Pfannkuchensack/sqlite_invokeai_db_tool

InvokeAI DB Tool

This project is not affiliated with InvokeAI in any way.

A tool for managing and analyzing InvokeAI database entries and PNG metadata.

Features

Database Management

• Synchronization: Match between database and output directory
• Restoration: Add missing database entries for existing images
• Cleanup: Remove database entries for non-existent images
• Metadata Extraction: Automatic extraction and storage of InvokeAI metadata
• Workflow Detection: Identification and flagging of images with workflow information
• Model Path Management: Management of model paths in the database
• Model Type Management: Management of model types in the database

Installation

Prerequisites

• Node.js (recommended: version 14 or higher)
• npm or yarn
• Electron

Installation

# Clone repository
git clone https://github.com/Pfannkuchensack/sqlite_invokeai_db_tool.git
cd sqlite_invokeai_db_tool

# Install dependencies
npm install

# Start application
npm start

Package Creation

# For Windows
npm run package-win
# or
npm run make

Usage

Main Application

1. Start the application with npm start or use the compiled version
2. Select the SQLite database file from InvokeAI
3. Choose the output directory where the images are stored
4. Use the various functions for synchronization and management

Technical Details

Metadata Detection

The tool recognizes different types of InvokeAI metadata in PNG files:

• invokeai_metadata: Contains general information about the image generation
• invokeai_graph: Contains the processing graph with nodes and connections
• invokeai_workflow: Contains workflow information for reuse

Image Classification

Images are classified based on their metadata:

• Final Images: Images with invokeai_metadata
• Intermediate Images: Images with exclusively is_intermediate=true nodes
• Unknown Images: Images without recognizable InvokeAI metadata

License

MIT

Author

Pfannkuchensack

German version available here