serialbench 0.3.1 → 0.5.0

data/lib/serialbench/test_data.rb

@@ -0,0 +1,315 @@
+# frozen_string_literal: true
+
+module Serialbench
+  # Fixture documents for every format x size the benchmark engine measures.
+  class TestData
+    def self.load(config)
+      new(config).load
+    end
+
+    def initialize(config)
+      @config = config
+    end
+
+    def load
+      @config.data_sizes.each_with_object({}) do |size, data|
+        data[size] = @config.formats.each_with_object({}) do |format, formats|
+          formats[format] = fixture_file(size, format) || generate("#{size}.#{format}")
+        end
+      end
+    end
+
+    private
+
+    def fixture_file(size, format)
+      path = "test_data/#{size}.#{format}"
+      File.read(path) if File.exist?(path)
+    end
+
+    def generate(key)
+      case key
+      when 'small.xml' then generate_small_xml
+      when 'medium.xml' then generate_medium_xml
+      when 'large.xml' then generate_large_xml
+      when 'small.json' then generate_small_json
+      when 'medium.json' then generate_medium_json
+      when 'large.json' then generate_large_json
+      when 'small.yaml' then generate_small_yaml
+      when 'medium.yaml' then generate_medium_yaml
+      when 'large.yaml' then generate_large_yaml
+      when 'small.toml' then generate_small_toml
+      when 'medium.toml' then generate_medium_toml
+      when 'large.toml' then generate_large_toml
+      else raise ArgumentError, "no test data generator for #{key}"
+      end
+    end
+
+    # Shared data structure generators
+    def small_test_data_structure
+      {
+        config: {
+          database: {
+            host: 'localhost',
+            port: 5432,
+            name: 'myapp',
+            user: 'admin',
+            password: 'secret'
+          },
+          cache: {
+            enabled: true,
+            ttl: 3600
+          }
+        }
+      }
+    end
+
+    def medium_test_data_structure
+      {
+        users: (1..1000).map do |i|
+          {
+            id: i,
+            name: "User #{i}",
+            email: "user#{i}@example.com",
+            created_at: "2023-01-#{(i % 28) + 1}T10:00:00Z",
+            profile: {
+              age: 20 + (i % 50),
+              city: "City #{i % 100}",
+              preferences: {
+                theme: i.even? ? 'dark' : 'light',
+                notifications: (i % 3).zero?
+              }
+            }
+          }
+        end
+      }
+    end
+
+    def large_test_data_structure
+      {
+        dataset: {
+          header: {
+            created: '2023-01-01T00:00:00Z',
+            count: 10_000,
+            format: 'data'
+          },
+          records: (1..10_000).map do |i|
+            {
+              id: i,
+              timestamp: "2023-01-01T#{format('%02d', i % 24)}:#{format('%02d', i % 60)}:#{format('%02d', i % 60)}Z",
+              data: {
+                field1: "Value #{i}",
+                field2: i * 2,
+                field3: (i % 100).zero? ? 'special' : 'normal',
+                nested: [
+                  "Item #{i}-1",
+                  "Item #{i}-2",
+                  "Item #{i}-3"
+                ]
+              },
+              metadata: {
+                source: 'generator',
+                version: '1.0',
+                checksum: i.to_s(16)
+              }
+            }
+          end
+        }
+      }
+    end
+
+    # XML test data generators
+    def generate_small_xml
+      data = small_test_data_structure
+      <<~XML
+        <?xml version="1.0" encoding="UTF-8"?>
+        <config>
+          <database>
+            <host>#{data[:config][:database][:host]}</host>
+            <port>#{data[:config][:database][:port]}</port>
+            <name>#{data[:config][:database][:name]}</name>
+            <user>#{data[:config][:database][:user]}</user>
+            <password>#{data[:config][:database][:password]}</password>
+          </database>
+          <cache>
+            <enabled>#{data[:config][:cache][:enabled]}</enabled>
+            <ttl>#{data[:config][:cache][:ttl]}</ttl>
+          </cache>
+        </config>
+      XML
+    end
+
+    def generate_medium_xml
+      data = medium_test_data_structure
+      users = data[:users].map do |user|
+        <<~USER
+          <user id="#{user[:id]}">
+            <name>#{user[:name]}</name>
+            <email>#{user[:email]}</email>
+            <created_at>#{user[:created_at]}</created_at>
+            <profile>
+              <age>#{user[:profile][:age]}</age>
+              <city>#{user[:profile][:city]}</city>
+              <preferences>
+                <theme>#{user[:profile][:preferences][:theme]}</theme>
+                <notifications>#{user[:profile][:preferences][:notifications]}</notifications>
+              </preferences>
+            </profile>
+          </user>
+        USER
+      end.join
+
+      <<~XML
+        <?xml version="1.0" encoding="UTF-8"?>
+        <users>
+          #{users}
+        </users>
+      XML
+    end
+
+    def generate_large_xml
+      data = large_test_data_structure
+      records = data[:dataset][:records].map do |record|
+        nested_items = record[:data][:nested].map { |item| "    <item>#{item}</item>" }.join("\n")
+        <<~RECORD
+            <record id="#{record[:id]}">
+              <timestamp>#{record[:timestamp]}</timestamp>
+              <data>
+                <field1>#{record[:data][:field1]}</field1>
+                <field2>#{record[:data][:field2]}</field2>
+                <field3>#{record[:data][:field3]}</field3>
+                <nested>
+          #{nested_items}
+                </nested>
+              </data>
+              <metadata>
+                <source>#{record[:metadata][:source]}</source>
+                <version>#{record[:metadata][:version]}</version>
+                <checksum>#{record[:metadata][:checksum]}</checksum>
+              </metadata>
+            </record>
+        RECORD
+      end.join
+
+      <<~XML
+        <?xml version="1.0" encoding="UTF-8"?>
+        <dataset>
+          <header>
+            <created>#{data[:dataset][:header][:created]}</created>
+            <count>#{data[:dataset][:header][:count]}</count>
+            <format>xml</format>
+          </header>
+          <records>
+            #{records}
+          </records>
+        </dataset>
+      XML
+    end
+
+    # JSON test data generators
+    def generate_small_json
+      JSON.generate(small_test_data_structure)
+    end
+
+    def generate_medium_json
+      JSON.generate(medium_test_data_structure)
+    end
+
+    def generate_large_json
+      data = large_test_data_structure
+      data[:dataset][:header][:format] = 'json'
+      JSON.generate(data)
+    end
+
+    # YAML test data generators
+    def generate_small_yaml
+      small_test_data_structure.to_yaml
+    end
+
+    def generate_medium_yaml
+      medium_test_data_structure.to_yaml
+    end
+
+    def generate_large_yaml
+      data = large_test_data_structure
+      data[:dataset][:header][:format] = 'yaml'
+      data.to_yaml
+    end
+
+    # TOML test data generators
+    def generate_small_toml
+      data = small_test_data_structure
+      <<~TOML
+        [config]
+
+        [config.database]
+        host = "#{data[:config][:database][:host]}"
+        port = #{data[:config][:database][:port]}
+        name = "#{data[:config][:database][:name]}"
+        user = "#{data[:config][:database][:user]}"
+        password = "#{data[:config][:database][:password]}"
+
+        [config.cache]
+        enabled = #{data[:config][:cache][:enabled]}
+        ttl = #{data[:config][:cache][:ttl]}
+      TOML
+    end
+
+    def generate_medium_toml
+      data = medium_test_data_structure
+      # Use smaller dataset for TOML due to verbosity
+      users = data[:users].first(100)
+      users.map do |user|
+        <<~USER
+          [[users]]
+          id = #{user[:id]}
+          name = "#{user[:name]}"
+          email = "#{user[:email]}"
+          created_at = "#{user[:created_at]}"
+
+          [users.profile]
+          age = #{user[:profile][:age]}
+          city = "#{user[:profile][:city]}"
+
+          [users.profile.preferences]
+          theme = "#{user[:profile][:preferences][:theme]}"
+          notifications = #{user[:profile][:preferences][:notifications]}
+        USER
+      end.join("\n")
+    end
+
+    def generate_large_toml
+      data = large_test_data_structure
+      # Use smaller dataset for TOML due to verbosity
+      records = data[:dataset][:records].first(1000)
+      records_toml = records.map do |record|
+        <<~RECORD
+          [[dataset.records]]
+          id = #{record[:id]}
+          timestamp = "#{record[:timestamp]}"
+
+          [dataset.records.data]
+          field1 = "#{record[:data][:field1]}"
+          field2 = #{record[:data][:field2]}
+          field3 = "#{record[:data][:field3]}"
+          nested = #{record[:data][:nested].inspect}
+
+          [dataset.records.metadata]
+          source = "#{record[:metadata][:source]}"
+          version = "#{record[:metadata][:version]}"
+          checksum = "#{record[:metadata][:checksum]}"
+        RECORD
+      end.join("\n")
+
+      <<~TOML
+        [dataset]
+
+        [dataset.header]
+        created = "#{data[:dataset][:header][:created]}"
+        count = #{records.length}
+        format = "toml"
+
+        #{records_toml}
+      TOML
+    end
+  end
+end

data/lib/serialbench/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Serialbench
-  VERSION = '0.3.1'
+  VERSION = '0.5.0'
 end

data/lib/serialbench.rb

@@ -1,7 +1,6 @@
 # frozen_string_literal: true
 
 require_relative 'serialbench/version'
-require_relative 'serialbench/site_generator'
 require_relative 'serialbench/serializers'
 require_relative 'serialbench/benchmark_runner'
 require_relative 'serialbench/cli'
@@ -9,5 +8,7 @@ require_relative 'serialbench/memory_profiler'
 require_relative 'serialbench/models'
 
 module Serialbench
+  autoload :TestData, 'serialbench/test_data'
+
   class Error < StandardError; end
 end

data/serialbench.gemspec

@@ -14,7 +14,7 @@ Gem::Specification.new do |spec|
 
   spec.summary = 'Comprehensive serialization benchmarking tool for Ruby'
   spec.description = 'A benchmarking suite for comparing performance of various serialization libraries in Ruby, including XML, JSON, and TOML parsers/generators.'
-  spec.homepage = 'https://github.com/metanorma/serialbench'
+  spec.homepage = 'https://github.com/serialbench/serialbench'
   spec.license = 'BSD-2-Clause'
   spec.required_ruby_version = Gem::Requirement.new('>= 2.7.0')
 
@@ -42,7 +42,6 @@ Gem::Specification.new do |spec|
 
   # Reporting dependencies
   spec.add_dependency 'csv' # Required from Ruby 3.4+
-  spec.add_dependency 'liquid' # Template engine for HTML reports
 
   # XML serializers
   spec.add_dependency 'oga'

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: serialbench
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.5.0
 platform: ruby
 authors:
 - Ribose
@@ -93,20 +93,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: liquid
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
 - !ruby/object:Gem::Dependency
   name: oga
   requirement: !ruby/object:Gem::Requirement
@@ -274,7 +260,6 @@ files:
 - ".github/workflows/benchmark.yml"
 - ".github/workflows/rake.yml"
 - ".github/workflows/release.yml"
-- ".github/workflows/windows-debug.yml"
 - ".gitignore"
 - ".rspec"
 - ".rubocop.yml"
@@ -311,7 +296,6 @@ files:
 - lib/serialbench/cli/base_cli.rb
 - lib/serialbench/cli/benchmark_cli.rb
 - lib/serialbench/cli/environment_cli.rb
-- lib/serialbench/cli/resultset_cli.rb
 - lib/serialbench/cli/ruby_build_cli.rb
 - lib/serialbench/cli/validate_cli.rb
 - lib/serialbench/config_manager.rb
@@ -322,7 +306,6 @@ files:
 - lib/serialbench/models/environment_config.rb
 - lib/serialbench/models/platform.rb
 - lib/serialbench/models/result.rb
-- lib/serialbench/models/result_set.rb
 - lib/serialbench/models/result_store.rb
 - lib/serialbench/ruby_build_manager.rb
 - lib/serialbench/runners/asdf_runner.rb
@@ -350,17 +333,17 @@ files:
 - lib/serialbench/serializers/yaml/base_yaml_serializer.rb
 - lib/serialbench/serializers/yaml/psych_serializer.rb
 - lib/serialbench/serializers/yaml/syck_serializer.rb
-- lib/serialbench/site_generator.rb
+- lib/serialbench/test_data.rb
 - lib/serialbench/version.rb
 - lib/serialbench/yaml_validator.rb
 - serialbench.gemspec
-homepage: https://github.com/metanorma/serialbench
+homepage: https://github.com/serialbench/serialbench
 licenses:
 - BSD-2-Clause
 metadata:
-  homepage_uri: https://github.com/metanorma/serialbench
-  source_code_uri: https://github.com/metanorma/serialbench
-  bug_tracker_uri: https://github.com/metanorma/serialbench/issues
+  homepage_uri: https://github.com/serialbench/serialbench
+  source_code_uri: https://github.com/serialbench/serialbench
+  bug_tracker_uri: https://github.com/serialbench/serialbench/issues
 rdoc_options: []
 require_paths:
 - lib

data/.github/workflows/windows-debug.yml

@@ -1,171 +0,0 @@
-name: Windows Debug
-
-on:
-  push:
-    branches: [ master, main ]
-    paths:
-      - '.github/workflows/windows-debug.yml'
-  pull_request:
-    paths:
-      - '.github/workflows/windows-debug.yml'
-  workflow_dispatch:
-
-jobs:
-  test-libxml2-install:
-    runs-on: windows-latest
-    strategy:
-      fail-fast: false
-      matrix:
-        ruby: ['3.1', '3.4']
-        method: ['vcpkg', 'msys2', 'direct']
-
-    name: Test ${{ matrix.method }} on Ruby ${{ matrix.ruby }}
-
-    steps:
-      - uses: actions/checkout@v4
-
-      - name: Set up Ruby
-        uses: ruby/setup-ruby@v1
-        with:
-          ruby-version: ${{ matrix.ruby }}
-
-      - name: Install via vcpkg
-        if: matrix.method == 'vcpkg'
-        shell: powershell
-        run: |
-          Write-Host "Installing libxml2 via vcpkg..."
-          vcpkg install libxml2:x64-windows
-          vcpkg integrate install
-
-          # Find the libxml2 installation
-          $vcpkgRoot = "C:\vcpkg"
-          $libxml2Path = Join-Path $vcpkgRoot "installed\x64-windows"
-
-          Write-Host "libxml2 installed at: $libxml2Path"
-
-          # Set environment variables
-          echo "LIBXML2_INCLUDE=$libxml2Path\include" >> $env:GITHUB_ENV
-          echo "LIBXML2_LIB=$libxml2Path\lib" >> $env:GITHUB_ENV
-          echo "PKG_CONFIG_PATH=$libxml2Path\lib\pkgconfig" >> $env:GITHUB_ENV
-
-          # Add to PATH
-          $currentPath = [Environment]::GetEnvironmentVariable("PATH", "Process")
-          $newPath = "$libxml2Path\bin;$currentPath"
-          echo "$newPath" >> $env:GITHUB_PATH
-
-          # Verify installation
-          Get-ChildItem -Path $libxml2Path -Recurse | Where-Object { $_.Name -like "*libxml*" } | Select-Object FullName
-
-      - name: Install via msys2
-        if: matrix.method == 'msys2'
-        shell: powershell
-        run: |
-          Write-Host "Installing libxml2 via msys2..."
-          choco install -y msys2
-
-          # Initialize msys2 and install libxml2
-          C:\tools\msys64\usr\bin\bash -lc "pacman -Syu --noconfirm"
-          C:\tools\msys64\usr\bin\bash -lc "pacman -S --noconfirm mingw-w64-x86_64-libxml2 mingw-w64-x86_64-pkg-config"
-
-          # Set environment variables
-          $msys2Root = "C:\tools\msys64\mingw64"
-          echo "LIBXML2_INCLUDE=$msys2Root\include\libxml2" >> $env:GITHUB_ENV
-          echo "LIBXML2_LIB=$msys2Root\lib" >> $env:GITHUB_ENV
-          echo "PKG_CONFIG_PATH=$msys2Root\lib\pkgconfig" >> $env:GITHUB_ENV
-
-          # Add to PATH
-          echo "$msys2Root\bin" >> $env:GITHUB_PATH
-
-          # Verify installation
-          Get-ChildItem -Path $msys2Root -Recurse | Where-Object { $_.Name -like "*libxml*" } | Select-Object FullName
-
-      - name: Install via direct download
-        if: matrix.method == 'direct'
-        shell: powershell
-        run: |
-          Write-Host "Installing libxml2 via direct download..."
-
-          # Download pre-built binaries from gnome.org
-          $downloadDir = "$env:TEMP\libxml2"
-          New-Item -ItemType Directory -Force -Path $downloadDir
-
-          # Download libxml2 and dependencies
-          $libxml2Url = "https://download.gnome.org/binaries/win64/libxml2/2.9/libxml2-2.9.14-x64_dll.zip"
-          $zlibUrl = "https://download.gnome.org/binaries/win64/dependencies/zlib-1.2.13-x64_dll.zip"
-          $iconvUrl = "https://download.gnome.org/binaries/win64/dependencies/libiconv-1.17-x64_dll.zip"
-
-          Invoke-WebRequest -Uri $libxml2Url -OutFile "$downloadDir\libxml2.zip"
-          Invoke-WebRequest -Uri $zlibUrl -OutFile "$downloadDir\zlib.zip"
-          Invoke-WebRequest -Uri $iconvUrl -OutFile "$downloadDir\iconv.zip"
-
-          # Extract archives
-          Expand-Archive -Path "$downloadDir\libxml2.zip" -DestinationPath "C:\libxml2" -Force
-          Expand-Archive -Path "$downloadDir\zlib.zip" -DestinationPath "C:\libxml2" -Force
-          Expand-Archive -Path "$downloadDir\iconv.zip" -DestinationPath "C:\libxml2" -Force
-
-          # Set environment variables
-          echo "LIBXML2_INCLUDE=C:\libxml2\include" >> $env:GITHUB_ENV
-          echo "LIBXML2_LIB=C:\libxml2\lib" >> $env:GITHUB_ENV
-
-          # Add to PATH
-          echo "C:\libxml2\bin" >> $env:GITHUB_PATH
-
-          # Verify installation
-          Get-ChildItem -Path "C:\libxml2" -Recurse | Where-Object { $_.Name -like "*libxml*" } | Select-Object FullName
-
-      - name: Verify pkg-config
-        shell: powershell
-        run: |
-          Write-Host "Checking for pkg-config..."
-          try {
-            $pkgConfigPath = Get-Command pkg-config -ErrorAction Stop
-            Write-Host "pkg-config found at: $($pkgConfigPath.Source)"
-            pkg-config --version
-          } catch {
-            Write-Host "pkg-config not found, installing pkgconfiglite..."
-            choco install -y pkgconfiglite
-            pkg-config --version
-          }
-
-          # Try to find libxml2 via pkg-config
-          Write-Host "`nSearching for libxml-2.0.pc..."
-          if ($env:PKG_CONFIG_PATH) {
-            Write-Host "PKG_CONFIG_PATH: $env:PKG_CONFIG_PATH"
-            pkg-config --list-all | Select-String libxml
-          }
-
-      - name: Install bundler
-        run: gem install bundler
-
-      - name: Try to install libxml-ruby gem
-        shell: powershell
-        continue-on-error: true
-        run: |
-          Write-Host "Attempting to install libxml-ruby..."
-
-          # Set additional environment variables for gem installation
-          if ($env:LIBXML2_INCLUDE) {
-            $env:CFLAGS = "-I$env:LIBXML2_INCLUDE"
-            $env:CPPFLAGS = "-I$env:LIBXML2_INCLUDE"
-          }
-          if ($env:LIBXML2_LIB) {
-            $env:LDFLAGS = "-L$env:LIBXML2_LIB"
-          }
-
-          gem install libxml-ruby --verbose
-
-      - name: Report results
-        if: always()
-        shell: powershell
-        run: |
-          Write-Host "`n=== Installation Method: ${{ matrix.method }} ==="
-          Write-Host "Ruby Version: ${{ matrix.ruby }}"
-          Write-Host "`nEnvironment Variables:"
-          Write-Host "LIBXML2_INCLUDE: $env:LIBXML2_INCLUDE"
-          Write-Host "LIBXML2_LIB: $env:LIBXML2_LIB"
-          Write-Host "PKG_CONFIG_PATH: $env:PKG_CONFIG_PATH"
-          Write-Host "`nPATH:"
-          $env:PATH -split ';' | ForEach-Object { Write-Host "  $_" }
-
-          Write-Host "`nChecking for libxml2 files..."
-          Get-ChildItem C:\ -Recurse -ErrorAction SilentlyContinue -Include libxml2.dll,libxml2.lib,xml2-config.h 2>$null | Select-Object FullName