discourse-ai/app/models/llm_model.rb
Sam 5e80f93e4c
FEATURE: PDF support for rag pipeline (#1118)
This PR introduces several enhancements and refactorings to the AI Persona and RAG (Retrieval-Augmented Generation) functionalities within the discourse-ai plugin. Here's a breakdown of the changes:

**1. LLM Model Association for RAG and Personas:**

-   **New Database Columns:** Adds `rag_llm_model_id` to both `ai_personas` and `ai_tools` tables. This allows specifying a dedicated LLM for RAG indexing, separate from the persona's primary LLM.  Adds `default_llm_id` and `question_consolidator_llm_id` to `ai_personas`.
-   **Migration:**  Includes a migration (`20250210032345_migrate_persona_to_llm_model_id.rb`) to populate the new `default_llm_id` and `question_consolidator_llm_id` columns in `ai_personas` based on the existing `default_llm` and `question_consolidator_llm` string columns, and a post migration to remove the latter.
-   **Model Changes:**  The `AiPersona` and `AiTool` models now `belong_to` an `LlmModel` via `rag_llm_model_id`. The `LlmModel.proxy` method now accepts an `LlmModel` instance instead of just an identifier.  `AiPersona` now has `default_llm_id` and `question_consolidator_llm_id` attributes.
-   **UI Updates:**  The AI Persona and AI Tool editors in the admin panel now allow selecting an LLM for RAG indexing (if PDF/image support is enabled).  The RAG options component displays an LLM selector.
-   **Serialization:** The serializers (`AiCustomToolSerializer`, `AiCustomToolListSerializer`, `LocalizedAiPersonaSerializer`) have been updated to include the new `rag_llm_model_id`, `default_llm_id` and `question_consolidator_llm_id` attributes.

**2. PDF and Image Support for RAG:**

-   **Site Setting:** Introduces a new hidden site setting, `ai_rag_pdf_images_enabled`, to control whether PDF and image files can be indexed for RAG. This defaults to `false`.
-   **File Upload Validation:** The `RagDocumentFragmentsController` now checks the `ai_rag_pdf_images_enabled` setting and allows PDF, PNG, JPG, and JPEG files if enabled.  Error handling is included for cases where PDF/image indexing is attempted with the setting disabled.
-   **PDF Processing:** Adds a new utility class, `DiscourseAi::Utils::PdfToImages`, which uses ImageMagick (`magick`) to convert PDF pages into individual PNG images. A maximum PDF size and conversion timeout are enforced.
-   **Image Processing:** A new utility class, `DiscourseAi::Utils::ImageToText`, is included to handle OCR for the images and PDFs.
-   **RAG Digestion Job:** The `DigestRagUpload` job now handles PDF and image uploads. It uses `PdfToImages` and `ImageToText` to extract text and create document fragments.
-   **UI Updates:**  The RAG uploader component now accepts PDF and image file types if `ai_rag_pdf_images_enabled` is true. The UI text is adjusted to indicate supported file types.

**3. Refactoring and Improvements:**

-   **LLM Enumeration:** The `DiscourseAi::Configuration::LlmEnumerator` now provides a `values_for_serialization` method, which returns a simplified array of LLM data (id, name, vision_enabled) suitable for use in serializers. This avoids exposing unnecessary details to the frontend.
-   **AI Helper:** The `AiHelper::Assistant` now takes optional `helper_llm` and `image_caption_llm` parameters in its constructor, allowing for greater flexibility.
-   **Bot and Persona Updates:** Several updates were made across the codebase, changing the string based association to a LLM to the new model based.
-   **Audit Logs:** The `DiscourseAi::Completions::Endpoints::Base` now formats raw request payloads as pretty JSON for easier auditing.
- **Eval Script:** An evaluation script is included.

**4. Testing:**

-    The PR introduces a new eval system for LLMs, this allows us to test how functionality works across various LLM providers. This lives in `/evals`
2025-02-14 12:15:07 +11:00

175 lines
4.5 KiB
Ruby

# frozen_string_literal: true
class LlmModel < ActiveRecord::Base
FIRST_BOT_USER_ID = -1200
BEDROCK_PROVIDER_NAME = "aws_bedrock"
has_many :llm_quotas, dependent: :destroy
belongs_to :user
validates :display_name, presence: true, length: { maximum: 100 }
validates :tokenizer, presence: true, inclusion: DiscourseAi::Completions::Llm.tokenizer_names
validates :provider, presence: true, inclusion: DiscourseAi::Completions::Llm.provider_names
validates :url, presence: true, unless: -> { provider == BEDROCK_PROVIDER_NAME }
validates_presence_of :name, :api_key
validates :max_prompt_tokens, numericality: { greater_than: 0 }
validate :required_provider_params
scope :in_use,
-> do
model_ids = DiscourseAi::Configuration::LlmEnumerator.global_usage.keys
where(id: model_ids)
end
def self.provider_params
{
aws_bedrock: {
access_key_id: :text,
region: :text,
disable_native_tools: :checkbox,
},
anthropic: {
disable_native_tools: :checkbox,
},
open_ai: {
organization: :text,
disable_native_tools: :checkbox,
disable_streaming: :checkbox,
reasoning_effort: {
type: :enum,
values: %w[default low medium high],
default: "default",
},
},
mistral: {
disable_native_tools: :checkbox,
},
google: {
disable_native_tools: :checkbox,
},
azure: {
disable_native_tools: :checkbox,
},
hugging_face: {
disable_system_prompt: :checkbox,
},
vllm: {
disable_system_prompt: :checkbox,
},
ollama: {
disable_system_prompt: :checkbox,
enable_native_tool: :checkbox,
disable_streaming: :checkbox,
},
open_router: {
disable_native_tools: :checkbox,
provider_order: :text,
provider_quantizations: :text,
disable_streaming: :checkbox,
},
}
end
def to_llm
DiscourseAi::Completions::Llm.proxy(self)
end
def identifier
"custom:#{id}"
end
def toggle_companion_user
return if name == "fake" && Rails.env.production?
enable_check = SiteSetting.ai_bot_enabled && enabled_chat_bot
if enable_check
if !user
next_id = DB.query_single(<<~SQL).first
SELECT min(id) - 1 FROM users
SQL
new_user =
User.new(
id: [FIRST_BOT_USER_ID, next_id].min,
email: "no_email_#{SecureRandom.hex}",
name: name.titleize,
username: UserNameSuggester.suggest(name),
active: true,
approved: true,
admin: true,
moderator: true,
trust_level: TrustLevel[4],
)
new_user.save!(validate: false)
self.update!(user: new_user)
else
user.active = true
user.save!(validate: false)
end
elsif user
# will include deleted
has_posts = DB.query_single("SELECT 1 FROM posts WHERE user_id = #{user.id} LIMIT 1").present?
if has_posts
user.update!(active: false) if user.active
else
user.destroy!
self.update!(user: nil)
end
end
end
def tokenizer_class
tokenizer.constantize
end
def lookup_custom_param(key)
provider_params&.dig(key)
end
def seeded?
id.present? && id < 0
end
def api_key
if seeded?
env_key = "DISCOURSE_AI_SEEDED_LLM_API_KEY_#{id.abs}"
ENV[env_key] || self[:api_key]
else
self[:api_key]
end
end
private
def required_provider_params
return if provider != BEDROCK_PROVIDER_NAME
%w[access_key_id region].each do |field|
if lookup_custom_param(field).blank?
errors.add(:base, I18n.t("discourse_ai.llm_models.missing_provider_param", param: field))
end
end
end
end
# == Schema Information
#
# Table name: llm_models
#
# id :bigint not null, primary key
# display_name :string
# name :string not null
# provider :string not null
# tokenizer :string not null
# max_prompt_tokens :integer not null
# created_at :datetime not null
# updated_at :datetime not null
# url :string
# api_key :string
# user_id :integer
# enabled_chat_bot :boolean default(FALSE), not null
# provider_params :jsonb
# vision_enabled :boolean default(FALSE), not null
#