RubyGems - branch_base - Versions diffs - 0.1.0 - Mend

branch_base 0.1.0

Files changed (29) hide show

checksums.yaml +7 -0
data/.github/dependabot.yml +11 -0
data/.github/workflows/ci.yaml +76 -0
data/.gitignore +19 -0
data/.prettierignore +4 -0
data/.rspec +5 -0
data/.rubocop.yml +259 -0
data/Dockerfile +5 -0
data/Gemfile +6 -0
data/Gemfile.lock +124 -0
data/LICENSE.txt +21 -0
data/README.md +152 -0
data/Rakefile +13 -0
data/branch_base.gemspec +43 -0
data/internal/git-wrapped.png +0 -0
data/internal/screenshot.jpg +0 -0
data/internal/screenshot.png +0 -0
data/internal/template.html.erb +84 -0
data/lib/branch_base/cli.rb +79 -0
data/lib/branch_base/database.rb +98 -0
data/lib/branch_base/repository.rb +43 -0
data/lib/branch_base/sync.rb +211 -0
data/lib/branch_base/version.rb +5 -0
data/lib/branch_base.rb +97 -0
data/package.json +13 -0
data/scripts/branch_base +6 -0
data/scripts/release_branch_base.sh +22 -0
data/yarn.lock +20 -0
metadata +297 -0

data/branch_base.gemspec ADDED Viewed

@@ -0,0 +1,43 @@
+# frozen_string_literal: true
+require_relative "lib/branch_base/version"
+Gem::Specification.new do |spec|
+  spec.name = "branch_base"
+  spec.version = BranchBase::VERSION
+  spec.authors = ["Shayon Mukherjee"]
+  spec.email = ["shayonj@gmail.com"]
+  spec.summary = "Sync Git Repository into a SQLite Database"
+  spec.description =
+    "BranchBase provides a CLI to synchronize a Git repository into a SQLite database."
+  spec.homepage = "https://github.com/shayonj/branch_base"
+  spec.license = "MIT"
+  spec.required_ruby_version = ">= 2.7.0"
+  spec.files =
+    `git ls-files -z`.split("\x0")
+      .reject { |f| f.match(%r{^(test|spec|features)/}) }
+  spec.bindir = "scripts"
+  spec.executables = ["branch_base"]
+  spec.require_paths = ["lib"]
+  spec.add_dependency("rugged")
+  spec.add_dependency("sqlite3")
+  spec.add_dependency("thor", "~> 1.0")
+  spec.metadata = { "rubygems_mfa_required" => "true" }
+  spec.add_development_dependency("git")
+  spec.add_development_dependency("prettier_print")
+  spec.add_development_dependency("pry")
+  spec.add_development_dependency("rake")
+  spec.add_development_dependency("rspec")
+  spec.add_development_dependency("rubocop")
+  spec.add_development_dependency("rubocop-packaging")
+  spec.add_development_dependency("rubocop-performance")
+  spec.add_development_dependency("rubocop-rake")
+  spec.add_development_dependency("rubocop-rspec")
+  spec.add_development_dependency("syntax_tree")
+  spec.add_development_dependency("syntax_tree-haml")
+  spec.add_development_dependency("syntax_tree-rbs")
+end

data/internal/git-wrapped.png ADDED Viewed

Binary file

data/internal/screenshot.jpg ADDED Viewed

Binary file

data/internal/screenshot.png ADDED Viewed

Binary file

data/internal/template.html.erb ADDED Viewed

@@ -0,0 +1,84 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Git Wrapped Statistics - <%= @repo_name.capitalize %> </title>
+    <style>
+      body {
+        font-family: "Fira Code", monospace;
+        background-color: #0C0A00;
+        color: #c1c1c1;
+        margin: 0;
+        padding: 0;
+        font-weight: 100;
+      }
+      header {
+        background-color: #0C0A00;
+        color: #fff;
+        text-align: center;
+        padding: 20px;
+        border-bottom: 1px solid #333;
+      }
+      h1 {
+        font-size: 30px;
+        margin: 0;
+      }
+      .container {
+        max-width: 1200px;
+        margin: 20px auto;
+        background-color: #0C0A00;
+        border-radius: 10px;
+        overflow: hidden;
+        display: grid;
+        grid-template-columns: repeat(auto-fill, minmax(500px, 1fr));
+        gap: 20px;
+      }
+      .section {
+        padding: 20px;
+        border-bottom: 1px solid #333;
+      }
+      h2 {
+        font-size: 20px;
+        margin: 0 0 20px;
+        color: #fff;
+      }
+      ul {
+        /* list-style: none; */
+        padding: 0;
+      }
+      li {
+        margin-bottom: 10px;
+        font-size: 18px;
+      }
+    </style>
+    <link
+      rel="stylesheet"
+      href="https://cdnjs.cloudflare.com/ajax/libs/firacode/5.2.0/fira_code.css"
+    />
+  </head>
+  <body>
+    <header>
+      <h1>Git Wrapped Statistics - <%= @repo_name.capitalize %></h1>
+    </header>
+    <div class="container">
+        <% @results.each do |key, values| %>
+            <div class="section">
+                <h2><%= @emojis[key] %> <%= key.gsub('_', ' ').split.map(&:capitalize).join(' ') %></h2>
+                <ul>
+                    <% values.each do |item| %>
+                        <li><%= item[0] %>: <%= item[1] %><%= item[2].nil? ? "" : "(#{item[2]})"%></li>
+                    <% end %>
+                </ul>
+            </div>
+        <% end %>
+    </div>
+  </body>
+</html>

data/lib/branch_base/cli.rb ADDED Viewed

@@ -0,0 +1,79 @@
+# frozen_string_literal: true
+require "thor"
+require "erb"
+require "fileutils"
+module BranchBase
+  class CLI < Thor
+    desc "sync REPO_PATH", "Synchronize a Git directory to a SQLite database"
+    def sync(repo_path)
+      BranchBase.logger.info("Starting sync process for #{repo_path}...")
+      full_repo_path = File.expand_path(repo_path)
+      unless File.directory?(File.join(full_repo_path, ".git"))
+        BranchBase.logger.error(
+          "The specified path is not a valid Git repository: #{full_repo_path}",
+        )
+        exit(1)
+      end
+      repo_name = File.basename(full_repo_path)
+      db_directory = full_repo_path
+      db_filename = File.join(db_directory, "#{repo_name}_git_data.db")
+      database = Database.new(db_filename)
+      repository = Repository.new(full_repo_path)
+      start_time = Time.now
+      sync = Sync.new(database, repository)
+      sync.run
+      elapsed_time = Time.now - start_time
+      BranchBase.logger.info(
+        "Repository data synced successfully in #{db_filename} in #{elapsed_time.round(2)} seconds",
+      )
+    end
+    desc "git-wrapped REPO_PATH",
+         "Generate Git wrapped statistics for the given repository"
+    def git_wrapped(repo_path)
+      BranchBase.logger.info("Generating Git wrapped for #{repo_path}...")
+      full_repo_path = File.expand_path(repo_path)
+      @repo_name = File.basename(full_repo_path)
+      db_filename = File.join(full_repo_path, "#{@repo_name}_git_data.db")
+      unless File.exist?(db_filename)
+        BranchBase.logger.error("Database file not found: #{db_filename}")
+        exit(1)
+      end
+      database = Database.new(db_filename)
+      @results = BranchBase.execute_git_wrapped_queries(database)
+      @emojis = BranchBase.emojis_for_titles
+      json_full_path = "#{full_repo_path}/git-wrapped.json"
+      File.write(json_full_path, JSON.pretty_generate(@results))
+      BranchBase.logger.info("Git wrapped JSON stored in #{json_full_path}")
+      erb_template = File.read("./internal/template.html.erb")
+      erb = ERB.new(erb_template)
+      generated_html = erb.result(binding)
+      html_full_path = "#{full_repo_path}/git-wrapped.html"
+      File.write(html_full_path, generated_html)
+      BranchBase.logger.info("Git wrapped HTML stored in #{html_full_path}")
+    end
+    desc "version", "Prints the version"
+    def version
+      puts BranchBase::VERSION
+    end
+    def self.exit_on_failure?
+      true
+    end
+  end
+end

data/lib/branch_base/database.rb ADDED Viewed

@@ -0,0 +1,98 @@
+# frozen_string_literal: true
+require "sqlite3"
+module BranchBase
+  class Database
+    def initialize(db_path)
+      @db = SQLite3::Database.new(db_path)
+      setup_schema
+    end
+    def execute(query, *params)
+      @db.execute(query, *params)
+    end
+    def prepare(statement)
+      @db.prepare(statement)
+    end
+    def transaction(&block)
+      @db.transaction(&block)
+    end
+    def last_insert_row_id
+      @db.last_insert_row_id
+    end
+    def setup_schema
+      @db.execute_batch(<<-SQL)
+        CREATE TABLE IF NOT EXISTS repositories (
+          repo_id INTEGER PRIMARY KEY AUTOINCREMENT,
+          name TEXT NOT NULL,
+          url TEXT NOT NULL
+        );
+        CREATE TABLE IF NOT EXISTS commits (
+          commit_hash TEXT PRIMARY KEY,
+          repo_id INTEGER NOT NULL,
+          author TEXT NOT NULL,
+          committer TEXT NOT NULL,
+          message TEXT NOT NULL,
+          timestamp TEXT NOT NULL,
+          FOREIGN KEY (repo_id) REFERENCES repositories (repo_id)
+        );
+        CREATE TABLE IF NOT EXISTS branches (
+          branch_id INTEGER PRIMARY KEY AUTOINCREMENT,
+          repo_id INTEGER NOT NULL,
+          name TEXT NOT NULL,
+          FOREIGN KEY (repo_id) REFERENCES repositories (repo_id)
+        );
+        CREATE TABLE IF NOT EXISTS branch_commits (
+          branch_id INTEGER NOT NULL,
+          commit_hash TEXT NOT NULL,
+          PRIMARY KEY (branch_id, commit_hash),
+          FOREIGN KEY (branch_id) REFERENCES branches (branch_id),
+          FOREIGN KEY (commit_hash) REFERENCES commits (commit_hash)
+        );
+        CREATE TABLE IF NOT EXISTS files (
+          file_id INTEGER PRIMARY KEY AUTOINCREMENT,
+          repo_id INTEGER NOT NULL,
+          file_path TEXT NOT NULL,
+          latest_commit TEXT NOT NULL,
+          FOREIGN KEY (repo_id) REFERENCES repositories (repo_id),
+          FOREIGN KEY (latest_commit) REFERENCES commits (commit_hash)
+        );
+        CREATE TABLE IF NOT EXISTS commit_files (
+          commit_hash TEXT NOT NULL,
+          file_id INTEGER NOT NULL,
+          changes TEXT NOT NULL,
+          PRIMARY KEY (commit_hash, file_id),
+          FOREIGN KEY (commit_hash) REFERENCES commits (commit_hash),
+          FOREIGN KEY (file_id) REFERENCES files (file_id)
+        );
+        CREATE TABLE IF NOT EXISTS commit_parents (
+          commit_hash TEXT NOT NULL,
+          parent_hash TEXT NOT NULL,
+          PRIMARY KEY (commit_hash, parent_hash),
+          FOREIGN KEY (commit_hash) REFERENCES commits (commit_hash),
+          FOREIGN KEY (parent_hash) REFERENCES commits (commit_hash)
+        );
+        CREATE INDEX IF NOT EXISTS idx_commits_repo_id ON commits (repo_id);
+        CREATE INDEX IF NOT EXISTS idx_commits_author ON commits (author);
+        CREATE INDEX IF NOT EXISTS idx_commits_committer ON commits (committer);
+        CREATE INDEX IF NOT EXISTS idx_branches_repo_id ON branches (repo_id);
+        CREATE INDEX IF NOT EXISTS idx_files_repo_id ON files (repo_id);
+        CREATE INDEX IF NOT EXISTS idx_files_file_path ON files (file_path);
+        CREATE INDEX IF NOT EXISTS idx_commit_parents_commit_hash ON commit_parents (commit_hash);
+        CREATE INDEX IF NOT EXISTS idx_commit_parents_parent_hash ON commit_parents (parent_hash);
+      SQL
+    end
+  end
+end

data/lib/branch_base/repository.rb ADDED Viewed

@@ -0,0 +1,43 @@
+# frozen_string_literal: true
+require "rugged"
+module BranchBase
+  class Repository
+    attr_reader :repo
+    def initialize(repo_path)
+      @repo = Rugged::Repository.new(repo_path)
+    end
+    def walk(branch_name = nil, &block)
+      # Use the provided branch's head commit OID if a branch name is given,
+      # otherwise, use the repository's HEAD commit OID.
+      oid =
+        if branch_name
+          branch = @repo.branches[branch_name]
+          raise ArgumentError, "Branch not found: #{branch_name}" unless branch
+          branch.target.oid
+        else
+          @repo.head.target.oid
+        end
+      @repo.walk(oid, Rugged::SORT_TOPO, &block)
+    end
+    def default_branch_name
+      head_ref = @repo.head.name
+      head_ref.sub(%r{^refs/heads/}, "")
+    rescue Rugged::ReferenceError
+      nil
+    end
+    def path
+      @repo.path
+    end
+    def branches
+      @repo.branches
+    end
+  end
+end

data/lib/branch_base/sync.rb ADDED Viewed

@@ -0,0 +1,211 @@
+# frozen_string_literal: true
+module BranchBase
+  class Sync
+    # TODO acctualy see if bulk inserts are faster
+    BATCH_SIZE = 1000
+    def initialize(database, repository)
+      @db = database
+      @repo = repository
+    end
+    def run
+      @db.execute("PRAGMA foreign_keys = OFF")
+      repo_id = sync_repository
+      sync_branches(repo_id)
+      sync_commits(repo_id)
+      @db.execute("PRAGMA foreign_keys = ON")
+    end
+    def sync_repository
+      repo_path = @repo.path.chomp(".git/")
+      repo_name = File.basename(repo_path)
+      existing_repo_id =
+        @db.execute(
+          "SELECT repo_id FROM repositories WHERE url = ?",
+          [repo_path],
+        ).first
+      return existing_repo_id[0] if existing_repo_id
+      @db.execute(
+        "INSERT INTO repositories (name, url) VALUES (?, ?)",
+        [repo_name, repo_path],
+      )
+      @db.last_insert_row_id
+    end
+    def sync_branches(repo_id)
+      BranchBase.logger.debug("Syncing branches for repository ID: #{repo_id}")
+      default_branch_name = @repo.default_branch_name
+      return unless default_branch_name
+      @repo.branches.each do |branch|
+        next if branch.name.nil? || branch.target.nil?
+        branch_id = insert_branch(repo_id, branch.name)
+        # Only sync commits for the default branch
+        if branch.name == default_branch_name
+          insert_branch_commits(branch_id, branch)
+        end
+      end
+    end
+    def insert_branch(repo_id, branch_name)
+      existing_branch_id =
+        @db.execute(
+          "SELECT branch_id FROM branches WHERE name = ? AND repo_id = ?",
+          [branch_name, repo_id],
+        ).first
+      return existing_branch_id[0] if existing_branch_id
+      @db.execute(
+        "INSERT INTO branches (repo_id, name) VALUES (?, ?)",
+        [repo_id, branch_name],
+      )
+      @db.last_insert_row_id
+    end
+    def insert_branch_commits(branch_id, branch)
+      BranchBase.logger.debug("Syncing branch commits for: #{branch.name}")
+      head_commit = branch.target
+      walker = Rugged::Walker.new(@repo.repo)
+      walker.push(head_commit)
+      walker.each do |commit|
+        next if commit_exists?(commit.oid)
+        @db.execute(
+          "INSERT OR IGNORE INTO branch_commits (branch_id, commit_hash) VALUES (?, ?)",
+          [branch_id, commit.oid],
+        )
+      end
+    end
+    def sync_commits(repo_id)
+      batched_commits = []
+      batched_files = []
+      @repo.walk do |commit|
+        next if commit_exists?(commit.oid)
+        batched_commits << [
+          commit.oid,
+          repo_id,
+          commit.author[:name],
+          commit.committer[:name],
+          commit.message,
+          commit.time.to_s,
+        ]
+        if batched_commits.size >= BATCH_SIZE
+          insert_commits(batched_commits)
+          batched_commits.clear
+        end
+        commit.diff.each_patch do |patch|
+          file_path = patch.delta.new_file[:path]
+          batched_files << [repo_id, file_path, commit.oid, patch.to_s]
+          if batched_files.size >= BATCH_SIZE
+            insert_files_and_commit_files(batched_files)
+            batched_files.clear
+          end
+        end
+        insert_commit_parents(commit)
+      end
+      insert_commits(batched_commits) unless batched_commits.empty?
+      insert_files_and_commit_files(batched_files) unless batched_files.empty?
+    end
+    private
+    def commit_exists?(commit_hash)
+      @db.execute(
+        "SELECT COUNT(*) FROM commits WHERE commit_hash = ?",
+        [commit_hash],
+      ).first[
+        0
+      ].positive?
+    end
+    def insert_commit_files(commit, repo_id)
+      commit.diff.each_patch do |patch|
+        file_path = patch.delta.new_file[:path]
+        @db.execute(
+          "INSERT OR IGNORE INTO files (repo_id, file_path, latest_commit) VALUES (?, ?, ?)",
+          [repo_id, file_path, commit.oid],
+        )
+        file_id = @db.last_insert_row_id
+        @db.execute(
+          "INSERT INTO commit_files (commit_hash, file_id, changes) VALUES (?, ?, ?)",
+          [commit.oid, file_id, patch.to_s],
+        )
+      end
+    end
+    def insert_commit_parents(commit)
+      BranchBase.logger.debug(
+        "Inserting parent commits for repository: #{@repo.path}",
+      )
+      commit.parent_ids.each do |parent_id|
+        @db.execute(
+          "INSERT INTO commit_parents (commit_hash, parent_hash) VALUES (?, ?)",
+          [commit.oid, parent_id],
+        )
+      end
+    end
+    def insert_branches(batched_branches)
+      @db.transaction do
+        batched_branches.each do |data|
+          @db.execute(
+            "INSERT OR IGNORE INTO branches (repo_id, name, head_commit) VALUES (?, ?, ?)",
+            data,
+          )
+        end
+      end
+    end
+    def insert_commits(batched_commits)
+      BranchBase.logger.debug(
+        "Inserting commits for repository ID: #{@repo.path}",
+      )
+      @db.transaction do
+        batched_commits.each do |data|
+          @db.execute(
+            "INSERT INTO commits (commit_hash, repo_id, author, committer, message, timestamp) VALUES (?, ?, ?, ?, ?, ?)",
+            data,
+          )
+        end
+      end
+    end
+    def insert_files_and_commit_files(batched_data)
+      @db.transaction do
+        batched_data.each do |data|
+          repo_id, file_path, commit_hash, changes = data
+          @db.execute(
+            "INSERT OR IGNORE INTO files (repo_id, file_path, latest_commit) VALUES (?, ?, ?)",
+            [repo_id, file_path, commit_hash],
+          )
+          file_id = @db.last_insert_row_id
+          @db.execute(
+            "INSERT INTO commit_files (commit_hash, file_id, changes) VALUES (?, ?, ?)",
+            [commit_hash, file_id, changes],
+          )
+        end
+      end
+    end
+  end
+end

data/lib/branch_base/version.rb ADDED Viewed

@@ -0,0 +1,5 @@
+# frozen_string_literal: true
+module BranchBase
+  VERSION = "0.1.0"
+end

data/lib/branch_base.rb ADDED Viewed

@@ -0,0 +1,97 @@
+# frozen_string_literal: true
+require "logger"
+require "json"
+require "branch_base/version"
+require "branch_base/database"
+require "branch_base/repository"
+require "branch_base/sync"
+require "branch_base/cli"
+module BranchBase
+  def self.logger
+    @logger ||=
+      Logger
+        .new($stdout)
+        .tap do |log|
+          log.progname = "BranchBase"
+          log.level = ENV["DEBUG"] ? Logger::DEBUG : Logger::INFO
+          log.formatter =
+            proc do |severity, datetime, progname, msg|
+              "#{datetime}: #{severity} - #{progname}: #{msg}\n"
+            end
+        end
+  end
+  def self.execute_git_wrapped_queries(database)
+    queries = {
+      "top_contributors_of_the_year" =>
+        "SELECT author, COUNT(*) AS commit_count
+      FROM commits
+      WHERE substr(commits.timestamp, 1, 4) = '2023'
+      GROUP BY author
+      ORDER BY commit_count DESC
+      LIMIT 10;
+      ",
+      "commits_per_day_of_the_week" =>
+        "SELECT
+    CASE strftime('%w', substr(timestamp, 1, 10))
+      WHEN '0' THEN 'Sunday'
+      WHEN '1' THEN 'Monday'
+      WHEN '2' THEN 'Tuesday'
+      WHEN '3' THEN 'Wednesday'
+      WHEN '4' THEN 'Thursday'
+      WHEN '5' THEN 'Friday'
+      WHEN '6' THEN 'Saturday'
+    END as day_of_week,
+    COUNT(*) as commit_count
+  FROM commits
+  WHERE substr(timestamp, 1, 4) = '2023'
+  GROUP BY day_of_week
+  ORDER BY commit_count DESC;
+  ",
+      "most_active_months" =>
+        "SELECT substr(commits.timestamp, 1, 7) AS month, COUNT(*) AS commit_count
+      FROM commits
+      WHERE substr(commits.timestamp, 1, 4) = '2023'
+      GROUP BY month
+      ORDER BY commit_count DESC
+      LIMIT 12;
+      ",
+      "commits_with_most_significant_number_of_changes" =>
+        "SELECT commits.commit_hash, COUNT(commit_files.file_id) AS files_changed
+      FROM commits
+      JOIN commit_files ON commits.commit_hash = commit_files.commit_hash
+      WHERE substr(commits.timestamp, 1, 4) = '2023'
+      GROUP BY commits.commit_hash
+      ORDER BY files_changed DESC
+      LIMIT 10;
+      ",
+      "most_edited_files" =>
+        "SELECT files.file_path, COUNT(*) AS edit_count
+        FROM commit_files
+        JOIN files ON commit_files.file_id = files.file_id
+        JOIN commits ON commit_files.commit_hash = commits.commit_hash
+        WHERE substr(commits.timestamp, 1, 4) = '2023'
+        GROUP BY files.file_path
+        ORDER BY edit_count DESC
+        LIMIT 10;
+        ;
+      ",
+    }
+    queries.transform_values { |query| database.execute(query) }
+  end
+  def self.emojis_for_titles
+    {
+      "top_contributors_of_the_year" => "🏆",
+      "commits_per_day_of_the_week" => "📅",
+      "most_active_months" => "📆",
+      "commits_with_most_significant_number_of_changes" => "📈",
+      "most_edited_files" => "📝",
+    }
+  end
+end

data/package.json ADDED Viewed

@@ -0,0 +1,13 @@
+{
+  "name": "pg-osc",
+  "version": "1.0.0",
+  "main": "index.js",
+  "repository": "git@github.com:shayonj/pg-osc.git",
+  "author": "Shayon Mukherjee <shayonj@gmail.com>",
+  "license": "MIT",
+  "private": true,
+  "dependencies": {
+    "@prettier/plugin-ruby": "^3.2.2",
+    "prettier": "^2.8.8"
+  }
+}

data/scripts/branch_base ADDED Viewed

@@ -0,0 +1,6 @@
+#!/usr/bin/env ruby
+# frozen_string_literal: true
+require "branch_base"
+BranchBase::CLI.start(ARGV)