html-to-markdown 2.9.2 → 2.11.1

data/Rakefile

@@ -1,24 +1,24 @@
-# frozen_string_literal: true
-
-require 'bundler/gem_tasks'
-require 'rb_sys/extensiontask'
-require 'rspec/core/rake_task'
-
-GEMSPEC = Gem::Specification.load(File.expand_path('html-to-markdown-rb.gemspec', __dir__))
-
-RbSys::ExtensionTask.new('html-to-markdown-rb', GEMSPEC) do |ext|
-  ext.lib_dir = 'lib'
-  ext.ext_dir = 'ext/html-to-markdown-rb'
-  ext.cross_compile = true
-  ext.cross_platform = %w[
-    x86_64-linux
-    x86_64-darwin
-    arm64-darwin
-    x64-mingw32
-  ]
-end
-
-RSpec::Core::RakeTask.new(:spec)
-
-task spec: :compile
-task default: :spec
+# frozen_string_literal: true
+
+require 'bundler/gem_tasks'
+require 'rb_sys/extensiontask'
+require 'rspec/core/rake_task'
+
+GEMSPEC = Gem::Specification.load(File.expand_path('html-to-markdown-rb.gemspec', __dir__))
+
+RbSys::ExtensionTask.new('html-to-markdown-rb', GEMSPEC) do |ext|
+  ext.lib_dir = 'lib'
+  ext.ext_dir = 'ext/html-to-markdown-rb'
+  ext.cross_compile = true
+  ext.cross_platform = %w[
+    x86_64-linux
+    x86_64-darwin
+    arm64-darwin
+    x64-mingw32
+  ]
+end
+
+RSpec::Core::RakeTask.new(:spec)
+
+task spec: :compile
+task default: :spec

data/Steepfile

@@ -1,26 +1,26 @@
-# frozen_string_literal: true
-
-# Steepfile for type checking html-to-markdown Ruby gem
-
-target :lib do
-  signature 'sig'
-
-  check 'lib'
-
-  # Configure libraries
-  library 'pathname'
-  library 'open3'
-
-  # Ignore vendor directory
-  ignore 'vendor'
-
-  # Ignore spec directory
-  ignore 'spec'
-
-  # Ignore bin directory
-  ignore 'bin'
-
-  # Ignore internal implementation modules (not public API)
-  ignore 'lib/html_to_markdown/cli.rb'
-  ignore 'lib/html_to_markdown/cli_proxy.rb'
-end
+# frozen_string_literal: true
+
+# Steepfile for type checking html-to-markdown Ruby gem
+
+target :lib do
+  signature 'sig'
+
+  check 'lib'
+
+  # Configure libraries
+  library 'pathname'
+  library 'open3'
+
+  # Ignore vendor directory
+  ignore 'vendor'
+
+  # Ignore spec directory
+  ignore 'spec'
+
+  # Ignore bin directory
+  ignore 'bin'
+
+  # Ignore internal implementation modules (not public API)
+  ignore 'lib/html_to_markdown/cli.rb'
+  ignore 'lib/html_to_markdown/cli_proxy.rb'
+end

data/bin/benchmark.rb

@@ -1,94 +1,94 @@
-#!/usr/bin/env ruby
-# frozen_string_literal: true
-
-require 'optparse'
-require 'time'
-
-$LOAD_PATH.unshift(File.expand_path('../lib', __dir__))
-require 'html_to_markdown'
-
-def json_escape(value)
-  value.to_s.gsub(/["\\\n\r]/) do |char|
-    case char
-    when '"', '\\'
-      "\\#{char}"
-    when "\n"
-      '\\n'
-    when "\r"
-      '\\r'
-    end
-  end
-end
-
-options = {
-  iterations: 50,
-  format: 'html'
-}
-
-OptionParser.new do |parser|
-  parser.banner = 'ruby benchmark.rb --file path/to/fixture.html [--iterations 200]'
-
-  parser.on('--file FILE', 'HTML fixture to convert repeatedly') do |file|
-    options[:file] = file
-  end
-
-  parser.on('--iterations N', Integer, 'Number of conversion iterations (default: 50)') do |n|
-    options[:iterations] = n.positive? ? n : 1
-  end
-
-  parser.on('--format FORMAT', 'Fixture format (html or hocr)') do |format|
-    options[:format] = format.downcase
-  end
-end.parse!
-
-fixture = options.fetch(:file) do
-  warn 'Missing --file parameter'
-  exit 1
-end
-
-unless File.exist?(fixture)
-  warn "Fixture not found: #{fixture}"
-  exit 1
-end
-
-unless %w[html hocr].include?(options[:format])
-  warn "Unsupported format: #{options[:format]}"
-  exit 1
-end
-
-html = File.binread(fixture)
-html.force_encoding(Encoding::UTF_8)
-html.freeze
-iterations = options[:iterations]
-options_handle = HtmlToMarkdown.options(
-  options[:format] == 'hocr' ? { hocr_spatial_tables: false } : nil
-)
-
-def convert_document(html, options_handle)
-  HtmlToMarkdown.convert_with_options(html, options_handle)
-end
-
-convert_document(html, options_handle)
-
-start = Process.clock_gettime(Process::CLOCK_MONOTONIC)
-iterations.times { convert_document(html, options_handle) }
-elapsed = Process.clock_gettime(Process::CLOCK_MONOTONIC) - start
-
-payload_size_bytes = html.bytesize
-bytes_processed = payload_size_bytes * iterations
-ops_per_sec = iterations / elapsed
-mb_per_sec = (bytes_processed.to_f / (1024 * 1024)) / elapsed
-
-payload = %({
-  "language":"ruby",
-  "fixture":"#{json_escape(File.basename(fixture))}",
-  "fixture_path":"#{json_escape(fixture)}",
-  "iterations":#{iterations},
-  "elapsed_seconds":#{format('%.8f', elapsed)},
-  "ops_per_sec":#{format('%.4f', ops_per_sec)},
-  "mb_per_sec":#{format('%.4f', mb_per_sec)},
-  "bytes_processed":#{bytes_processed},
-  "payload_size_bytes":#{payload_size_bytes}
-})
-
-puts payload.strip
+#!/usr/bin/env ruby
+# frozen_string_literal: true
+
+require 'optparse'
+require 'time'
+
+$LOAD_PATH.unshift(File.expand_path('../lib', __dir__))
+require 'html_to_markdown'
+
+def json_escape(value)
+  value.to_s.gsub(/["\\\n\r]/) do |char|
+    case char
+    when '"', '\\'
+      "\\#{char}"
+    when "\n"
+      '\\n'
+    when "\r"
+      '\\r'
+    end
+  end
+end
+
+options = {
+  iterations: 50,
+  format: 'html'
+}
+
+OptionParser.new do |parser|
+  parser.banner = 'ruby benchmark.rb --file path/to/fixture.html [--iterations 200]'
+
+  parser.on('--file FILE', 'HTML fixture to convert repeatedly') do |file|
+    options[:file] = file
+  end
+
+  parser.on('--iterations N', Integer, 'Number of conversion iterations (default: 50)') do |n|
+    options[:iterations] = n.positive? ? n : 1
+  end
+
+  parser.on('--format FORMAT', 'Fixture format (html or hocr)') do |format|
+    options[:format] = format.downcase
+  end
+end.parse!
+
+fixture = options.fetch(:file) do
+  warn 'Missing --file parameter'
+  exit 1
+end
+
+unless File.exist?(fixture)
+  warn "Fixture not found: #{fixture}"
+  exit 1
+end
+
+unless %w[html hocr].include?(options[:format])
+  warn "Unsupported format: #{options[:format]}"
+  exit 1
+end
+
+html = File.binread(fixture)
+html.force_encoding(Encoding::UTF_8)
+html.freeze
+iterations = options[:iterations]
+options_handle = HtmlToMarkdown.options(
+  options[:format] == 'hocr' ? { hocr_spatial_tables: false } : nil
+)
+
+def convert_document(html, options_handle)
+  HtmlToMarkdown.convert_with_options(html, options_handle)
+end
+
+convert_document(html, options_handle)
+
+start = Process.clock_gettime(Process::CLOCK_MONOTONIC)
+iterations.times { convert_document(html, options_handle) }
+elapsed = Process.clock_gettime(Process::CLOCK_MONOTONIC) - start
+
+payload_size_bytes = html.bytesize
+bytes_processed = payload_size_bytes * iterations
+ops_per_sec = iterations / elapsed
+mb_per_sec = (bytes_processed.to_f / (1024 * 1024)) / elapsed
+
+payload = %({
+  "language":"ruby",
+  "fixture":"#{json_escape(File.basename(fixture))}",
+  "fixture_path":"#{json_escape(fixture)}",
+  "iterations":#{iterations},
+  "elapsed_seconds":#{format('%.8f', elapsed)},
+  "ops_per_sec":#{format('%.4f', ops_per_sec)},
+  "mb_per_sec":#{format('%.4f', mb_per_sec)},
+  "bytes_processed":#{bytes_processed},
+  "payload_size_bytes":#{payload_size_bytes}
+})
+
+puts payload.strip

data/exe/html-to-markdown

@@ -1,6 +1,6 @@
-#!/usr/bin/env ruby
-# frozen_string_literal: true
-
-require 'html_to_markdown/cli'
-
-exit HtmlToMarkdown::CLI.run(ARGV)
+#!/usr/bin/env ruby
+# frozen_string_literal: true
+
+require 'html_to_markdown/cli'
+
+exit HtmlToMarkdown::CLI.run(ARGV)

data/ext/html-to-markdown-rb/extconf.rb

@@ -1,38 +1,38 @@
-# frozen_string_literal: true
-
-require 'mkmf'
-require 'rb_sys/mkmf'
-require 'rbconfig'
-require 'pathname'
-
-if RbConfig::CONFIG['host_os'] =~ /mswin|mingw/
-  devkit = ENV.fetch('RI_DEVKIT', nil)
-  prefix = ENV['MSYSTEM_PREFIX'] || '/ucrt64'
-
-  if devkit
-    sysroot = "#{devkit}#{prefix}".tr('\\\\', '/')
-    extra_args = [
-      '--target=x86_64-pc-windows-gnu',
-      "--sysroot=#{sysroot}"
-    ]
-
-    existing = ENV['BINDGEN_EXTRA_CLANG_ARGS'].to_s.split(/\s+/)
-    ENV['BINDGEN_EXTRA_CLANG_ARGS'] = (existing + extra_args).uniq.join(' ')
-  end
-end
-
-default_profile = ENV.fetch('CARGO_PROFILE', 'release')
-
-create_rust_makefile('html_to_markdown_rb') do |config|
-  config.profile = default_profile.to_sym
-
-  native_dir = File.expand_path('native', __dir__)
-  relative_native =
-    begin
-      Pathname.new(native_dir).relative_path_from(Pathname.new(__dir__)).to_s
-    rescue ArgumentError
-      native_dir
-    end
-
-  config.ext_dir = relative_native
-end
+# frozen_string_literal: true
+
+require 'mkmf'
+require 'rb_sys/mkmf'
+require 'rbconfig'
+require 'pathname'
+
+if RbConfig::CONFIG['host_os'] =~ /mswin|mingw/
+  devkit = ENV.fetch('RI_DEVKIT', nil)
+  prefix = ENV['MSYSTEM_PREFIX'] || '/ucrt64'
+
+  if devkit
+    sysroot = "#{devkit}#{prefix}".tr('\\\\', '/')
+    extra_args = [
+      '--target=x86_64-pc-windows-gnu',
+      "--sysroot=#{sysroot}"
+    ]
+
+    existing = ENV['BINDGEN_EXTRA_CLANG_ARGS'].to_s.split(/\s+/)
+    ENV['BINDGEN_EXTRA_CLANG_ARGS'] = (existing + extra_args).uniq.join(' ')
+  end
+end
+
+default_profile = ENV.fetch('CARGO_PROFILE', 'release')
+
+create_rust_makefile('html_to_markdown_rb') do |config|
+  config.profile = default_profile.to_sym
+
+  native_dir = File.expand_path('native', __dir__)
+  relative_native =
+    begin
+      Pathname.new(native_dir).relative_path_from(Pathname.new(__dir__)).to_s
+    rescue ArgumentError
+      native_dir
+    end
+
+  config.ext_dir = relative_native
+end

data/ext/html-to-markdown-rb/native/Cargo.toml

@@ -1,28 +1,28 @@
-[package]
-name = "html-to-markdown-rb"
-version = "2.9.2"
-edition = "2024"
-authors = ["Na'aman Hirschfeld <nhirschfeld@gmail.com>"]
-license = "MIT"
-repository = "https://github.com/Goldziher/html-to-markdown"
-homepage = "https://github.com/Goldziher/html-to-markdown"
-documentation = "https://docs.rs/html-to-markdown-rs"
-readme = "README.md"
-rust-version = "1.85"
-description = "Ruby bindings (Magnus) for html-to-markdown - high-performance HTML to Markdown converter"
-keywords = ["html", "markdown", "ruby", "magnus", "bindings"]
-categories = ["api-bindings"]
-
-[lib]
-name = "html_to_markdown_rb"
-crate-type = ["cdylib", "rlib"]
-
-[features]
-default = []
-
-[dependencies]
-html-to-markdown-rs = { version = "2.9.2", features = ["inline-images"] }
-magnus = { git = "https://github.com/matsadler/magnus", rev = "f6db11769efb517427bf7f121f9c32e18b059b38", features = ["rb-sys"] }
-
-[dev-dependencies]
-pretty_assertions = "1.4"
+[package]
+name = "html-to-markdown-rb"
+version = "2.11.1"
+edition = "2024"
+authors = ["Na'aman Hirschfeld <nhirschfeld@gmail.com>"]
+license = "MIT"
+repository = "https://github.com/Goldziher/html-to-markdown"
+homepage = "https://github.com/Goldziher/html-to-markdown"
+documentation = "https://docs.rs/html-to-markdown-rs"
+readme = "README.md"
+rust-version = "1.85"
+description = "Ruby bindings (Magnus) for html-to-markdown - high-performance HTML to Markdown converter"
+keywords = ["html", "markdown", "ruby", "magnus", "bindings"]
+categories = ["api-bindings"]
+
+[lib]
+name = "html_to_markdown_rb"
+crate-type = ["cdylib", "rlib"]
+
+[features]
+default = []
+
+[dependencies]
+html-to-markdown-rs = { version = "2.11.1", features = ["inline-images"] }
+magnus = { git = "https://github.com/matsadler/magnus", rev = "f6db11769efb517427bf7f121f9c32e18b059b38", features = ["rb-sys"] }
+
+[dev-dependencies]
+pretty_assertions = "1.4"