rsyntaxtree 1.3.2 → 1.4.0

data/lib/rsyntaxtree/tikz_generator.rb

@@ -0,0 +1,131 @@
+# frozen_string_literal: true
+
+#==========================
+# tikz_generator.rb
+#==========================
+#
+# Generates TikZ/forest LaTeX code from parsed tree elements
+# Copyright (c) 2007-2026 Yoichiro Hasebe <yohasebe@gmail.com>
+
+module RSyntaxTree
+  class TikZGenerator
+    LATEX_ESCAPE_MAP = {
+      '&' => '\\&',
+      '%' => '\\%',
+      '$' => '\\$',
+      '#' => '\\#',
+      '_' => '\\_',
+      '{' => '\\{',
+      '}' => '\\}',
+      '~' => '\\textasciitilde{}',
+      '^' => '\\textasciicircum{}'
+    }.freeze
+
+    def initialize(element_list, params)
+      @element_list = element_list
+      @params = params
+    end
+
+    # Generate TikZ forest code
+    # @param standalone [Boolean] whether to include LaTeX preamble
+    # @param font [String, nil] font name for XeLaTeX/fontspec (enables fontspec when specified)
+    # @return [String] TikZ/forest code
+    def generate(standalone: false, font: nil)
+      tree_code = generate_tree(1)
+
+      if standalone
+        generate_standalone(tree_code, font: font)
+      else
+        generate_forest_only(tree_code)
+      end
+    end
+
+    private
+
+    def generate_standalone(tree_code, font: nil)
+      if font
+        <<~LATEX
+          \\documentclass[border=10pt]{standalone}
+          \\usepackage{forest}
+          \\usepackage{fontspec}
+          \\setmainfont{#{font}}
+
+          \\begin{document}
+          #{generate_forest_only(tree_code)}
+          \\end{document}
+        LATEX
+      else
+        <<~LATEX
+          \\documentclass[border=10pt]{standalone}
+          \\usepackage{forest}
+
+          \\begin{document}
+          #{generate_forest_only(tree_code)}
+          \\end{document}
+        LATEX
+      end
+    end
+
+    def generate_forest_only(tree_code)
+      <<~LATEX
+        \\begin{forest}
+        for tree={
+          parent anchor=south,
+          child anchor=north,
+          align=center,
+          base=top
+        }
+        #{tree_code}
+        \\end{forest}
+      LATEX
+    end
+
+    # Recursively generate tree structure
+    def generate_tree(element_id, indent = 0)
+      element = @element_list.get_id(element_id)
+      return "" unless element
+
+      label = extract_label(element)
+      escaped_label = escape_latex(label)
+
+      children = element.children
+      indent_str = "  " * indent
+
+      if children.empty?
+        "#{indent_str}[#{escaped_label}]"
+      else
+        child_code = children.map { |child_id| generate_tree(child_id, indent + 1) }.join("\n")
+        "#{indent_str}[#{escaped_label}\n#{child_code}\n#{indent_str}]"
+      end
+    end
+
+    # Extract plain text label from element content
+    def extract_label(element)
+      content = element.content
+      return "" if content.nil? || content.empty?
+
+      # content is an array of hashes with :type and :elements
+      texts = []
+      content.each do |line|
+        next unless line[:type] == :text
+
+        line[:elements].each do |el|
+          text = el[:text].to_s
+          # Remove RSyntaxTree-specific whitespace block
+          text = text.gsub("￭", " ")
+          texts << text unless text.strip.empty?
+        end
+      end
+      texts.join(" ")
+    end
+
+    # Escape LaTeX special characters
+    def escape_latex(text)
+      result = text.dup
+      LATEX_ESCAPE_MAP.each do |char, escaped|
+        result.gsub!(char, escaped)
+      end
+      result
+    end
+  end
+end

data/lib/rsyntaxtree/utils.rb

@@ -5,7 +5,7 @@
 #==========================
 #
 # Image utility functions to inspect text font metrics
-# Copyright (c) 2007-2024 Yoichiro Hasebe <yohasebe@gmail.com>
+# Copyright (c) 2007-2026 Yoichiro Hasebe <yohasebe@gmail.com>
 
 require 'rmagick'
 

data/lib/rsyntaxtree/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module RSyntaxTree
-  VERSION = "1.3.2"
+  VERSION = "1.4.0"
 end

data/lib/rsyntaxtree.rb

@@ -6,7 +6,7 @@
 #
 # Facade of rsyntaxtree library.  When loaded by a driver script, it does all
 # the necessary 'require' to use the library.
-# Copyright (c) 2007-2024 Yoichiro Hasebe <yohasebe@gmail.com>
+# Copyright (c) 2007-2026 Yoichiro Hasebe <yohasebe@gmail.com>
 
 FONT_DIR = File.expand_path(File.join(__dir__, "/../fonts"))
 ETYPE_NODE = 1
@@ -24,7 +24,7 @@ DEFAULT_OPTS = {
   linewidth: 1,
   vheight: 2.0,
   color: "modern",
-  symmetrize: "on",
+  symmetrize: "off",
   transparent: "off",
   polyline: "off",
   hide_default_connectors: "off"
@@ -41,6 +41,7 @@ require_relative 'rsyntaxtree/utils'
 require_relative 'rsyntaxtree/element'
 require_relative 'rsyntaxtree/elementlist'
 require_relative 'rsyntaxtree/svg_graph'
+require_relative 'rsyntaxtree/tikz_generator'
 require_relative 'rsyntaxtree/version'
 require_relative 'rsyntaxtree/string_parser'
 
@@ -152,6 +153,9 @@ module RSyntaxTree
     end
 
     def draw_png(binary = false)
+      surface = nil
+      context = nil
+      b = nil
       svg = draw_svg
       rsvg = RSVG::Handle.new_from_data(svg)
       dim = rsvg.dimensions
@@ -160,17 +164,19 @@ module RSyntaxTree
       context.render_rsvg_handle(rsvg)
       b = StringIO.new
       surface.write_to_png(b)
-      result = binary ? b : b.string
-      # Clean up resources
-      b.close unless binary
-      surface.finish
-      context.destroy
-      result
+      binary ? b : b.string
     rescue Cairo::InvalidSize
       raise RSTError, +"Error: the result syntree is too big"
+    ensure
+      b&.close unless binary
+      surface&.finish
+      context&.destroy
     end
 
     def draw_pdf(binary = false)
+      surface = nil
+      context = nil
+      b = nil
       b = StringIO.new
       svg = draw_svg
       rsvg = RSVG::Handle.new_from_data(svg)
@@ -179,13 +185,12 @@ module RSyntaxTree
       context = Cairo::Context.new(surface)
       context.render_rsvg_handle(rsvg)
       surface.finish
-      result = binary ? b : b.string
-      # Clean up resources
-      b.close unless binary
-      context.destroy
-      result
+      binary ? b : b.string
     rescue Cairo::InvalidSize
       raise RSTError, +"Error: the result syntree is too big"
+    ensure
+      b&.close unless binary
+      context&.destroy
     end
 
     def draw_svg
@@ -195,18 +200,31 @@ module RSyntaxTree
       graph.svg_data
     end
 
-    # Currently not used
-    def draw_tree
-      svg = draw_svg
-      image, _data = Magick::Image.from_blob(svg) do |im|
-        im.format = 'svg'
-      end
-      blob = image.to_blob do |im|
-        im.format = @params[:format].upcase
-      end
-      # Clean up resources
-      image.destroy!
+    def draw_jpg
+      png_data = draw_png
+      images = Magick::Image.from_blob(png_data)
+      image = images.first
+      image.format = 'JPEG'
+      blob = image.to_blob
+      images.each(&:destroy!)
       blob
     end
+
+    def draw_gif
+      png_data = draw_png
+      images = Magick::Image.from_blob(png_data)
+      image = images.first
+      image.format = 'GIF'
+      blob = image.to_blob
+      images.each(&:destroy!)
+      blob
+    end
+
+    def draw_tikz(standalone: false, font: nil)
+      sp = StringParser.new(@params[:data].gsub('&', '&amp;'), @params[:fontset], @params[:fontsize], @global)
+      sp.parse
+      generator = TikZGenerator.new(sp.get_elementlist, @params)
+      generator.generate(standalone: standalone, font: font)
+    end
   end
 end

data/rsyntaxtree.gemspec

@@ -22,6 +22,8 @@ Gem::Specification.new do |s|
   s.add_runtime_dependency "rmagick", ">= 4.3"
   s.add_runtime_dependency "rsvg2"
 
+  s.add_development_dependency "minitest"
   s.add_development_dependency "nokogiri"
+  s.add_development_dependency "rake"
   s.add_development_dependency "yaml"
 end

data/syntree.svg

@@ -0,0 +1,41 @@
+<?xml version="1.0" standalone="no"?>
+<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN" "http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd">
+  <svg width="118.60000000000001" height="434.25" viewBox="-15.200000000000001, 0, 133.8, 445.5" version="1.1" xmlns="http://www.w3.org/2000/svg">
+  <defs>
+    <marker id="arrow" markerUnits="userSpaceOnUse" viewBox="0 0 10 10" refX="10" refY="5" markerWidth="15.200000000000001" markerHeight="15.200000000000001" orient="auto">
+      <path d="M 0 0 L 10 5 L 0 10" fill="#CC79A7"/>
+    </marker>
+    <marker id="arrowBackward" markerUnits="userSpaceOnUse" viewBox="0 0 10 10" refX="5" refY="5" markerWidth="15.200000000000001" markerHeight="15.200000000000001" orient="auto">
+      <path d="M 0 0 L 10 5 L 0 10 z" fill="#CC79A7"/>
+    </marker>
+    <marker id="arrowForward" markerUnits="userSpaceOnUse" viewBox="0 0 10 10" refX="5" refY="5" markerWidth="15.200000000000001" markerHeight="15.200000000000001" orient="auto">
+      <path d="M 10 0 L 0 5 L 10 10 z" fill="#CC79A7"/>
+    </marker>
+    <marker id="arrowBothways" markerUnits="userSpaceOnUse" viewBox="0 0 30 10" refX="15" refY="5" markerWidth="45.6" markerHeight="15.200000000000001" orient="auto">
+      <path d="M 0 5 L 10 0 L 10 5 L 20 5 L 20 0 L 30 5 L 20 10 L 20 5 L 10 5 L 10 10 z" fill="#CC79A7"/>
+    </marker>
+    <pattern id="hatchBlack" x="10" y="10" width="10" height="10" patternUnits="userSpaceOnUse" patternTransform="rotate(45)">
+      <line x1="0" y="0" x2="0" y2="10" stroke="black" stroke-width="4"></line>
+    </pattern>
+    <pattern id="hatchForNode" x="10" y="10" width="10" height="10" patternUnits="userSpaceOnUse" patternTransform="rotate(45)">
+      <line x1="0" y="0" x2="0" y2="10" stroke="#0072B2" stroke-width="4"></line>
+    </pattern>
+    <pattern id="hatchForLeaf" x="10" y="10" width="10" height="10" patternUnits="userSpaceOnUse" patternTransform="rotate(45)">
+      <line x1="0" y="0" x2="0" y2="10" stroke="#009E73" stroke-width="4"></line>
+    </pattern>
+  </defs>
+<rect x="-15.200000000000001" y="0" width="133.8" height="445.5" stroke="none" fill="white" />"
+<text white-space='pre' alignment-baseline='text-top' style='fill: #0072B2; storoke-width: 0; font-size: 32px;' x='42.7' y='67.5'><tspan x='42.7' y='67.5' style="" text-decoration="" font-family="'Noto Sans', 'Noto Sans JP', OpenMoji, 'OpenMoji Color', 'OpenMoji Black', sans-serif">S</tspan>
+</text>
+<text white-space='pre' alignment-baseline='text-top' style='fill: red; storoke-width: 0; font-size: 32px;' x='30.200000000000003' y='238.5'><tspan x='30.200000000000003' y='238.5' style="" text-decoration="" font-family="'Noto Sans', 'Noto Sans JP', OpenMoji, 'OpenMoji Color', 'OpenMoji Black', sans-serif">NP</tspan>
+</text>
+<text white-space='pre' alignment-baseline='text-top' style='fill: #009E73; storoke-width: 0; font-size: 32px;' x='15.200000000000003' y='409.5'><tspan x='15.200000000000003' y='409.5' style="" text-decoration="" font-family="'Noto Sans', 'Noto Sans JP', OpenMoji, 'OpenMoji Color', 'OpenMoji Black', sans-serif">hello</tspan>
+</text>
+<line style='fill: none; stroke:black; stroke-width:2; stroke-linejoin:round; stroke-linecap:round;' x1='51.7' y1='182.25' x2='51.7' y2='95.625' />
+<line style='fill: none; stroke:black; stroke-width:2; stroke-linejoin:round; stroke-linecap:round;' x1='51.7' y1='353.25' x2='51.7' y2='266.625' />
+<polyline style='stroke:red; stroke-width:2; fill:none; stroke-linejoin:round; stroke-linecap:round;'
+                            points='30.200000000000003,187.875 22.6,187.875 22.6,261.0 30.200000000000003,261.0' />
+
+<polyline style='stroke:red; stroke-width:2; fill:none; stroke-linejoin:round; stroke-linecap:round;'
+                            points='73.20000000000002,187.875 80.80000000000001,187.875 80.80000000000001,261.0 73.20000000000002,261.0' />
+</svg>

data/test/cli_test.rb

@@ -0,0 +1,262 @@
+# frozen_string_literal: true
+
+require "minitest/autorun"
+require "minitest/pride"
+require "tempfile"
+require "open3"
+
+class CLITest < Minitest::Test
+  BIN_PATH = File.expand_path("../bin/rsyntaxtree", __dir__)
+  RUBY = "ruby"
+
+  def setup
+    @tmpdir = Dir.mktmpdir
+  end
+
+  def teardown
+    FileUtils.remove_entry(@tmpdir) if @tmpdir && Dir.exist?(@tmpdir)
+  end
+
+  # Helper to run CLI and capture output (defined at bottom as private)
+
+  # ===================
+  # Feature 1: stdin support
+  # ===================
+
+  def test_stdin_input_generates_output
+    tree_data = "[S [NP test] [VP works]]"
+    outfile = File.join(@tmpdir, "syntree.svg")
+
+    result = run_cli("-f", "svg", "-o", @tmpdir, stdin_data: tree_data)
+
+    assert result[:status].success?, "CLI should succeed with stdin input: #{result[:stderr]}"
+    assert File.exist?(outfile), "Output file should be created"
+    assert File.size(outfile) > 0, "Output file should not be empty"
+  end
+
+  def test_stdin_input_with_pipe_simulation
+    tree_data = "[S [NP hello] [VP world]]"
+    outfile = File.join(@tmpdir, "syntree.png")
+
+    result = run_cli("-f", "png", "-o", @tmpdir, stdin_data: tree_data)
+
+    assert result[:status].success?, "CLI should succeed with piped input"
+    assert File.exist?(outfile), "PNG output file should be created"
+  end
+
+  def test_argument_takes_precedence_over_stdin
+    tree_data_stdin = "[S [NP stdin] [VP data]]"
+    tree_data_arg = "[S [NP arg] [VP data]]"
+    outfile = File.join(@tmpdir, "syntree.svg")
+
+    # When both arg and stdin provided, arg should win
+    result = run_cli("-f", "svg", "-o", @tmpdir, tree_data_arg, stdin_data: tree_data_stdin)
+
+    assert result[:status].success?, "CLI should succeed"
+    assert File.exist?(outfile), "Output file should be created"
+    content = File.read(outfile)
+    assert content.include?("arg"), "Output should contain data from argument, not stdin"
+  end
+
+  def test_file_input_still_works
+    tree_file = File.join(@tmpdir, "input.txt")
+    File.write(tree_file, "[S [NP file] [VP input]]")
+    outfile = File.join(@tmpdir, "syntree.svg")
+
+    result = run_cli("-f", "svg", "-o", @tmpdir, tree_file)
+
+    assert result[:status].success?, "CLI should succeed with file input"
+    assert File.exist?(outfile), "Output file should be created"
+    content = File.read(outfile)
+    assert content.include?("file"), "Output should contain data from file"
+  end
+
+  # ===================
+  # Feature 2: Config file support
+  # ===================
+
+  def test_config_file_invalid_format_shows_error
+    config_file = File.join(@tmpdir, ".rsyntaxtreerc")
+    File.write(config_file, <<~YAML)
+      format: invalid_format
+    YAML
+
+    tree_data = "[S [NP test] [VP config]]"
+    result = run_cli("-o", @tmpdir, tree_data, chdir: @tmpdir)
+
+    refute result[:status].success?, "CLI should fail with invalid format in config"
+    assert result[:stderr].include?("format") || result[:stdout].include?("format"),
+           "Error should mention 'format'"
+  end
+
+  def test_config_file_invalid_fontsize_shows_error
+    config_file = File.join(@tmpdir, ".rsyntaxtreerc")
+    File.write(config_file, <<~YAML)
+      fontsize: 999
+    YAML
+
+    tree_data = "[S [NP test] [VP config]]"
+    result = run_cli("-o", @tmpdir, tree_data, chdir: @tmpdir)
+
+    refute result[:status].success?, "CLI should fail with invalid fontsize in config"
+    assert result[:stderr].include?("fontsize") || result[:stdout].include?("fontsize"),
+           "Error should mention 'fontsize'"
+  end
+
+  def test_config_file_invalid_color_shows_error
+    config_file = File.join(@tmpdir, ".rsyntaxtreerc")
+    File.write(config_file, <<~YAML)
+      color: rainbow
+    YAML
+
+    tree_data = "[S [NP test] [VP config]]"
+    result = run_cli("-o", @tmpdir, tree_data, chdir: @tmpdir)
+
+    refute result[:status].success?, "CLI should fail with invalid color in config"
+    assert result[:stderr].include?("color") || result[:stdout].include?("color"),
+           "Error should mention 'color'"
+  end
+
+  def test_config_file_unknown_key_shows_warning
+    config_file = File.join(@tmpdir, ".rsyntaxtreerc")
+    File.write(config_file, <<~YAML)
+      format: svg
+      unknown_option: value
+    YAML
+
+    tree_data = "[S [NP test] [VP config]]"
+    outfile = File.join(@tmpdir, "syntree.svg")
+    result = run_cli("-o", @tmpdir, tree_data, chdir: @tmpdir)
+
+    # Should succeed but warn about unknown key
+    assert result[:status].success?, "CLI should succeed with unknown key"
+    assert File.exist?(outfile), "Output should be created"
+    assert result[:stderr].include?("unknown_option") || result[:stdout].include?("unknown_option"),
+           "Should warn about unknown option"
+  end
+
+  def test_config_file_shows_path_in_error
+    config_file = File.join(@tmpdir, ".rsyntaxtreerc")
+    File.write(config_file, <<~YAML)
+      format: invalid
+    YAML
+
+    tree_data = "[S [NP test] [VP config]]"
+    result = run_cli("-o", @tmpdir, tree_data, chdir: @tmpdir)
+
+    refute result[:status].success?, "CLI should fail"
+    # Error message should include the config file path
+    assert result[:stderr].include?(".rsyntaxtreerc") || result[:stdout].include?(".rsyntaxtreerc"),
+           "Error should mention config file path"
+  end
+
+  def test_config_file_in_current_dir
+    # Create config file in tmpdir
+    config_file = File.join(@tmpdir, ".rsyntaxtreerc")
+    File.write(config_file, <<~YAML)
+      format: svg
+      color: off
+      fontsize: 20
+    YAML
+
+    tree_data = "[S [NP config] [VP test]]"
+    outfile = File.join(@tmpdir, "syntree.svg")
+
+    # Run from tmpdir with config
+    result = run_cli("-o", @tmpdir, tree_data, chdir: @tmpdir)
+
+    assert result[:status].success?, "CLI should succeed with config file: #{result[:stderr]}"
+    assert File.exist?(outfile), "Should output SVG as specified in config"
+  end
+
+  def test_cli_args_override_config
+    config_file = File.join(@tmpdir, ".rsyntaxtreerc")
+    File.write(config_file, <<~YAML)
+      format: svg
+    YAML
+
+    tree_data = "[S [NP override] [VP test]]"
+    outfile_png = File.join(@tmpdir, "syntree.png")
+
+    # CLI arg -f png should override config's svg
+    result = run_cli("-f", "png", "-o", @tmpdir, tree_data, chdir: @tmpdir)
+
+    assert result[:status].success?, "CLI should succeed"
+    assert File.exist?(outfile_png), "Should output PNG (CLI override)"
+  end
+
+  def test_config_file_with_custom_filename
+    config_file = File.join(@tmpdir, ".rsyntaxtreerc")
+    File.write(config_file, <<~YAML)
+      format: svg
+      outfilename: custom_tree
+    YAML
+
+    tree_data = "[S [NP custom] [VP name]]"
+    outfile = File.join(@tmpdir, "custom_tree.svg")
+
+    result = run_cli("-o", @tmpdir, tree_data, chdir: @tmpdir)
+
+    assert result[:status].success?, "CLI should succeed"
+    assert File.exist?(outfile), "Should use custom filename from config"
+  end
+
+  # ===================
+  # Feature 3: Penn TreeBank format
+  # ===================
+
+  def test_penn_treebank_input
+    tree_data = "(S (NP hello) (VP world))"
+    outfile = File.join(@tmpdir, "syntree.svg")
+
+    result = run_cli("-f", "svg", "-o", @tmpdir, tree_data)
+
+    assert result[:status].success?, "CLI should succeed with Penn TreeBank input: #{result[:stderr]}"
+    assert File.exist?(outfile), "Output file should be created"
+    content = File.read(outfile)
+    assert content.include?("hello"), "Output should contain tree content"
+  end
+
+  def test_penn_treebank_nested
+    tree_data = "(S (NP (Det the) (N dog)) (VP (V runs)))"
+    outfile = File.join(@tmpdir, "syntree.svg")
+
+    result = run_cli("-f", "svg", "-o", @tmpdir, tree_data)
+
+    assert result[:status].success?, "CLI should succeed with nested Penn TreeBank"
+    assert File.exist?(outfile), "Output file should be created"
+  end
+
+  def test_penn_treebank_from_file
+    tree_file = File.join(@tmpdir, "tree.penn")
+    File.write(tree_file, "(S (NP test) (VP file))")
+    outfile = File.join(@tmpdir, "syntree.svg")
+
+    result = run_cli("-f", "svg", "-o", @tmpdir, tree_file)
+
+    assert result[:status].success?, "CLI should succeed with Penn file input"
+    assert File.exist?(outfile), "Output file should be created"
+  end
+
+  def test_penn_treebank_from_stdin
+    tree_data = "(S (NP stdin) (VP penn))"
+    outfile = File.join(@tmpdir, "syntree.svg")
+
+    result = run_cli("-f", "svg", "-o", @tmpdir, stdin_data: tree_data)
+
+    assert result[:status].success?, "CLI should succeed with Penn from stdin"
+    assert File.exist?(outfile), "Output file should be created"
+  end
+
+  private
+
+  # Updated helper to support chdir
+  def run_cli(*args, stdin_data: nil, chdir: nil)
+    cmd = [RUBY, BIN_PATH] + args
+    opts = {}
+    opts[:stdin_data] = stdin_data if stdin_data
+    opts[:chdir] = chdir if chdir
+    stdout, stderr, status = Open3.capture3(*cmd, **opts)
+    { stdout: stdout, stderr: stderr, status: status }
+  end
+end