cbeta 3.5.5 → 3.6.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2eb898277b45bb91b77ad88a0a064ec1780647cfbb8956fa6d407aa5250fef6d
-  data.tar.gz: 70d0c1f69410cff20427dc4b0cb5016c11a360c47b2fd0f24e0b38d84b9b6068
+  metadata.gz: d57589aae785253515b3d2e73f7e5ee0f2267eb5cfd78ffc07a31cda853e0dc4
+  data.tar.gz: 636713370e0c81c2b25121be8aa58f4d959eb24c0297f64d6a6300431848b805
 SHA512:
-  metadata.gz: 2e37c7495f13476adf58ad47eb5cb435402085af2dbca6b20821b63f23b00c907c2fca028035b1ba1e365adddcb04fd281df6f04daa13a8c2132a94bd33fc1d1
-  data.tar.gz: 7dbb43982e33c8a5d533d8fee814f418aebf0a2e8c40df9d59a68bd726f9de67ddb8c15aee521f37c14c8e792cc0f53ca2035598819076c630468b25d0fb4c9e
+  metadata.gz: de9137c36aad2f633b1b682d7b22a32d99d260b7e2d5a506bd404a2a13596098755884da9292dacf6692f9751b1744c9f8eccdd5aacc8670c66cadd0ab94a287
+  data.tar.gz: 8885756792aeb5e473fbf4ffa638c03ba330156a10e7f800464146c93b81defb9a7b8ed5103c6b9cee010c6714b3e361eb1ae0e685dd9f94fc29eb9f2dab1fdc

data/lib/cbeta/cbeta_share.rb

@@ -1,4 +1,10 @@
 module CbetaShare
+  def each_canon(xml_root)
+    Dir.entries(xml_root).sort.each do |c|
+      next unless c.match(/^#{CBETA::CANON}$/)
+      yield(c)
+    end
+  end
 
   def to_html(e)
     e.to_xml(
@@ -8,4 +14,4 @@ module CbetaShare
     )
   end
 
-end
+end

data/lib/cbeta/p5a_checker.rb

@@ -0,0 +1,162 @@
+require_relative 'cbeta_share'
+
+# 檢查 CBETA XML P5a
+class CBETA::P5aChecker
+  # @param xml_root [String] 來源 CBETA XML P5a 路徑
+  # @param figures [String] 插圖 路徑 (可由 https://github.com/cbeta-git/CBR2X-figures 取得)
+  # @param log [String] Log file path
+  def initialize(xml_root: nil, figures: nil, log: nil)
+    @gaijis = CBETA::Gaiji.new
+    @xml_root = xml_root
+    @figures = figures
+    @log = log
+  end
+  
+  def check
+    @errors = ''
+    @g_errors = {}
+    puts "xml: #{@xml_root}"
+    each_canon(@xml_root) do |c|
+      @canon = c
+      path = File.join(@xml_root, @canon)
+      handle_canon(path)
+    end
+
+    @g_errors.keys.sort.each do |k|
+      s = @g_errors[k].to_a.join(',')
+      @errors << "#{k} 無缺字資料，出現於：#{s}\n"
+    end
+    
+    if @errors.empty?
+      puts "檢查完成，未發現錯誤。"
+    elsif @log.nil?
+      puts "\n發現錯誤："
+      puts @errors
+    else
+      File.write(@log, @errors)
+      puts "\n發現錯誤，請查看 #{@log}"
+    end
+  end
+  
+  private
+
+  include CbetaShare
+
+  def chk_text(node)
+    return if node.text.strip.empty?
+    if node.parent.name == 'div'
+      error "lb: #{@lb}, text: #{node.text.inspect}", type: "[E02] 文字直接出現在 div 下"
+    end
+  end
+
+  def e_g(e)
+    gid = e['ref'][1..-1]
+    unless @gaijis.key? gid
+      @g_errors[gid] = Set.new unless @g_errors.key? gid
+      @g_errors[gid] << @basename
+    end
+  end
+  
+  def e_graphic(e)
+    url = File.basename(e['url'])
+    fn = File.join(@figures, @canon, url)
+    unless File.exist? fn
+      error "圖檔 #{url} 不存在"
+    end
+  end
+  
+  def e_lb(e)
+    return if e['type']=='old'
+    unless e['n'].match(/^[a-z\d]\d{3}[a-z]\d+$/)
+      error "lb format error: #{e['n']}"
+    end
+
+    @lb = e['n']
+    ed_lb = "#{e['ed']}#{@lb}"
+    if @lbs.include? ed_lb
+      unless e['ed'].start_with?('R')
+        error "lb: #{@lb}, ed: #{e['ed']}", type: "[E01] 行號重複"
+      end
+    else
+      @lbs << ed_lb
+    end
+  end
+  
+  def e_lem(e)
+    unless e.key?('wit')
+      error "lem 缺少 wit 屬性"
+    end
+  end
+
+  def e_rdg(e)
+    return if e['type'] == 'cbetaRemark'
+    unless e.key?('wit')
+      error "rdg 缺少 wit 屬性, lb: #{@lb}"
+    end
+  end
+
+  def error(msg, type: nil)
+    s = ''
+    s << "#{type}: " unless type.nil?
+    s << "#{@basename}, #{msg}"
+    puts s
+    @errors << s + "\n"
+  end
+  
+  def handle_canon(folder)
+    Dir.entries(folder).sort.each do |f|
+      next if f.start_with? '.'
+      @vol = f
+      $stderr.puts @vol + ' '
+      path = File.join(folder, @vol)
+      handle_vol(path)
+    end
+  end
+  
+  def handle_file(fn)
+    @basename = File.basename(fn)
+    
+    s = File.read(fn)
+    if s.include? "\u200B"
+      @errors << "#{@basename} 含有 U+200B Zero Width Space 字元\n"
+    end
+    
+    doc = Nokogiri::XML(s)
+    if doc.errors.empty?
+      doc.remove_namespaces!
+      @lbs = Set.new
+      traverse(doc.root)
+    else
+      @errors << "錯誤: #{@basename} not well-formed\n"
+    end
+  end
+
+  def handle_node(e)
+    case e.name
+    when 'g'       then e_g(e)
+    when 'graphic' then e_graphic(e)
+    when 'lb'      then e_lb(e)
+    when 'lem'     then e_lem(e)
+    when 'rdg'     then e_rdg(e)
+    else traverse(e)
+    end
+  end
+  
+  def handle_vol(folder)
+    Dir.entries(folder).sort.each do |f|
+      next if f.start_with? '.'
+      path = File.join(folder, f)
+      handle_file(path)
+    end
+  end
+
+  def traverse(e)
+    e.children.each { |c| 
+      if c.text?
+        chk_text(c)
+      elsif e.element?
+        handle_node(c)
+      end
+    }
+  end
+end

data/lib/cbeta.rb

@@ -282,6 +282,7 @@ require 'cbeta/canon'
 require 'cbeta/char_count'
 require 'cbeta/char_freq'
 require 'cbeta/html_to_pdf'
+require 'cbeta/p5a_checker'
 require 'cbeta/p5a_to_html'
 require 'cbeta/p5a_to_html_for_every_edition'
 require 'cbeta/p5a_to_html_for_pdf'

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: cbeta
 version: !ruby/object:Gem::Version
-  version: 3.5.5
+  version: 3.6.0
 platform: ruby
 authors:
 - Ray Chou
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-03-31 00:00:00.000000000 Z
+date: 2025-04-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: unihan2
@@ -16,20 +16,20 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.1'
+        version: '1.2'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.1.0
+        version: 1.2.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.1'
+        version: '1.2'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.1.0
+        version: 1.2.0
 - !ruby/object:Gem::Dependency
   name: nokogiri
   requirement: !ruby/object:Gem::Requirement
@@ -85,6 +85,7 @@ files:
 - lib/cbeta/gaiji.rb
 - lib/cbeta/html_to_pdf.rb
 - lib/cbeta/html_to_text.rb
+- lib/cbeta/p5a_checker.rb
 - lib/cbeta/p5a_to_html.rb
 - lib/cbeta/p5a_to_html_for_every_edition.rb
 - lib/cbeta/p5a_to_html_for_pdf.rb