ai-featureify the rest of the translators

Author: nattsw

config/locales/server.en.yml

@@ -374,7 +374,16 @@ en:
           description: "Default persona powering the Helper's image caption feature"
         locale_detection:
           name: "Locale detection"
-          description: "Powers the translation feature by detecting the locale of a given text"
+          description: "Powers the translation feature by detecting the locale of a given text (posts, titles, etc.)"
+        post_raw_translator:
+          name: "Post translator"
+          description: "Powers the translation feature by translating posts containing Discourse Markdown"
+        topic_title_translator:
+          name: "Topic title translator"
+          description: "Powers the translation feature by translating topic titles"
+        short_text_translator:
+          name: "Short text translator"
+          description: "Powers the translation feature by as a generic text translator, used for short texts like category names or tags"
 
       topic_not_found: "Summary unavailable, topic not found!"
       summarizing: "Summarizing topic"

config/settings.yml

@@ -460,6 +460,26 @@ discourse_ai:
     enum: "DiscourseAi::Configuration::LlmEnumerator"
     validator: "DiscourseAi::Configuration::LlmValidator"
     area: "ai-features/translation"
+  ai_translation_locale_detection_persona:
+    default: "-27"
+    type: enum
+    enum: "DiscourseAi::Configuration::PersonaEnumerator"
+    area: "ai-features/translation"
+  ai_translation_post_raw_translator_persona:
+    default: "-28"
+    type: enum
+    enum: "DiscourseAi::Configuration::PersonaEnumerator"
+    area: "ai-features/translation"
+  ai_translation_topic_title_translator_persona:
+    default: "-29"
+    type: enum
+    enum: "DiscourseAi::Configuration::PersonaEnumerator"
+    area: "ai-features/translation"
+  ai_translation_short_text_translator_persona:
+    default: "-30"
+    type: enum
+    enum: "DiscourseAi::Configuration::PersonaEnumerator"
+    area: "ai-features/translation"
   ai_translation_backfill_rate:
     default: 0
     min: 0
@@ -480,11 +500,6 @@ discourse_ai:
     client: false
     hidden: true
     area: "ai-features/translation"
-  ai_translation_locale_detection_persona:
-    default: "-27"
-    type: enum
-    enum: "DiscourseAi::Configuration::PersonaEnumerator"
-    area: "ai-features/translation"
 
   inferred_concepts_enabled:
     default: false

lib/configuration/feature.rb

@@ -126,6 +126,24 @@ def translation_features
               DiscourseAi::Configuration::Module::TRANSLATION_ID,
               DiscourseAi::Configuration::Module::TRANSLATION,
             ),
+            new(
+              "post_raw_translator",
+              "ai_translation_post_raw_translator_persona",
+              DiscourseAi::Configuration::Module::TRANSLATION_ID,
+              DiscourseAi::Configuration::Module::TRANSLATION,
+            ),
+            new(
+              "topic_title_translator",
+              "ai_translation_topic_title_translator_persona",
+              DiscourseAi::Configuration::Module::TRANSLATION_ID,
+              DiscourseAi::Configuration::Module::TRANSLATION,
+            ),
+            new(
+              "short_text_translator",
+              "ai_translation_short_text_translator_persona",
+              DiscourseAi::Configuration::Module::TRANSLATION_ID,
+              DiscourseAi::Configuration::Module::TRANSLATION,
+            ),
           ]
         end
 

lib/personas/bot_context.rb

@@ -25,6 +25,7 @@ class BotContext
 
       def initialize(
         post: nil,
+        topic: nil,
         participants: nil,
         user: nil,
         skip_tool_details: nil,
@@ -70,7 +71,14 @@ def initialize(
           @topic_id = post.topic_id
           @private_message = post.topic.private_message?
           @participants ||= post.topic.allowed_users.map(&:username).join(", ") if @private_message
-          @user = post.user
+          @user ||= post.user
+        end
+
+        if topic
+          @topic_id ||= topic.id
+          @private_message ||= topic.private_message?
+          @participants ||= topic.allowed_users.map(&:username).join(", ") if @private_message
+          @user ||= topic.user
         end
       end
 

lib/personas/persona.rb

@@ -65,6 +65,9 @@ def system_personas
             Translator => -25,
             ImageCaptioner => -26,
             LocaleDetection => -27,
+            PostRawTranslator => -28,
+            TopicTitleTranslator => -29,
+            ShortTextTranslator => -30,
           }
         end
 

lib/personas/post_raw_translator.rb

@@ -0,0 +1,58 @@
+# frozen_string_literal: true
+
+module DiscourseAi
+  module Personas
+    class PostRawTranslator < Persona
+      def self.default_enabled
+        false
+      end
+
+      def system_prompt
+        <<~PROMPT.strip
+          You are a highly skilled translator tasked with translating content from one language to another. Your goal is to provide accurate and contextually appropriate translations while preserving the original structure and formatting of the content. Follow these instructions carefully:
+
+          Translation Instructions:
+          1. Translate the content accurately while preserving any Markdown, HTML elements, or newlines.
+          2. Maintain the original document structure including headings, lists, tables, code blocks, etc.
+          3. Preserve all links, images, and other media references without translation.
+          4. Handle code snippets appropriately:
+            - Do not translate variable names, functions, or syntax within code blocks (```).
+            - Translate comments within code blocks.
+          5. For technical terminology:
+            - Provide the accepted target language term if it exists.
+            - If no equivalent exists, transliterate the term and include the original term in parentheses.
+          6. For ambiguous terms or phrases, choose the most contextually appropriate translation.
+          7. Do not add any content besides the translation.
+          8. Ensure the translation only contains the original language and the target language.
+
+          The text to translate will be provided in JSON format with the following structure:
+          {"content": "Text to translate", "target_locale": "Target language code"}
+
+          Output your translation in the following JSON format:
+          {"translation": "Your translated text here"}
+
+          Here are three examples of correct translations:
+
+          Original: {"content":"New Update for Minecraft Adds Underwater Temples", "target_locale":"Spanish"}
+          Correct translation: {"translation": "Nueva actualización para Minecraft añade templos submarinos"}
+
+          Original: {"content": "# Machine Learning 101\n\nMachine Learning (ML) is a subset of Artificial Intelligence (AI) that focuses on the development of algorithms and statistical models that enable computer systems to improve their performance on a specific task through experience.\n\n## Key Concepts\n\n1. **Supervised Learning**: The algorithm learns from labeled training data.\n2. **Unsupervised Learning**: The algorithm finds patterns in unlabeled data.\n3. **Reinforcement Learning**: The algorithm learns through interaction with an environment.\n\n```python\n# Simple example of a machine learning model\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.linear_model import LogisticRegression\n\n# Assuming X and y are your features and target variables\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n\nmodel = LogisticRegression()\nmodel.fit(X_train, y_train)\n\n# Evaluate the model\naccuracy = model.score(X_test, y_test)\nprint(f'Model accuracy: {accuracy}')\n```\n\nFor more information, visit [Machine Learning on Wikipedia](https://en.wikipedia.org/wiki/Machine_learning).", "target_locale":"French"}
+          Correct translation: {"translation": "# Machine Learning 101\n\nLe Machine Learning (ML) est un sous-ensemble de l'Intelligence Artificielle (IA) qui se concentre sur le développement d'algorithmes et de modèles statistiques permettant aux systèmes informatiques d'améliorer leurs performances sur une tâche spécifique grâce à l'expérience.\n\n## Concepts clés\n\n1. **Apprentissage supervisé** : L'algorithme apprend à partir de données d'entraînement étiquetées.\n2. **Apprentissage non supervisé** : L'algorithme trouve des motifs dans des données non étiquetées.\n3. **Apprentissage par renforcement** : L'algorithme apprend à travers l'interaction avec un environnement.\n\n```python\n# Exemple simple d'un modèle de machine learning\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.linear_model import LogisticRegression\n\n# En supposant que X et y sont vos variables de caractéristiques et cibles\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n\nmodel = LogisticRegression()\nmodel.fit(X_train, y_train)\n\n# Évaluer le modèle\naccuracy = model.score(X_test, y_test)\nprint(f'Model accuracy: {accuracy}')\n```\n\nPour plus d'informations, visitez [Machine Learning sur Wikipedia](https://en.wikipedia.org/wiki/Machine_learning)."}
+
+          Original: {"content": "**Heathrow fechado**: paralisação de voos deve continuar nos próximos dias, diz gestora do aeroporto de *Londres*", "target_locale": "English"}
+          Correct translation: {"translation": "**Heathrow closed**: flight disruption expected to continue in coming days, says *London* airport management"}
+
+          Remember, you are being consumed via an API. Only return the translated text in the specified JSON format. Do not include any additional information or explanations in your response.
+        PROMPT
+      end
+
+      def response_format
+        [{ "key" => "translation", "type" => "string" }]
+      end
+
+      def temperature
+        0.3
+      end
+    end
+  end
+end

lib/personas/short_text_translator.rb

@@ -0,0 +1,50 @@
+# frozen_string_literal: true
+
+module DiscourseAi
+  module Personas
+    class ShortTextTranslator < Persona
+      def self.default_enabled
+        false
+      end
+
+      def system_prompt
+        <<~PROMPT.strip
+          You are a translation service specializing in translating short pieces of text or a few words.
+          These words may be things like a name, description, or title. Adhere to the following guidelines:
+
+          1. Keep proper nouns and technical terms in their original language
+          2. Keep the translated content close to the original length
+          3. Translation maintains the original meaning
+          4. Preserving any Markdown, HTML elements, links, parenthesis, or newlines
+
+          The text to translate will be provided in JSON format with the following structure:
+          {"content": "Text to translate", "target_locale": "Target language code"}
+
+          Provide your translation in the following JSON format:
+          {"translation": "target_locale translation here"}
+
+          Here are three examples of correct translation
+
+          Original: {"content":"Japan", "target_locale":"es"}
+          Correct translation: {"translation": "Japón"}
+
+          Original: {"content":"Cats and Dogs", "target_locale":"zh_CN"}
+          Correct translation: {"translation": "猫和狗"}
+
+          Original: {"content": "Q&A", "target_locale": "pt"}
+          Correct translation: {"translation": "Perguntas e Respostas"}
+
+          Remember to keep proper nouns like "Minecraft" and "Toyota" in their original form. Translate the text now and provide your answer in the specified JSON format.
+        PROMPT
+      end
+
+      def response_format
+        [{ "key" => "translation", "type" => "string" }]
+      end
+
+      def temperature
+        0.3
+      end
+    end
+  end
+end

lib/personas/topic_title_translator.rb

@@ -0,0 +1,57 @@
+# frozen_string_literal: true
+
+module DiscourseAi
+  module Personas
+    class TopicTitleTranslator < Persona
+      def self.default_enabled
+        false
+      end
+
+      def system_prompt
+        <<~PROMPT.strip
+          You are a translation service specializing in translating forum post titles from English to the asked target_locale. Your task is to provide accurate and contextually appropriate translations while adhering to the following guidelines:
+
+          1. Translate the given title from English to target_locale asked.
+          2. Keep proper nouns and technical terms in their original language.
+          3. Attempt to keep the translated title length close to the original when possible.
+          4. Ensure the translation maintains the original meaning and tone.
+
+          To complete this task:
+
+          1. Read and understand the title carefully.
+          2. Identify any proper nouns or technical terms that should remain untranslated.
+          3. Translate the remaining words and phrases into the target_locale, ensuring the meaning is preserved.
+          4. Adjust the translation if necessary to keep the length similar to the original title.
+          5. Review your translation for accuracy and naturalness in the target_locale.
+
+          The text to translate will be provided in JSON format with the following structure:
+          {"content": "Title to translate", "target_locale": "Target language code"}
+
+          Provide your translation in the following JSON format:
+          {"translation": "Your target_locale translation here"}
+
+          Here are three examples of correct translation
+
+          Original: {"content":"New Update for Minecraft Adds Underwater Temples", "target_locale":"es"}
+          Correct translation: {"translation": "Nueva actualización para Minecraft añade templos submarinos"}
+
+          Original: {"content":"Toyota announces revolutionary battery technology", "target_locale":"fr"}
+          Correct translation: {"translation": "Toyota annonce une technologie de batteries révolutionnaire"}
+
+          Original: {"content": "Heathrow fechado: paralisação de voos deve continuar nos próximos dias, diz gestora do aeroporto de Londres", "target_locale": "en"}
+          Correct translation: {"translation": "Heathrow closed: flight disruption expected to continue in coming days, says London airport management"}
+
+          Remember to keep proper nouns like "Minecraft" and "Toyota" in their original form. Translate the title now and provide your answer in the specified JSON format.
+        PROMPT
+      end
+
+      def response_format
+        [{ "key" => "translation", "type" => "string" }]
+      end
+
+      def temperature
+        0.3
+      end
+    end
+  end
+end

lib/translation/base_translator.rb

@@ -3,62 +3,63 @@
 module DiscourseAi
   module Translation
     class BaseTranslator
-      def initialize(text:, target_locale:, topic_id: nil, post_id: nil)
+      def initialize(text:, target_locale:, topic: nil, post: nil)
         @text = text
         @target_locale = target_locale
-        @topic_id = topic_id
-        @post_id = post_id
+        @topic = topic
+        @post = post
       end
 
       def translate
-        prompt =
-          DiscourseAi::Completions::Prompt.new(
-            prompt_template,
-            messages: [{ type: :user, content: formatted_content, id: "user" }],
-            topic_id: @topic_id,
-            post_id: @post_id,
+        return nil if !SiteSetting.ai_translation_enabled
+        if (ai_persona = AiPersona.find_by(id: persona_setting)).blank?
+          return nil
+        end
+
+        persona_klass = ai_persona.class_instance
+        persona = persona_klass.new
+
+        llm_model = LlmModel.find_by(id: preferred_llm_model(persona_klass))
+        return nil if llm_model.blank?
+
+        bot =
+          DiscourseAi::Personas::Bot.as(
+            ai_persona.user || Discourse.system_user,
+            persona: persona,
+            model: llm_model,
           )
 
-        structured_output =
-          DiscourseAi::Completions::Llm.proxy(SiteSetting.ai_translation_model).generate(
-            prompt,
-            user: Discourse.system_user,
+        context =
+          DiscourseAi::Personas::BotContext.new(
+            user: ai_persona.user || Discourse.system_user,
+            skip_tool_details: true,
             feature_name: "translation",
-            response_format: response_format,
+            messages: [{ type: :user, content: formatted_content }],
+            topic: @topic,
+            post: @post,
           )
 
+        structured_output = nil
+        bot.reply(context) do |partial, _, type|
+          structured_output = partial if type == :structured_output
+        end
+
         structured_output&.read_buffered_property(:translation)
       end
 
       def formatted_content
         { content: @text, target_locale: @target_locale }.to_json
       end
 
-      def response_format
-        {
-          type: "json_schema",
-          json_schema: {
-            name: "reply",
-            schema: {
-              type: "object",
-              properties: {
-                translation: {
-                  type: "string",
-                },
-              },
-              required: ["translation"],
-              additionalProperties: false,
-            },
-            strict: true,
-          },
-        }
-      end
-
       private
 
-      def prompt_template
+      def persona_setting
         raise NotImplementedError
       end
+
+      def preferred_llm_model(persona_klass)
+        persona_klass.default_llm_id || SiteSetting.ai_translation_model&.split(":")&.last
+      end
     end
   end
 end