jekyll-chatgpt-translate 0.0.35 → 0.0.37

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 841ab37ebdb5e4049f301883bd5d3c411dbf7719824c52f8fa3f3604aebb0109
-  data.tar.gz: e553ef7cdf4564adba5128bf9bb7eea90c59f7724fa72986427a5e31e7bf2bb2
+  metadata.gz: c0f1eb7be508b40853f8fc04d94cd0c6f30ddfe2be81b6d3fc7f6a45101d10b1
+  data.tar.gz: cb4d6e55f8d93e00307135f52cfba873a7ece62a01cb785f0d0e4ada293aaa20
 SHA512:
-  metadata.gz: a9a6c85190c260ba4770691f4810cbe775f6ae3b2aafba67f5f0081fb540e2a86dc817ae65fccb857643e39d446e4c62b2a98dabe8ce43a46df8a6786ac59b1b
-  data.tar.gz: dba3c7a5bd214a8f970fa623d7dc46779928ddf334fa10f3ca42cfafb61f9a7fae580cd86ed0f11d55f1a08082e0f04eead95ab098481231940e94216cdacaa0
+  metadata.gz: 053d40f293680a7b4ec3f78f5b09b1ddef17183b1dbfce93ab10fffca3cbb1b1384644e4da3377c5c1abc0ff96fdb1b55d8f58d7e30788aa7ddeba44dc6e0d65
+  data.tar.gz: 0cc7dd560f374a400b8c8ff2b7782daca793a2db48eac1fc698806506f0341713751241a2feb739d485145efbe57ba6c5d2d43d188208fa0e8142993447a6928

data/Gemfile

@@ -25,11 +25,11 @@
 source 'https://rubygems.org'
 gemspec
 
-gem 'cucumber', '9.0.1', require: false
+gem 'cucumber', '9.0.2', require: false
 gem 'kramdown-parser-gfm', '1.1.0', require: false
 gem 'minitest', '5.19.0', require: false
 gem 'rake', '13.0.6', require: false
-gem 'rubocop', '1.56.2', require: false
-gem 'rubocop-rspec', '2.23.2', require: false
+gem 'rubocop', '1.56.4', require: false
+gem 'rubocop-rspec', '2.24.0', require: false
 gem 'simplecov', '0.22.0', require: false
 gem 'webmock', '3.19.1', require: false

data/jekyll-chatgpt-translate.gemspec

@@ -28,7 +28,7 @@ Gem::Specification.new do |s|
   s.required_rubygems_version = Gem::Requirement.new('>= 0') if s.respond_to? :required_rubygems_version=
   s.required_ruby_version = '>= 2.6'
   s.name = 'jekyll-chatgpt-translate'
-  s.version = '0.0.35'
+  s.version = '0.0.37'
   s.license = 'MIT'
   s.summary = 'Translate Jekyll Pages Through ChatGPT'
   s.description = [
@@ -42,6 +42,7 @@ Gem::Specification.new do |s|
   s.executables = s.files.grep(%r{^bin/}) { |f| File.basename(f) }
   s.rdoc_options = ['--charset=UTF-8']
   s.extra_rdoc_files = %w[README.md LICENSE.txt]
+  s.add_runtime_dependency 'humanize', '>= 2'
   s.add_runtime_dependency 'iri', '>= 0'
   s.add_runtime_dependency 'iso-639', '>= 0'
   s.add_runtime_dependency 'jekyll', '>= 3'

data/lib/jekyll-chatgpt-translate/chatgpt.rb

@@ -51,36 +51,62 @@ class GptTranslate::ChatGPT
     @target = target
   end
 
-  def translate(markdown, min: 32)
-    GptTranslate::Pars.new(markdown).to_a.map do |pa|
+  def translate(markdown, min: 32, window_length: 2000)
+    pars = GptTranslate::Pars.new(markdown).to_a
+    ready = []
+    later = []
+    pars.each_with_index do |pa, i|
       par = pa.dup
       par.strip!
       if @source == @target
         Jekyll.logger.debug("No need to translate from #{@source.inspect} to #{@target.inspect}: #{par.inspect}")
-        par
+        ready[i] = par
       elsif par.length < min
         Jekyll.logger.debug("Not translating this, b/c too short: #{par.inspect}")
-        par
+        ready[i] = par
       elsif par.start_with?('```')
         Jekyll.logger.debug("Not translating this code block: #{par.inspect}")
-        par
-      elsif par =~ /^[^\p{Alnum}\*'"\[]/
-        Jekyll.logger.debug("Not translating this, b/c it's not a plain text: #{par.inspect}")
-        par
+        ready[i] = par
       elsif @key.empty?
-        par
+        ready[i] = par
+      elsif par.start_with?('> ')
+        ready[i] = "> #{translate_par(par[2..])}"
       elsif par.start_with?('* ')
-        "* #{translate_par(par[2..])}"
+        ready[i] = "* #{translate_par(par[2..])}"
       elsif par =~ /^[0-9]+\. /
-        "1. #{translate_par(par.split('.', 2)[1])}"
+        ready[i] = "1. #{translate_par(par.split('.', 2)[1])}"
+      elsif par =~ /^[^\p{Alnum}\*'"\[]/
+        Jekyll.logger.debug("Not translating this, b/c it's not a plain text: #{par.inspect}")
+        ready[i] = par
       else
-        translate_par(par)
+        later[i] = par
+      end
+    end
+    out = []
+    i = 0
+    while i < pars.length
+      unless ready[i].nil?
+        out << ready[i]
+        i += 1
+        next
+      end
+      accum = []
+      until later[i].nil?
+        accum << later[i]
+        break if accum.join.split.count > window_length
+        i += 1
       end
-    end.join("\n\n")
+      out << translate_pars(accum)
+    end
+    out.join("\n\n")
   end
 
   private
 
+  def translate_pars(accum)
+    translate_par(accum.join("\n\n"))
+  end
+
   def translate_par(par)
     client = OpenAI::Client.new(access_token: @key)
     if @@models_printed

data/lib/jekyll-chatgpt-translate/generator.rb

@@ -63,7 +63,7 @@ class GptTranslate::Generator < Jekyll::Generator
     site.posts.docs.shuffle.each_with_index do |doc, pos|
       plain = GptTranslate::Plain.new(doc.content).to_s
       config['targets'].each do |target|
-        start = Time.now
+        pstart = Time.now
         link = GptTranslate::Permalink.new(doc, target['permalink']).to_path
         lang = target['language']
         raise 'Language must be defined for each target' if target.nil?
@@ -129,7 +129,7 @@ class GptTranslate::Generator < Jekyll::Generator
           added = true
           translated += 1
           Jekyll.logger.info("Translated via ChatGPT \
-in #{(Time.now - start).round(2)}s: #{path} (#{File.size(path)} bytes)")
+in #{(Time.now - pstart).round(2)}s: #{path} (#{File.size(path)} bytes)")
         end
         next unless added
         doc.data['chatgpt-translate'] ||= {}

data/lib/jekyll-chatgpt-translate/plain.rb

@@ -55,8 +55,6 @@ class GptTranslate::Plain
   # Motivated by https://github.com/vmg/redcarpet/blob/master/lib/redcarpet/render_strip.rb
   class Strip < Redcarpet::Render::Base
     %i[
-      block_quote
-      block_html
       autolink
       underline
       triple_emphasis
@@ -78,6 +76,10 @@ class GptTranslate::Plain
       code
     end
 
+    def block_quote(txt)
+      "> #{txt}"
+    end
+
     def emphasis(txt)
       "*#{txt}*"
     end
@@ -100,6 +102,10 @@ class GptTranslate::Plain
       "![#{alt}](#{link} \"#{title}\")"
     end
 
+    def block_html(html)
+      "#{html}\n"
+    end
+
     def raw_html(html)
       html
     end

data/lib/jekyll-chatgpt-translate/prompt.rb

@@ -23,6 +23,7 @@
 # SOFTWARE.
 
 require 'iso-639'
+require 'humanize'
 
 # The module we are in.
 module GptTranslate; end
@@ -47,8 +48,11 @@ class GptTranslate::Prompt
     raise "Unknown source language ISO-639 code: #{@source.inspect}" if from.nil?
     to = ISO_639.find_by_code(@target)
     raise "Unknown source language ISO-639 code: #{@target.inspect}" if to.nil?
+    md = @par
+    parts = md.split("\n\n")
+    label = parts.size > 1 ? "#{parts.size.humanize(locale: :en)} Markdown paragraphs" : 'Markdown paragraph'
     head = [
-      'Please, translate the following Markdown paragraph from ',
+      "Please, translate the following #{label} from ",
       from[3],
       ' to ',
       to[3],

data/lib/jekyll-chatgpt-translate/version.rb

@@ -23,5 +23,5 @@
 # SOFTWARE.
 
 module GptTranslate
-  VERSION = '0.0.35'
+  VERSION = '0.0.37'
 end

data/test/test_chatgpt.rb

@@ -64,9 +64,9 @@ class GptTranslate::ChatGPTTest < Minitest::Test
     chat = GptTranslate::ChatGPT.new('', 'foo', 'xx', 'xx')
     chat.translate(
       "
-      How are you, my friend?
+      How are you, my friend? This text must be translated through ChatGPT.
 
-      Read this Java code:
+      Read this Java code (this paragraph must also be translated through ChatGPT):
 
       ```
       System.out.println(\"Hello, dude!\");
@@ -76,7 +76,8 @@ class GptTranslate::ChatGPTTest < Minitest::Test
 
       This is it.
       ",
-      min: 40
+      min: 40,
+      window_length: 10
     )
   end
 

data/test/test_plain.rb

@@ -85,6 +85,13 @@ class GptTranslate::PlainTest < Minitest::Test
     )
   end
 
+  def test_quote
+    assert_equal(
+      "He said this:\n\n> Life is great!",
+      GptTranslate::Plain.new("He said this:\n\n\n> Life is great!\n\n").to_s
+    )
+  end
+
   def test_code
     assert_equal(
       'Hello, `Java`!',
@@ -136,6 +143,11 @@ class GptTranslate::PlainTest < Minitest::Test
     assert_equal('<img src="a"/>', GptTranslate::Plain.new('<img src="a"/>').to_s)
   end
 
+  def test_html_hr
+    md = "First\n\n<hr/>\n\nsecond!"
+    assert_equal(md, GptTranslate::Plain.new(md).to_s)
+  end
+
   def test_liquid_tags
     assert_equal(
       'Hello, !',

data/test/test_prompt.rb

@@ -52,4 +52,10 @@ class GptTranslate::PromptTest < Minitest::Test
       GptTranslate::Prompt.new('Hello, Jeff!', 'en', 'zh').to_s
     )
   end
+
+  def test_multiple_paragraphs
+    assert(
+      GptTranslate::Prompt.new("Hello,\n\nJeff!", 'en', 'zh').to_s.include?("\"Hello,\n\nJeff!\"")
+    )
+  end
 end

metadata

@@ -1,15 +1,29 @@
 --- !ruby/object:Gem::Specification
 name: jekyll-chatgpt-translate
 version: !ruby/object:Gem::Version
-  version: 0.0.35
+  version: 0.0.37
 platform: ruby
 authors:
 - Yegor Bugayenko
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2023-09-06 00:00:00.000000000 Z
+date: 2023-10-02 00:00:00.000000000 Z
 dependencies:
+- !ruby/object:Gem::Dependency
+  name: humanize
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '2'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '2'
 - !ruby/object:Gem::Dependency
   name: iri
   requirement: !ruby/object:Gem::Requirement