Key Features

How It Works

The module creates a seamless integration between the Document Page knowledge base and the LLM RAG system:

Usage Guide

Creating RAG Documents from the UI

Processing Entire Categories

You can also add entire categories of documents to your LLM RAG system programmatically:

from odoo import api, SUPERUSER_ID

def add_knowledge_base_to_rag(env, category_id, collection_id):
    # Get all pages in this category
    pages = env['document.page'].search([
        ('parent_id', '=', category_id),
        ('type', '=', 'content')
    ])

    # Get model ID for document.page
    model_id = env['ir.model'].search([('model', '=', 'document.page')], limit=1).id

    # Create documents
    created_docs = env['llm.document']
    for page in pages:
        # Create document with reference to the page
        document = env['llm.document'].create({
            'name': f"{page.name} - Knowledge Base",
            'model_id': model_id,
            'res_id': page.id,
            'collection_ids': [(4, collection_id)]
        })

        # Process through RAG pipeline
        document.process_document()
        created_docs |= document

    return created_docs

# Example usage
env = api.Environment(cr, SUPERUSER_ID, {})
category_id = 1  # ID of your document.page category
collection_id = 1  # ID of your llm.document.collection

docs = add_knowledge_base_to_rag(env, category_id, collection_id)

Technical Implementation

The module works by implementing a custom rag_parse method on the document.page model. This method is called by the LLM RAG module during document processing.

Sample Output

Here's an example of how a document page is converted to the RAG format:

# Product Return Policy

## Metadata
**Last Contributor:** Administrator
**Last Updated:** 2023-05-15 14:30:22
**Category:** Policies

## Content
<p>Our product return policy allows customers to return items within 30 days of purchase.</p>
<p>To qualify for a return, items must be:</p>
<ul>
  <li>In original condition</li>
  <li>With original packaging</li>
  <li>Accompanied by a receipt or proof of purchase</li>
</ul>

## Related Pages
- [Return Process](/web#id=12&model=document.page&view_type=form)
- [Refund Policy](/web#id=15&model=document.page&view_type=form)

Benefits for Your Organization

• Unified Knowledge Base: Make your existing documentation available to your AI systems
• Improved AI Responses: LLMs can answer questions based on your internal knowledge
• Reduced Duplication: Maintain documentation in one place and use it everywhere
• Up-to-date Information: Changes to your knowledge base can be reflected in the LLM system

Installation

1. Ensure both document_page and llm_rag modules are installed
2. Install this module (llm_document_page)
3. No additional configuration is needed - the integration is automatic