github_repo_statistics 2.3.5 → 2.3.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 538bcf5774623f5443c771c731035806d40bc7ecced742899a95168dca809094
-  data.tar.gz: '063853f3714c6a18e70d9eb936e88d2ef7c592203936412f6f9981960ae5db7f'
+  metadata.gz: 722258cfcbda685136e97c5b844f38b4da2937dbf3693b5131c628aad5131672
+  data.tar.gz: e0377225ed1e088b3eef9ff283ebf254225b36546eae619378f28df4f64cd800
 SHA512:
-  metadata.gz: 7073aca1430b973a1a4ef07be6436843d532539bcd0aa5e12caaa22cf6b9c2e6152538d366934e4436aae95d3a64a6fecbe408ad5c9770e1ca6dda3a3aadaafe
-  data.tar.gz: 603ecfda33d86d85f8c41a4ea40860830c6e2311b37cb4e8c31d9d58b1a80fd1204c961a2b0326329df42298a9a7b95dea7a05d39da711792f2ccd4119ad450c
+  metadata.gz: 37cc51e5dd9c5c40ee4b816fcbfa39e0c288fe6fa17c298890e6f6dbc508923dd094f2bd75969bc906695c69a6e5bec6d36ccdb380b8c1d9a455247b86ce2af4
+  data.tar.gz: 355b86a47bb3f55ac5426256a840e3e4a77320a093524f77c93627d73b08c53df3762a5c9c8bab34992f182ba7a74575806442933855858a4f17d245d3986cf2

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    github_repo_statistics (2.3.5)
+    github_repo_statistics (2.3.7)
       date
       faraday-retry
       google-cloud-bigquery

data/bin/github_repo_statistics

@@ -78,6 +78,16 @@ OptionParser.new do |opts|
     options[:code_extension] = code_extension
   end
 
+  opts.on('--team-to-focus STRING',
+          'The team identifier to focus for the metric collection [default: ""]') do |team_to_focus|
+    options[:team_to_focus] = team_to_focus
+  end
+
+  opts.on('--codeowner-to-focus STRING',
+          'The codeowner identifier to focus for the metric collection [default: ""]') do |codeowner_to_focus|
+    options[:codeowner_to_focus] = codeowner_to_focus
+  end
+
   opts.on('--output-to-files',
           'Puts the output for hotspot and codeowners into files instead of the STDOUT (useful for CI and big amount of data) [default:false]') do
     options[:file_output] = true
@@ -112,6 +122,8 @@ FILE_OUTPUT = options[:file_output] || false
 CODE_EXTENSIONS = options[:code_extension] ? options[:code_extension].split(',') : ['.swift', '.kt']
 EXCLUDED_FILES = options[:excluded_files]
 EXCLUDED_PRS = options[:excluded_prs]
+TEAM_TO_FOCUS = options[:team_to_focus]
+CODEOWNER_TO_FOCUS = options[:codeowner_to_focus]
 
 unless CI
   puts "\nDirectory: #{REPO_PATH}\n"

data/lib/github_repo_statistics/github_repo_statistics.rb

@@ -150,10 +150,24 @@ class GithubRepoStatistics
     File.foreach(file).reject { |line| line.match(%r{^\s*(//|/\*.*\*/|\s*$)}) }.count
   end
 
-  def filter_existing_code_files(files)
+  def filter_existing_code_files(files, start_date, end_date)
     files.select do |f|
       next unless File.exist?(f)
 
+      git_log = git_commit_info(file: f, start_date:, end_date:).split("\n")
+
+      teams = git_log.map do |team|
+        team.match(/#{TEAM_REGEX}/)[0].upcase
+      end.reject { |e| EXCLUSIONS&.include?(e) }
+
+      if TEAM_TO_FOCUS && CODEOWNER_TO_FOCUS
+        next if !teams.include?(TEAM_TO_FOCUS) && !find_owner(file: f).include?(CODEOWNER_TO_FOCUS)
+      elsif TEAM_TO_FOCUS
+        next unless teams.include?(TEAM_TO_FOCUS)
+      elsif CODEOWNER_TO_FOCUS
+        next unless find_owner(file: f).include?(CODEOWNER_TO_FOCUS)
+      end
+
       if EXCLUDED_FILES
         excluded_patterns = EXCLUDED_FILES.split(',')
         next if excluded_patterns.any? { |pattern| f.include?(pattern) }
@@ -213,9 +227,10 @@ class GithubRepoStatistics
         team.match(/#{TEAM_REGEX}/)[0].upcase
       end.reject { |e| EXCLUSIONS&.include?(e) }
 
+      teams = calculate_percentile(teams, 90)
+
       total_changes += commit_count
       all_teams << teams
-      teams = teams.uniq
 
       if teams.count > 1
         files_changed_by_many_teams += 1
@@ -231,6 +246,30 @@ class GithubRepoStatistics
      file_team_map]
   end
 
+  def calculate_percentile(arr, percentile)
+    # Count occurrences of each unique element
+    counts = arr.each_with_object(Hash.new(0)) { |item, hash| hash[item] += 1 }
+
+    # Sort elements by their counts in descending order
+    sorted_counts = counts.sort_by { |k, v| -v }.to_h
+
+    # Calculate the cut-off for the percentile
+    total_count = arr.size
+    cutoff = total_count * (percentile / 100.0)
+
+    # Select elements that meet the percentile criteria
+    selected_elements = []
+    cumulative_count = 0
+
+    sorted_counts.each do |item, count|
+      cumulative_count += count
+      selected_elements << item
+      break if cumulative_count >= cutoff
+    end
+
+    selected_elements
+  end
+
   def filter_files(file_team_map:, size: BIG_FILE_SIZE)
     file_team_map.select do |file_path|
       next unless File.exist?(file_path)
@@ -245,10 +284,11 @@ class GithubRepoStatistics
     start_date = @begin_time.to_time.to_i - duration_in_days * 86_400
     end_date = @begin_time.to_time.to_i
     git_ls = git_files(directory_path: @directory_path)
-    file_count = filter_existing_code_files(git_ls.split).count
+    file_count = filter_existing_code_files(git_ls.split, start_date, end_date).count
     all_files_with_changes = files_with_changes(directory_path: @directory_path, start_date:, end_date:).split.sort
-    code_files_with_changes = filter_existing_code_files(all_files_with_changes)
+    code_files_with_changes = filter_existing_code_files(all_files_with_changes, start_date, end_date)
     uniq_code_files_with_changes = code_files_with_changes.uniq
+
     all_teams, cross_teams_count, single_ownership_teams_count, files_changed_by_many_teams, total_changes, file_team_map = analyze_changed_files(
       uniq_code_files_with_changes:, start_date:, end_date:
     )

data/lib/github_repo_statistics/release_merge_report.rb

@@ -20,7 +20,7 @@ class ReleaseMergeReport
       puts "#{branch}: #{count}"
     end
 
-    # ENV['BQ_CREDENTIALS'] = `cat /Users/serghei.moret/.config/gcloud/application_default_credentials.json`
+    ENV['BQ_CREDENTIALS'] = `cat /Users/serghei.moret/.config/gcloud/application_default_credentials.json`
 
     export_to_bigquery(grouped_branch_counts) if ENV['BQ_CREDENTIALS']
 

data/lib/github_repo_statistics/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 class GithubRepoStatistics
-  VERSION = '2.3.5'
+  VERSION = '2.3.7'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: github_repo_statistics
 version: !ruby/object:Gem::Version
-  version: 2.3.5
+  version: 2.3.7
 platform: ruby
 authors:
 - Serghei Moret
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-05-22 00:00:00.000000000 Z
+date: 2024-06-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: date