # Print the grouped branch counts
+    puts 'Branches with Merged Pull Requests:'
+    grouped_branch_counts.each do |branch, count|
+      puts "#{branch}: #{count}"
+    end
+
+    ENV['BQ_CREDENTIALS'] = `cat /Users/serghei.moret/.config/gcloud/application_default_credentials.json`
+
+    export_to_bigquery(grouped_branch_counts) if ENV['BQ_CREDENTIALS']
+
+    grouped_branch_counts
+  end
+
+  private
+
+  def count_merged_pull_requests_per_branch
+    client = Octokit::Client.new(access_token: @token)
+    client.auto_paginate = true
+
+    tags = client.tags(@repo)
+    branch_counts = Hash.new(0)
+
+    tags.each do |tag|
+      next if !tag.name.match?(/^(v23|v24)\./) && !tag.name.match?(/^(23|24)\./)
+
+      # Extract release version from the tag name
+      release_version = tag.name.gsub('v', '')
+
+      # Construct branch name with the specified prefix
+      branch_name = "#{@branch_prefix}#{release_version}"
+
+      # Count merged pull requests associated with the branch
+      pull_requests = client.pull_requests(@repo, state: 'closed', sort: 'updated', direction: 'desc', base: branch_name)
+                            .select { |pr| pr.merged_at }
+
+      branch_counts[branch_name] = pull_requests.size
+    end
+
+    branch_counts
+  end
+
+  def group_branch_counts(branch_counts)
+    patch_counts = Hash.new(0)
+    hotfix_counts = Hash.new { |hash, key| hash[key] = Hash.new(0) }
+
+    branch_counts.each do |branch, count|
+      major_minor_version, patch_version = branch.match(/^#{@branch_prefix}(\d+\.\d+)(?:\.(\d+))?/)&.captures
+
+      if patch_version.nil?
+        # Branch is a patch version
+        patch_counts[major_minor_version] += count
+      elsif count > 0
+        # Branch is a hotfix version
+        hotfix_counts[major_minor_version][patch_version] += count
+      end
+    end
+
+    # Sum up the counts for hotfix versions within the same major and minor version
+    hotfix_counts.each do |major_minor_version, hotfixes|
+      hotfix_counts[major_minor_version] = hotfixes.values.sum
+    end
+
+    { patch_versions: patch_counts, hotfix_versions: hotfix_counts }
+  end
+
+  def export_to_bigquery(branch_counts)
+    require "google/cloud/bigquery"
+    require "json"
+    creds = JSON.parse(ENV['BQ_CREDENTIALS'])
+    bigquery = Google::Cloud::Bigquery.new(
+      project_id: "hellofresh-android",
+      credentials: creds
+    )
+    dataset = bigquery.dataset "github_data"
+
+    date = DateTime.now
+
+    branch_counts.each do |branch, count|
+      query = <<~SQL
+        MERGE INTO release_merges AS target
+        USING (SELECT '#{branch}' AS release, '#{@repo}' AS platform) AS source
+        ON target.release = source.release AND target.platform = source.platform
+        WHEN MATCHED THEN
+          UPDATE SET
+            target.merge_count = #{count},
+            target.timestamp = '#{date}'
+        WHEN NOT MATCHED THEN
+          INSERT (release, merge_count, platform, timestamp)
+          VALUES ('#{branch}', #{count}, '#{@repo}', '#{date}');
+      SQL
+      date -= 7
+      dataset.query(query)
+    end
+  end
+end

data/lib/github_repo_statistics/review_report.rb

@@ -87,6 +87,8 @@ class ReviewReport
         puts "  Reviews with comments: #{reviews_with_comments}"
         puts "  Change requested reviews: #{change_requested_reviews}"
 
+        # ENV['BQ_CREDENTIALS'] = `cat /Users/serghei.moret/.config/gcloud/application_default_credentials.json`
+
         if ENV['BQ_CREDENTIALS']
           require "google/cloud/bigquery"
           require "json"

data/lib/github_repo_statistics/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 class GithubRepoStatistics
-  VERSION = '2.2.14'
+  VERSION = '2.2.16'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: github_repo_statistics
 version: !ruby/object:Gem::Version
-  version: 2.2.14
+  version: 2.2.16
 platform: ruby
 authors:
 - Serghei Moret
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-03-15 00:00:00.000000000 Z
+date: 2024-04-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: date
@@ -87,6 +87,7 @@ executables:
 - console
 - force_merge_report
 - github_repo_statistics
+- release_merge_report
 - review_report
 - setup
 extensions: []
@@ -108,12 +109,14 @@ files:
 - bin/console
 - bin/force_merge_report
 - bin/github_repo_statistics
+- bin/release_merge_report
 - bin/review_report
 - bin/setup
 - github_repo_statistics.gemspec
 - lib/github_repo_statistics.rb
 - lib/github_repo_statistics/force_merge_report.rb
 - lib/github_repo_statistics/github_repo_statistics.rb
+- lib/github_repo_statistics/release_merge_report.rb
 - lib/github_repo_statistics/review_report.rb
 - lib/github_repo_statistics/version.rb
 - sig/github_repo_statistics.rbs