jekyll-uj-powertools 1.3.0 → 1.3.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2db2b27f78f8588283c589c8d6c06e342ba6a11a1b1b55669ea81b722b5fe64b
-  data.tar.gz: 47b62725a184ea99d3f25daec3620e57539dbd8fb42159bdd81bb1f68176e5f9
+  metadata.gz: f91b80879708adc5fb116e4ce3722237453ddd4ba20c9c1ebf5f798446f45c64
+  data.tar.gz: e35de7eebf30b3f535a64abb5f09f06c7c0ea2b75e0c304ccc05571927dca464
 SHA512:
-  metadata.gz: 81e5f82be027417b008c9a6d3cbc2bb86bc5f992a60447fb42dc94011da6abb4cb5374756bca80141c36cad263e92e16c2a7ee18a94c4c5122173e9c1464a5ae
-  data.tar.gz: 6a3ea9d448ba3a543cf70623e6eb67ab72dc25eecbd761d9959e66cf6a9db92342271307713a95ea6bb268b761e2f23e0565d107c3c4fa8f6528543b3118f036
+  metadata.gz: 2ed5f204a4864131bbdbd54fa8838a058e385fa60826274d2208cfee36e3c23b57978df8fcdcd3b972a9629a3825c55618a5b437e012be88768ccb1d4fd09ab6
+  data.tar.gz: f4a1853571d6c486aa054e09ed725eb2af9035ada859fac30446a3356943c9b63a665a0d673aba75695471df0f30a1b49c308a6f62060feba91c27207e61c9ee

data/README.md

@@ -92,11 +92,17 @@ bundle exec rspec
 
 ## 💎 Build + Publish the Gem
 ```shell
-# Build the gem
-gem build jekyll-uj-powertools.gemspec
+# Ensure dist folder exists
+mkdir -p dist
 
-# Publish the gem where X.X.X is the version number
-gem push jekyll-uj-powertools-X.X.X.gem
+# Build the gem and push it to RubyGems
+gem build jekyll-uj-powertools.gemspec -o dist/jekyll-uj-powertools-latest.gem
+
+# Publish the latest gem
+gem push dist/jekyll-uj-powertools-latest.gem
+
+# Clear the files in the dist folder
+rm -rf dist/*
 ```
 
 ## 🗨️ Contributing

data/lib/{generators → _old/generators}/translate-pages.rb

@@ -16,11 +16,12 @@ module Jekyll
     RECHECK_DAYS = 30
 
     def generate(site)
+      # Get target languages from site config
       target_langs = site.config.dig('translation', 'languages') || []
 
       # Log
-      puts "🔁 Starting translation process for supported languages: #{target_langs.join(', ')}"
-      puts "📂 Cache directory: #{CACHE_DIR}"
+      puts "🔁 Starting translation process for supported languages (#{target_langs.length}): #{target_langs.join(', ')}"
+      # puts "📂 Cache directory: #{CACHE_DIR}"
       # puts "🔍 All environment variables:"
       # ENV.each { |k, v| puts "   #{k}=#{v}" }
       puts "🔍 UJ_ environment variables:"
@@ -32,6 +33,12 @@ module Jekyll
         return
       end
 
+      # Quit if UJ_BUILD_MODE is false
+      if ENV['UJ_BUILD_MODE'] == 'false' && ENV['UJ_TRANSLATION_FORCE'] != 'true'
+        puts "🚫 UJ_BUILD_MODE is set to 'false' (set UJ_TRANSLATION_FORCE=true). Exiting translation process."
+        return
+      end
+
       # Ensure OpenAI API key is set
       unless ENV['OPENAI_API_KEY'] && !ENV['OPENAI_API_KEY'].strip.empty?
         puts "❌ OPENAI_API_KEY not found in environment. Exiting translation process."
@@ -72,8 +79,9 @@ module Jekyll
 
           # @TODO: Remove this
           # Unless its pages/legal/terms.md, QUIT
-          if page_path != 'pages/legal/terms.md'
-            skipped_files << "#{page_path} (only 'pages/legal/terms.md' is processed)"
+          uj_translation_only = ENV['UJ_TRANSLATION_ONLY']
+          if uj_translation_only && page_path != uj_translation_only
+            skipped_files << "#{page_path} (UJ_TRANSLATION_ONLY is set)"
             next
           end
 
@@ -132,6 +140,10 @@ module Jekyll
         puts "📭 No cache found, generating translation..."
       end
 
+      # Log before/after content
+      puts "\n--- BEFORE CONTENT (#{lang}) ---\n#{content[0..500]}..."
+
+      # Translate the content using OpenAI API
       begin
         result = translate_with_api(content, lang)
       rescue => e
@@ -139,6 +151,10 @@ module Jekyll
         return nil
       end
 
+      # Log the first 500 characters of the result
+      puts "\n--- AFTER TRANSLATION (#{lang}) ---\n#{result[0..500]}..."
+
+      # Save the translation and metadata
       FileUtils.mkdir_p(File.dirname(path))
       File.write(path, result)
       File.write(meta_path, {
@@ -148,10 +164,6 @@ module Jekyll
 
       puts "📝 Cached translation and metadata written to: #{path}"
 
-      # Log before/after content
-      puts "\n--- BEFORE CONTENT (#{lang}) ---\n#{content[0..500]}..."
-      puts "\n--- AFTER TRANSLATION (#{lang}) ---\n#{result[0..500]}..."
-
       result
     end
 
@@ -183,6 +195,8 @@ module Jekyll
       end
 
       json = JSON.parse(res.body)
+      # Log json
+      puts "🔍 API response: #{json.inspect}"
       result = json.dig('choices', 0, 'message', 'content')
 
       if result.nil? || result.strip.empty?

data/lib/_old/hooks/translate-pages.rb

@@ -0,0 +1,253 @@
+# Libraries
+require 'json'
+require 'net/http'
+require 'fileutils'
+require 'nokogiri'
+require 'digest'
+
+# Hook
+Jekyll::Hooks.register :site, :post_write do |site|
+  # Variables
+  # Translation path
+  CACHE_DIR = '.temp/translations'
+  # Re-translate pages older than this many days
+  RECHECK_DAYS = 30
+
+  # Get target languages from site config
+  target_langs = site.config.dig('translation', 'languages') || []
+
+  # Log
+  puts "🔁 Starting translation process for supported languages (#{target_langs.length}): #{target_langs.join(', ')}"
+  puts "🔍 UJ_ environment variables:"
+  ENV.select { |k, _| k.start_with?('UJ_') }.each { |k, v| puts "   #{k}=#{v}" }
+
+  # Skip if site config translation is disabled
+  unless site.config.dig('translation', 'enabled')
+    puts "🚫 Translation is disabled in _config.yml (translation.enabled: false)"
+    next
+  end
+
+  # Quit if UJ_BUILD_MODE is false
+  if ENV['UJ_BUILD_MODE'] == 'false' && ENV['UJ_TRANSLATION_FORCE'] != 'true'
+    puts "🚫 UJ_BUILD_MODE is set to 'false' (set UJ_TRANSLATION_FORCE=true). Exiting translation process."
+    next
+  end
+
+  # Ensure OpenAI API key is set
+  unless ENV['OPENAI_API_KEY'] && !ENV['OPENAI_API_KEY'].strip.empty?
+    puts "❌ OPENAI_API_KEY not found in environment. Exiting translation process."
+    next
+  end
+
+  # Quit if no languages are configured
+  if target_langs.empty?
+    puts "🚫 No target languages configured in _config.yml (translation.languages). Exiting translation process."
+    next
+  end
+
+  # Keep track of skipped files
+  skipped_files = []
+
+  # Loop through all pages in the site
+  site.pages.clone.each do |page|
+    # Quit if its not an HTML page
+    next unless page.output_ext == '.html'
+
+    # Get original content
+    original_content = page.output
+
+    # Extract body content
+    doc = Nokogiri::HTML(original_content)
+    original_content_body = doc.at('body')&.inner_html.to_s
+
+    # Compute original hash
+    original_hash = Digest::SHA256.hexdigest(original_content_body)
+
+    # Get page path and URL
+    page_path = page.path.sub(/^_?site\//, '')
+    page_url = page.url
+
+    # Skip if page.translation.enabled is false
+    if page.data['translation'] && page.data['translation']['enabled'] == false
+      skipped_files << "#{page_path} (translation disabled)"
+      next
+    end
+
+    # Skip if page.redirect.url is set
+    if page.data['redirect'] && page.data['redirect']['url']
+      skipped_files << "#{page_path} (redirect set)"
+      next
+    end
+
+    # Loop through target languages
+    target_langs.each do |lang|
+      page_new_url = "/#{lang}#{page.url}"
+      page_new_path = File.join(CACHE_DIR, lang, page_new_url)
+      page_new_meta_path = "#{page_new_path}.meta.json"
+
+      # See if we only want to test a specific page
+      uj_translation_only = ENV['UJ_TRANSLATION_ONLY']
+      if uj_translation_only && page_path != uj_translation_only
+        skipped_files << "#{page_path} (UJ_TRANSLATION_ONLY is set)"
+        next
+      end
+
+      # Log
+      puts "🌐 Processing page '#{page_url}' for language '#{lang}'"
+
+      # LOG new_page.data
+      # Log permalink
+      puts "🔗 New permalink: #{page_new_url}"
+
+      # Either read cached translation or generate a new one
+      translated = read_or_translate(original_content_body, original_hash, lang, page_new_path, page_new_meta_path)
+
+      # Fallback if translation failed
+      if translated.nil?
+        puts "⚠️ Translation failed for #{page_url}, using original content and marking for retry"
+
+        # Force a retry next time by setting bad hash + old timestamp
+        FileUtils.mkdir_p(File.dirname(page_new_meta_path))
+        File.write(page_new_meta_path, {
+          timestamp: 0,
+          hash: '__fail__'
+        }.to_json)
+
+        translated = original_content_body
+      end
+
+      # Rewrite internal links
+      translated_html = rewrite_links(translated, lang)
+
+      # Inject translated content into original HTML structure
+      translated_doc = Nokogiri::HTML(original_content)
+      translated_doc.at('body').inner_html = translated_html
+      final_html = translated_doc.to_html
+
+      # Determine output path
+      output_dir = site.config['destination']
+      translated_output_path = File.join(output_dir, lang, page.url)
+      translated_output_path = File.join(translated_output_path, 'index.html') if translated_output_path.end_with?('/')
+
+      # Write translated page to disk
+      FileUtils.mkdir_p(File.dirname(translated_output_path))
+      File.write(translated_output_path, final_html)
+      puts "✅ Wrote translated file: #{translated_output_path}"
+    end
+  end
+
+  # Log skipped files at the end
+  if skipped_files.any?
+    puts "\n🚫 Skipped files:"
+    skipped_files.each { |f| puts " - #{f}" }
+  end
+
+  # Log
+  puts "🎉 Translation process complete."
+end
+
+def read_or_translate(content, hash, lang, path, page_new_meta_path)
+  if File.exist?(path) && File.exist?(page_new_meta_path)
+    meta = JSON.parse(File.read(page_new_meta_path)) rescue {}
+    last_hash = meta['hash']
+    last_time = Time.at(meta['timestamp'].to_i) rescue Time.at(0)
+
+    age_days = ((Time.now - last_time) / (60 * 60 * 24)).round
+    puts "📅 Cache age: #{age_days}/#{RECHECK_DAYS} days"
+
+    # Determine whether we should re-check based on age
+    recheck_due_to_age = RECHECK_DAYS && RECHECK_DAYS > 0 && age_days >= RECHECK_DAYS
+
+    # Re-translate if hash changed or translation is too old (only if RECHECK_DAYS is truthy)
+    if last_hash == hash && !recheck_due_to_age
+      puts "📦 Using cached translation at: #{path}"
+      return File.read(path)
+    else
+      puts "🔁 Cache stale or hash changed, regenerating translation..."
+    end
+  else
+    puts "📭 No cache found, generating translation..."
+  end
+
+  # Log before/after content
+  puts "\n--- BEFORE CONTENT (#{lang}) ---\n#{content[0..500]}..."
+
+  # Translate the content using OpenAI API
+  begin
+    result = translate_with_api(content, lang)
+  rescue => e
+    puts "❌ Skipping translation for '#{lang}' due to error: #{e.message}"
+    return nil
+  end
+
+  # Log the first 500 characters of the result
+  puts "\n--- AFTER TRANSLATION (#{lang}) ---\n#{result[0..500]}..."
+
+  # Save the translation and metadata
+  FileUtils.mkdir_p(File.dirname(path))
+  File.write(path, result)
+  File.write(page_new_meta_path, {
+    timestamp: Time.now.to_i,
+    hash: hash
+  }.to_json)
+
+  puts "📝 Cached translation and metadata written to: #{path}"
+
+  result
+end
+
+def translate_with_api(content, lang)
+  system_prompt = "You are a professional translator. Translate the provided HTML content, preserving all original formatting, HTML structure, metadata, and links. Do not explain anything — just return the translated HTML. Translate to #{lang}."
+  user_message = content
+
+  uri = URI('https://api.openai.com/v1/chat/completions')
+
+  http = Net::HTTP.new(uri.host, uri.port)
+  http.use_ssl = true
+  http.read_timeout = 30  # seconds
+  http.open_timeout = 10  # seconds
+
+  request = Net::HTTP::Post.new(uri.path, {
+    'Authorization' => "Bearer #{ENV['OPENAI_API_KEY']}",
+    'Content-Type' => 'application/json'
+  })
+
+  request.body = {
+    model: 'gpt-4o',
+    messages: [
+      { role: 'system', content: system_prompt },
+      { role: 'user', content: user_message }
+    ],
+    temperature: 0.3,
+    max_tokens: 4096
+  }.to_json
+
+  response = http.request(request)
+
+  if response.code.to_i != 200
+    raise "HTTP #{response.code}: #{response.body}"
+  end
+
+  json = JSON.parse(response.body)
+  puts "🔍 API response: #{json.inspect}"
+  result = json.dig('choices', 0, 'message', 'content')
+
+  if result.nil? || result.strip.empty?
+    raise "Translation returned empty or invalid content"
+  end
+
+  puts "🔤 Translation complete."
+  result
+end
+
+def rewrite_links(html, lang)
+  doc = Nokogiri::HTML::DocumentFragment.parse(html)
+  doc.css('a[href^="/"]').each do |a|
+    href = a['href']
+    next if href.start_with?("/#{lang}") || href.include?('.') || href.start_with?('//')
+    new_href = "/#{lang}#{href}"
+    puts "🔗 Rewriting link: #{href} -> #{new_href}"
+    a['href'] = new_href
+  end
+  doc.to_html
+end

data/lib/generators/{inject-data.rb → inject-properties.rb}

@@ -1,3 +1,6 @@
+# Libraries
+# ...
+
 module Jekyll
   class InjectData < Generator
     safe true
@@ -26,7 +29,13 @@ module Jekyll
       # Inject a random number into the item's data
       item.data['random_id'] = rand(100) # Random number between 0 and 99
 
-      return unless item.data['layout'] # Skip items without layouts
+      # Inject the file extension into the item's data
+      if item.respond_to?(:path)
+        item.data['extension'] = File.extname(item.path)
+      end
+
+      # Skip items without layouts
+      return unless item.data['layout']
 
       # Find the layout file by its name
       layout_name = item.data['layout']

data/lib/hooks/inject-properties.rb

@@ -0,0 +1,8 @@
+# Libraries
+# ...
+
+# Hook
+Jekyll::Hooks.register :site, :pre_render do |site|
+  site.config['uj'] ||= {}
+  site.config['uj']['cache_breaker'] = Jekyll::UJPowertools.cache_timestamp
+end

data/lib/jekyll-uj-powertools/version.rb

@@ -1,5 +1,5 @@
 module Jekyll
   module UJPowertools
-    VERSION = "1.3.0"
+    VERSION = "1.3.2"
   end
 end

data/lib/jekyll-uj-powertools.rb

@@ -1,5 +1,7 @@
+# Libraries
 require "jekyll"
 
+# Module
 module Jekyll
   module UJPowertools
     # Initialize a timestamp that will remain consistent across calls
@@ -60,18 +62,20 @@ module Jekyll
     def self.cache_timestamp
       @cache_timestamp
     end
+
+    # Check if a string ends with a specific suffix
+    # def uj_ends_with(input, suffix)
+    #   input.end_with?(suffix)
+    # end
   end
 
   # Load Generators
-  require_relative "generators/inject-data"
-  require_relative "generators/translate-pages"
+  require_relative "generators/inject-properties"
+
+  # Load Hooks
+  require_relative "hooks/inject-properties"
+  # require_relative "hooks/translate-pages"
 end
 
 # Register the filter
 Liquid::Template.register_filter(Jekyll::UJPowertools)
-
-# Register hook
-Jekyll::Hooks.register :site, :pre_render do |site|
-  site.config['uj'] ||= {}
-  site.config['uj']['cache_breaker'] = Jekyll::UJPowertools.cache_timestamp
-end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: jekyll-uj-powertools
 version: !ruby/object:Gem::Version
-  version: 1.3.0
+  version: 1.3.2
 platform: ruby
 authors:
 - ITW Creative Works
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-06-16 00:00:00.000000000 Z
+date: 2025-06-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: jekyll
@@ -99,8 +99,10 @@ files:
 - README.md
 - Rakefile
 - jekyll-uj-powertools.gemspec
-- lib/generators/inject-data.rb
-- lib/generators/translate-pages.rb
+- lib/_old/generators/translate-pages.rb
+- lib/_old/hooks/translate-pages.rb
+- lib/generators/inject-properties.rb
+- lib/hooks/inject-properties.rb
 - lib/jekyll-uj-powertools.rb
 - lib/jekyll-uj-powertools/version.rb
 - spec/jekyll-uj-powertools_spec.rb