πΌοΈ Terminus Tag Manager πΉ - v5.1 π
Welcome to the Image & Video Tag Hub, a comprehensive tool for organizing, tagging, and managing your local and remote media collections. This version integrates the new Civitai Resource Panel, enhances SFTP support, and lays the foundation for future AI-assisted organization.
A Look Ahead: Planned Features for the Next Major Release v5.10
We are actively developing the next version, which will introduce several powerful new capabilities and foundational improvements to enhance your workflow and user experience. Here is what is planned:
π οΈ Complete Code Refactor: The entire project is being re-coded from the ground up. This major undertaking will result in a cleaner, more efficient codebase, leading to significant improvements in performance, stability, and maintainability for all future updates.
π GitHub Integration & Automatic Updates: To streamline maintenance and ensure you always have the latest improvements, the project will be hosted on GitHub. An automatic update feature will be integrated, which will notify you of new versions and allow for easy, one-click updates. This feature can be disabled in the settings for users who prefer to update manually.
βοΈ AI-Powered Tag-to-Caption Conversion: For users who prefer descriptive captions over keyword lists, a new feature will provide automatic conversion. The system will intelligently interpret existing tags to generate a coherent, descriptive sentence, offering greater flexibility in metadata management.
β¨ Streamlined DeepDanbooru Training Interface: To make custom tag training accessible to everyone, we are building a new, user-friendly interface on top of the existing DeepDanbooru functionality. This will simplify the process of managing datasets and fine-tuning the model, especially for users who may be new to Python or find the current methods complex.
π§ Enhanced Civitai Integration: The Civitai panel will receive a significant upgrade. It will automatically scan tags for keywords and search for the corresponding assets on the Civitai website. For offline collections, it will support a local JSON file as a fallback. Results will be displayed with a preview image and name, with new options to navigate directly to the resource page or download the asset from within the application.
π·οΈ Custom Tag Sets with Autocomplete: This feature will introduce the ability to create and load custom sets of tags as a "project." When tagging, an autocomplete dropdown menu will suggest tags from your active set, significantly increasing tagging speed and ensuring consistency across your collection.
πΌοΈ Batch Tag Editing: A powerful multi-image editing tool will be introduced, allowing users to perform operations on a selection of files at once. This includes finding and replacing specific tags or adding new tags to multiple files simultaneously, eliminating the need for repetitive manual edits.
β¨ What's New in the Current v4.5?
β‘ Progressive SFTP Loading: Browse large directories on a remote server is now seamless. The application initially loads the first 50 images and automatically loads the next 50 as you scroll, preventing freezes and improving performance.
βΉοΈ Civitai Info Panel: The right-hand panel can be configured to display Civitai resources. When an image generated by Stable Diffusion is selected, this panel automatically searches for and displays the Checkpoint and LoRA models used, providing direct links to their Civitai pages.
βοΈ Revamped Settings: The settings window has been updated to allow you to choose which panel to display on the right: the Civitai Info panel, the File Info panel, or the Downloader.
π§ Requirements
Python: Python 3.x installed on your system.
VLC Media Player: For video playback, a standard installation of VLC is required.
Microsoft Visual C++: The Python version of this application requires the Microsoft Visual C++ Redistributable to be installed.
π₯ Installation
Download Files: Ensure the main application script (
image_text_viewer.py) and theextensionssubfolder are in the same directory.Install Libraries: You can either run the provided
Installation.batfile, or open a terminal in the application's directory and run:pip install -r requirements.txt
π‘ Tip: If you run into any trouble with the installation (like pip command issues or other errors), don't hesitate to use an AI assistant for help. π€ Just copy and paste the error message you receive and ask it for step-by-step instructions on how to fix it for your specific setup.
π Launching the Application
To start the hub, run the provided batch file from the application's root directory: image_text_viewer_launch.bat
β¨ Features & Usage
The main window is divided into three primary sections:
π Left Panel: Media thumbnails, tag search, and page navigation.
π₯οΈ Center Panel: A large preview of the selected image or video.
π Right Panel: Tag and metadata display, file operations, and a modular section for the Civitai Panel, SFTP client, or Downloader.
Getting Started
π Local Mode: Click the "..." button at the top to select a local folder containing your media. Thumbnails will then appear in the left panel.
π SFTP Mode: Go to Settings -> Viewer and enable "FTP Mode." The right-hand panel will transform into an SFTP client where you can connect to a remote server.
Navigating Your Media
Local Files: Click any thumbnail to select it for preview. Use the Prev / Next buttons to navigate through pages.
SFTP Files:
Click a directory name to navigate into it. The file list will update automatically.
Click the "Load" button next to a directory to scan and progressively load all of its images into the thumbnail panel on the left.
Double-click any preview in the center panel to open a full-screen view.
Managing Tags & Metadata
π·οΈ Tags: The "Tags" text box on the right displays the contents of the
.txtfile associated with the selected image. You can edit these tags directly.π€ SD Metadata: For AI-generated images, this section displays embedded generation parameters.
π Search: Use the search bar in the left panel to filter your loaded media by keywords found in their tag files.
File Operations
π Copy Tags: Copies the current tags to your clipboard.
π Move File: (Local Mode Only) Moves the selected media file and its
.txtfile to your designated output folder.ποΈ Delete File: (Local Mode Only) Permanently deletes the selected media file and its
.txtfile after confirmation.πΎ Save Tags: (SFTP Mode Only) Saves the current text in the tag editor back to the corresponding
.txtfile on the remote server.π Backup Encryption: An encryption option is available when backing up your files for added security.
π Known Issues
AI Downloader: The "Edit Text" function within the AI Downloader is currently not working.
EXE Application: The current executable file has known stability issues. For a better and cleaner experience, using the Python script is recommended.
π¨ Themes & Future Plans
Themes: The ability to change themes is included, but this feature is low-priority and will likely not see major updates unless a special edition is released.
π§ AI Assistant (Coming Soon): The groundwork for a local, offline AI assistant is being laid. This optional feature will help automate tasks like file conversion, cleanup, and organization.
Description
FAQ
Comments (5)
Seems like a pretty neat tool. I do have a habit of downloading a lot of images for reference/inspiration and then it's a case of figuring out what assets were used to achieve the effect after the fact. So I like the link out to see which Loras/models are used.
Shower thoughts: -
Direct Web-forge integration would cool, as an active extension.
Ability to scan your existing model folders and highlight which models you already have/which ones you don't.
Ability to add commonly visited Folders as Favourites. I download to a Civit image folder for instance, so that's always going to be my go to.
I'd expand the gallery view so it shows more than one row of thumbnails.
Navigate the gallery up and down and across pages of the galley using the arrow keys.
Option to have the gallery display at the bottom.
Ability to favourite images and just see favourites, and/or build custom themed collections. You might be looking across a whole bunch of images for an idea for a particular look, this way you can keep tabs on them.
Mark images without meta data and maybe the option to exclude them.
Anyway, cool stuff.
Thanks for feedback! These are excellent suggestions, and I'll definitely be working on adding them in a future update.
Just wanted to give you a heads-up that since it's a solo project just me, it'll probably take me about a month or so to get everything built and tested.
Keep an eye out for updates!
Does not work with Python embedded using ComfyUI. Gives an error (ModuleNotFoundError: No module named 'tkinter'). Used the batch file installer. No worky. Apparently it will only work with a full install of Python in Windows and ticking the option 'td/tk and IDLE' during Python installation. But I don't like running two versions of Python on Windows. Any idea how I can get this working with python_embedded?
my application uses CustomTkinter, which relies on a standard Python library called Tkinter to draw all its windows, buttons, and text boxes.
ComfyUI's Python: The "python_embedded" that comes with ComfyUI's easy installer is a special, stripped-down version of Python. To keep the file size small, it does not include Tkinter or other components needed for creating graphical user interface (GUI) applications.
using ComfyUI's Python, it immediately fails because the required Tkinter module is missing, leading to the ModuleNotFoundError: No module named 'tkinter'.
installing a full version of Python will not interfere with ComfyUI. The ComfyUI batch file is specifically coded to use its own internal python_embedded folder, so the two Python installations will operate independently without causing issues.
Python 3.10.6 Is recommended
there is no reliable way to fix this issue without installing a full version of Python.
Iβve found a workaround for the tkinter issue with the embedded Python setup.
The script just needs an update, and Iβll be including the fix in the next official release.
It will be available in version 5.10, which I plan to upload to GitHub within the next month.
That version will fully support the embedded Python environment without needing a full install.
