feat: Route system LLM jobs (4/6) (#14843)
## Description Routes the remaining system-only and legacy-sensitive LLM jobs through feature-level model configuration, while preserving system credential usage and usage-accounting behavior. This adds dedicated defaults for help center article generation, onboarding content generation, query translation, transcription, and search embeddings so these flows can be configured per account without falling back to installation-wide model settings. Fixes https://linear.app/chatwoot/issue/CW-7425/test-new-models ## Type of change - [ ] Bug fix (non-breaking change which fixes an issue) - [x] New feature (non-breaking change which adds functionality) - [ ] Breaking change (fix or feature that would cause existing functionality not to work as expected) - [ ] This change requires a documentation update ## How Has This Been Tested? Verified the feature routing defaults and account overrides for the touched Captain/system LLM paths, including the legacy OpenAI transcription and paginated FAQ services. - `eval "$(rbenv init -)" && bundle exec rspec spec/lib/captain/base_task_service_spec.rb spec/lib/llm/models_spec.rb spec/models/concerns/captain_featurable_spec.rb spec/controllers/api/v1/accounts/captain/preferences_controller_spec.rb spec/enterprise/services/captain/llm/paginated_faq_generator_service_spec.rb spec/enterprise/services/captain/llm/pdf_processing_service_spec.rb spec/enterprise/services/messages/audio_transcription_service_spec.rb spec/enterprise/services/onboarding/help_center_article_builder_spec.rb spec/enterprise/services/captain/onboarding/website_analyzer_service_spec.rb` - `eval "$(rbenv init -)" && bundle exec rubocop app/controllers/api/v1/accounts/captain/preferences_controller.rb app/models/concerns/account_settings_schema.rb lib/captain/base_task_service.rb enterprise/app/services/captain/llm/article_translation_service.rb enterprise/app/services/captain/llm/article_writer_service.rb enterprise/app/services/captain/llm/embedding_service.rb enterprise/app/services/captain/llm/help_center_curation_service.rb enterprise/app/services/captain/llm/paginated_faq_generator_service.rb enterprise/app/services/captain/llm/translate_query_service.rb enterprise/app/services/captain/llm/widget_tagline_service.rb enterprise/app/services/captain/onboarding/website_analyzer_service.rb enterprise/app/services/messages/audio_transcription_service.rb spec/enterprise/services/messages/audio_transcription_service_spec.rb spec/lib/captain/base_task_service_spec.rb` - `ruby -e "require 'yaml'; config = YAML.load_file('config/llm.yml'); %w[document_faq_generation help_center_article_generation onboarding_content_generation help_center_query_translation audio_transcription help_center_search].each { |feature| abort(%(missing #{feature})) unless config.dig('features', feature) }; abort('wrong article default') unless config.dig('features', 'help_center_article_generation', 'default') == 'gpt-5.2'; puts 'llm.yml ok'"` - `git diff --check` ## Checklist: - [x] My code follows the style guidelines of this project - [x] I have performed a self-review of my code - [x] I have commented on my code, particularly in hard-to-understand areas - [ ] I have made corresponding changes to the documentation - [x] My changes generate no new warnings - [x] I have added tests that prove my fix is effective or that my feature works - [x] New and existing unit tests pass locally with my changes - [ ] Any dependent changes have been merged and published in downstream modules
This commit is contained in:
Sony Mathew
@@ -126,6 +126,27 @@ features:
|
||||
pdf_faq_generation:
|
||||
models: [gpt-4.1-mini, gpt-5-mini, gpt-4.1, gpt-5.1, gpt-5.2]
|
||||
default: gpt-4.1-mini
|
||||
help_center_article_generation:
|
||||
models:
|
||||
[
|
||||
gpt-4.1-mini,
|
||||
gpt-5-mini,
|
||||
gpt-4.1,
|
||||
gpt-5.1,
|
||||
gpt-5.2,
|
||||
claude-haiku-4.5,
|
||||
claude-sonnet-4.5,
|
||||
gemini-3-flash,
|
||||
gemini-3-pro,
|
||||
]
|
||||
default: gpt-5.2
|
||||
onboarding_content_generation:
|
||||
models:
|
||||
[gpt-4.1, gpt-4.1-mini, gpt-5-mini, gpt-5.1, gpt-5.2]
|
||||
default: gpt-4.1
|
||||
help_center_query_translation:
|
||||
models: [gpt-4.1-nano, gpt-4.1-mini, gpt-5-mini]
|
||||
default: gpt-4.1-nano
|
||||
audio_transcription:
|
||||
models: [whisper-1]
|
||||
default: whisper-1
|
||||
|
||||
@@ -6,7 +6,7 @@ class Captain::Llm::ArticleTranslationService < Captain::BaseTaskService
|
||||
def perform
|
||||
raise ArgumentError, "Invalid type: #{type}" unless TYPES.include?(type)
|
||||
|
||||
response = make_api_call(model: translation_model, messages: messages)
|
||||
response = make_api_call(feature: 'help_center_article_generation', messages: messages)
|
||||
return response if response[:error]
|
||||
|
||||
response.merge(message: response[:message].strip)
|
||||
@@ -33,10 +33,6 @@ class Captain::Llm::ArticleTranslationService < Captain::BaseTaskService
|
||||
@llm_credential ||= system_llm_credential
|
||||
end
|
||||
|
||||
def translation_model
|
||||
@translation_model ||= InstallationConfig.find_by(name: 'CAPTAIN_OPEN_AI_MODEL')&.value.presence || GPT_MODEL
|
||||
end
|
||||
|
||||
def title_system_prompt
|
||||
<<~SYSTEM_PROMPT_MESSAGE
|
||||
You are a professional translator.
|
||||
|
||||
@@ -6,7 +6,7 @@ class Captain::Llm::ArticleWriterService < Captain::BaseTaskService
|
||||
pattr_initialize [:account!, :source_pages!, { hint_title: nil }]
|
||||
|
||||
def perform
|
||||
response = make_api_call(model: writer_model, messages: messages, schema: RESPONSE_SCHEMA)
|
||||
response = make_api_call(feature: 'help_center_article_generation', messages: messages, schema: RESPONSE_SCHEMA)
|
||||
return response if response[:error]
|
||||
|
||||
response.merge(message: extract_payload(response[:message]))
|
||||
@@ -92,10 +92,6 @@ class Captain::Llm::ArticleWriterService < Captain::BaseTaskService
|
||||
false
|
||||
end
|
||||
|
||||
def writer_model
|
||||
'gpt-5.2'
|
||||
end
|
||||
|
||||
def build_follow_up_context?
|
||||
false
|
||||
end
|
||||
|
||||
@@ -6,7 +6,7 @@ class Captain::Llm::EmbeddingService
|
||||
def initialize(account_id: nil)
|
||||
Llm::Config.initialize!
|
||||
@account_id = account_id
|
||||
@embedding_model = InstallationConfig.find_by(name: 'CAPTAIN_EMBEDDING_MODEL')&.value.presence || LlmConstants::DEFAULT_EMBEDDING_MODEL
|
||||
@embedding_model = self.class.embedding_model
|
||||
end
|
||||
|
||||
def self.embedding_model
|
||||
|
||||
@@ -9,7 +9,7 @@ class Captain::Llm::HelpCenterCurationService < Captain::BaseTaskService
|
||||
pattr_initialize [:account!, :links!]
|
||||
|
||||
def perform
|
||||
response = make_api_call(model: CURATION_MODEL, messages: messages, schema: RESPONSE_SCHEMA)
|
||||
response = make_api_call(feature: 'onboarding_content_generation', model: CURATION_MODEL, messages: messages, schema: RESPONSE_SCHEMA)
|
||||
return response if response[:error]
|
||||
|
||||
response.merge(message: extract_payload(response[:message]))
|
||||
|
||||
@@ -1,6 +1,4 @@
|
||||
class Captain::Llm::TranslateQueryService < Captain::BaseTaskService
|
||||
MODEL = 'gpt-4.1-nano'.freeze
|
||||
|
||||
pattr_initialize [:account!]
|
||||
|
||||
def translate(query, target_language:)
|
||||
@@ -11,7 +9,7 @@ class Captain::Llm::TranslateQueryService < Captain::BaseTaskService
|
||||
{ role: 'user', content: query }
|
||||
]
|
||||
|
||||
response = make_api_call(model: MODEL, messages: messages)
|
||||
response = make_api_call(feature: 'help_center_query_translation', messages: messages)
|
||||
return query if response[:error]
|
||||
|
||||
response[:message].strip
|
||||
|
||||
@@ -4,7 +4,7 @@ class Captain::Llm::WidgetTaglineService < Captain::BaseTaskService
|
||||
pattr_initialize [:account!]
|
||||
|
||||
def perform
|
||||
response = make_api_call(model: tagline_model, messages: messages, schema: RESPONSE_SCHEMA)
|
||||
response = make_api_call(feature: 'onboarding_content_generation', messages: messages, schema: RESPONSE_SCHEMA)
|
||||
return response if response[:error]
|
||||
|
||||
response.merge(message: extract_tagline(response[:message]))
|
||||
@@ -68,10 +68,6 @@ class Captain::Llm::WidgetTaglineService < Captain::BaseTaskService
|
||||
false
|
||||
end
|
||||
|
||||
def tagline_model
|
||||
@tagline_model ||= InstallationConfig.find_by(name: 'CAPTAIN_OPEN_AI_MODEL')&.value.presence || GPT_MODEL
|
||||
end
|
||||
|
||||
def build_follow_up_context?
|
||||
false
|
||||
end
|
||||
|
||||
@@ -4,7 +4,7 @@ class Captain::Onboarding::WebsiteAnalyzerService < Llm::BaseAiService
|
||||
MAX_CONTENT_LENGTH = 8000
|
||||
|
||||
def initialize(website_url)
|
||||
super()
|
||||
super(feature: 'onboarding_content_generation')
|
||||
@website_url = normalize_url(website_url)
|
||||
@website_content = nil
|
||||
@favicon_url = nil
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
class Messages::AudioTranscriptionService< Llm::LegacyBaseOpenAiService
|
||||
include Integrations::LlmInstrumentation
|
||||
|
||||
TRANSCRIPTION_MODEL = 'gpt-4o-mini-transcribe'.freeze
|
||||
TRANSCRIPTION_MODEL = 'whisper-1'.freeze
|
||||
# OpenAI's transcription endpoint hard limit is 25 MB *decimal* (25_000_000), not
|
||||
# binary (25.megabytes = 26_214_400) — using the binary form leaks the 25.0–26.2 MB
|
||||
# range to the API as 413s. Long audio (~70+ min Opus) keeps the attachment but skips
|
||||
@@ -15,6 +15,7 @@ class Messages::AudioTranscriptionService< Llm::LegacyBaseOpenAiService
|
||||
@attachment = attachment
|
||||
@message = attachment.message
|
||||
@account = message.account
|
||||
@transcription_model = Llm::FeatureRouter.resolve(feature: 'audio_transcription', account: account)[:model]
|
||||
end
|
||||
|
||||
def perform
|
||||
@@ -81,7 +82,7 @@ class Messages::AudioTranscriptionService< Llm::LegacyBaseOpenAiService
|
||||
# behaviour across OpenAI transcription models.
|
||||
response = @client.audio.transcribe(
|
||||
parameters: {
|
||||
model: TRANSCRIPTION_MODEL,
|
||||
model: transcription_model,
|
||||
file: file,
|
||||
temperature: 0.0
|
||||
}
|
||||
@@ -98,7 +99,7 @@ class Messages::AudioTranscriptionService< Llm::LegacyBaseOpenAiService
|
||||
def instrumentation_params(file_path)
|
||||
{
|
||||
span_name: 'llm.messages.audio_transcription',
|
||||
model: TRANSCRIPTION_MODEL,
|
||||
model: transcription_model,
|
||||
account_id: account&.id,
|
||||
feature_name: 'audio_transcription',
|
||||
file_path: file_path
|
||||
@@ -127,4 +128,8 @@ class Messages::AudioTranscriptionService< Llm::LegacyBaseOpenAiService
|
||||
'x-mp3' => 'mp3'
|
||||
}.fetch(subtype, subtype)
|
||||
end
|
||||
|
||||
def transcription_model
|
||||
@transcription_model || TRANSCRIPTION_MODEL
|
||||
end
|
||||
end
|
||||
|
||||
@@ -59,7 +59,10 @@ class Captain::BaseTaskService
|
||||
def resolved_model(model:, feature:)
|
||||
return model if feature.blank?
|
||||
|
||||
Llm::FeatureRouter.resolve(feature: feature, account: account)[:model]
|
||||
route = Llm::FeatureRouter.resolve(feature: feature, account: account)
|
||||
return model if model.present? && route[:source] == :default
|
||||
|
||||
route[:model]
|
||||
end
|
||||
|
||||
def execute_ruby_llm_request(model:, messages:, schema: nil, tools: [])
|
||||
|
||||
@@ -17,6 +17,7 @@ RSpec.describe Captain::Llm::ArticleTranslationService do
|
||||
|
||||
it 'returns the stripped translated title' do
|
||||
expect(service).to receive(:make_api_call) do |args|
|
||||
expect(args[:feature]).to eq('help_center_article_generation')
|
||||
expect(args[:messages][0][:content]).to include('professional translator')
|
||||
expect(args[:messages][0][:content]).to include(target_language)
|
||||
expect(args[:messages][1][:content]).to eq('Getting Started')
|
||||
|
||||
@@ -0,0 +1,38 @@
|
||||
require 'rails_helper'
|
||||
|
||||
RSpec.describe Captain::Llm::EmbeddingService, type: :service do
|
||||
def configure_embedding_model(value)
|
||||
InstallationConfig.find_or_initialize_by(name: 'CAPTAIN_EMBEDDING_MODEL').tap do |config|
|
||||
config.value = value
|
||||
config.locked = false
|
||||
config.save!
|
||||
end
|
||||
end
|
||||
|
||||
describe '.embedding_model' do
|
||||
it 'uses the installation embedding model when configured' do
|
||||
configure_embedding_model('custom-embedding-model')
|
||||
|
||||
expect(described_class.embedding_model).to eq('custom-embedding-model')
|
||||
end
|
||||
|
||||
it 'falls back to the default embedding model when the installation value is blank' do
|
||||
configure_embedding_model('')
|
||||
|
||||
expect(described_class.embedding_model).to eq(LlmConstants::DEFAULT_EMBEDDING_MODEL)
|
||||
end
|
||||
end
|
||||
|
||||
describe '#get_embedding' do
|
||||
let(:account) { create(:account) }
|
||||
let(:embedding_response) { double('embedding_response', vectors: [0.1, 0.2]) } # rubocop:disable RSpec/VerifiedDoubles
|
||||
|
||||
it 'sends the installation embedding model to RubyLLM' do
|
||||
configure_embedding_model('custom-embedding-model')
|
||||
|
||||
expect(RubyLLM).to receive(:embed).with('search text', model: 'custom-embedding-model').and_return(embedding_response)
|
||||
|
||||
expect(described_class.new(account_id: account.id).get_embedding('search text')).to eq([0.1, 0.2])
|
||||
end
|
||||
end
|
||||
end
|
||||
@@ -3,7 +3,7 @@ require 'rails_helper'
|
||||
RSpec.describe Messages::AudioTranscriptionService, type: :service do
|
||||
let(:account) { create(:account, audio_transcriptions: true) }
|
||||
let(:conversation) { create(:conversation, account: account) }
|
||||
let(:message) { create(:message, conversation: conversation) }
|
||||
let(:message) { create(:message, account: account, conversation: conversation) }
|
||||
let(:attachment) { message.attachments.create!(account: account, file_type: :audio) }
|
||||
|
||||
before do
|
||||
@@ -101,4 +101,29 @@ RSpec.describe Messages::AudioTranscriptionService, type: :service do
|
||||
FileUtils.rm_f(temp_file_path) if temp_file_path.present?
|
||||
end
|
||||
end
|
||||
|
||||
describe '#transcribe_audio' do
|
||||
let(:service) { described_class.new(attachment) }
|
||||
let(:audio_api) { double('audio_api') } # rubocop:disable RSpec/VerifiedDoubles
|
||||
let(:audio_file_path) { Rails.root.join('tmp/audio_transcription_service_spec.mp3').to_s }
|
||||
|
||||
before do
|
||||
File.binwrite(audio_file_path, 'audio')
|
||||
allow(service).to receive(:fetch_audio_file).and_return(audio_file_path)
|
||||
allow(service).to receive(:update_transcription)
|
||||
allow(service.client).to receive(:audio).and_return(audio_api)
|
||||
end
|
||||
|
||||
after do
|
||||
FileUtils.rm_f(audio_file_path)
|
||||
end
|
||||
|
||||
it 'uses the audio transcription feature model' do
|
||||
expect(audio_api).to receive(:transcribe).with(
|
||||
parameters: hash_including(model: 'whisper-1', temperature: 0.0)
|
||||
).and_return({ 'text' => 'Audio transcript' })
|
||||
|
||||
expect(service.send(:transcribe_audio)).to eq('Audio transcript')
|
||||
end
|
||||
end
|
||||
end
|
||||
|
||||
@@ -21,6 +21,7 @@ RSpec.describe Captain::BaseTaskService do
|
||||
let(:service) { test_service_class.new(account: account, conversation_display_id: conversation.display_id) }
|
||||
|
||||
before do
|
||||
InstallationConfig.where(name: 'CAPTAIN_OPEN_AI_API_KEY').destroy_all
|
||||
create(:installation_config, name: 'CAPTAIN_OPEN_AI_API_KEY', value: 'test-key')
|
||||
# Stub captain enabled check to allow OSS specs to test base functionality
|
||||
# without enterprise module interference
|
||||
@@ -178,6 +179,26 @@ RSpec.describe Captain::BaseTaskService do
|
||||
service.send(:make_api_call, feature: 'editor', messages: messages)
|
||||
end
|
||||
|
||||
it 'uses the supplied model as a feature fallback when there is no account override' do
|
||||
expect(mock_context).to receive(:chat).with(model: 'gpt-5.2').and_return(mock_chat)
|
||||
|
||||
service.send(:make_api_call, feature: 'document_faq_generation', model: 'gpt-5.2', messages: messages)
|
||||
end
|
||||
|
||||
it 'uses the help center article generation feature default' do
|
||||
expect(mock_context).to receive(:chat).with(model: 'gpt-5.2').and_return(mock_chat)
|
||||
|
||||
service.send(:make_api_call, feature: 'help_center_article_generation', messages: messages)
|
||||
end
|
||||
|
||||
it 'prefers account overrides over supplied feature fallback models' do
|
||||
account.update!(captain_models: { 'help_center_article_generation' => 'gpt-4.1' })
|
||||
|
||||
expect(mock_context).to receive(:chat).with(model: 'gpt-4.1').and_return(mock_chat)
|
||||
|
||||
service.send(:make_api_call, feature: 'help_center_article_generation', model: 'gpt-5.2', messages: messages)
|
||||
end
|
||||
|
||||
it 'returns formatted response with tokens' do
|
||||
result = service.send(:make_api_call, model: model, messages: messages)
|
||||
|
||||
|
||||
Reference in New Issue
Block a user