EAEmbedFileToAIDB

This document was generated using Claude.ai

Overview

Processes PDF files attached to entities and converts them into AI-searchable format by storing them in a vector database. Enables AI assistants to search through document content.

When This Action Runs

  • Trigger: Manual execution through entity flows
  • Target: Entities with PDF attachments
  • Purpose: Make PDF content searchable by AI assistants
  • Timing: On-demand when AI document search is needed

How It Works

  1. Extracts PDF files from specified attachment fields
  2. Converts content to text and splits into 800-character chunks (200-character overlap)
  3. Creates vector embeddings using OpenAI's text-embedding-ada-002 model
  4. Stores embeddings in Milvus vector database for AI retrieval

Parameters

Parameter 1: Attachment IDs (Required)

Format: Field names separated by commas or newlines Examples:

  • attachment1 (single header field)
  • attachment1,attachment2 (multiple fields)
  • details.attachment1 (collection field)
  • attachment1,details.attachment1 (mixed fields)

Prerequisites

Required AI Module Configuration (System Settings)

  • Chat Provider API Key - AI service authentication
  • Chat Provider - GPT-4, GPT-4 Turbo, or DeepSeek
  • Vector Store URI - Milvus database connection
  • Vector Store Token - Database authentication
  • OpenAI Embedding Key - Required for embeddings

Database Requirements

  • Accessible Milvus vector database
  • Network connectivity to OpenAI services
  • Sufficient vector database storage

File Requirements

  • Supported: PDF files only
  • Duplicate Prevention: SHA-256 hash identification
  • Processing: 800-character chunks with 200-character overlap

Expected Results

Success Cases

  • New File: Successfully embedded into AI database
  • Duplicate File: Warning "The file already exists" (skipped)

Error Cases

  • Invalid File Type: "Only PDF files are supported"
  • Configuration Issues: Missing API keys or database connectivity
  • Field Issues: Attachment fields don't exist or are empty

Important Warnings

⚠️ Configuration Requirements

  • Complete Setup: All AI module settings must be configured
  • API Keys: Valid OpenAI and vector database credentials required
  • Network Access: Requires connectivity to external AI services

⚠️ Security Considerations

  • Content Exposure: PDF content becomes searchable by AI assistants
  • Sensitive Documents: Ensure AI access is appropriate for document content
  • API Usage: Each file generates OpenAI API calls (cost implications)
  • AI Module Setup - Required configuration for AI features
  • Vector Database Management - Manages searchable document storage

Module: ai

Full Class Name: com.namasoft.modules.ai.util.actions.EAEmbedFileToAIDB