88 lines
5.4 KiB
Markdown
88 lines
5.4 KiB
Markdown
# 🎭 ComfyUI Dynamic Dropdowns & Assistant
|
|
|
|
|
|
|
|
A self-contained, high-efficiency custom node system and Windows desktop companion app for ComfyUI. This tool eliminates canvas clutter by consolidating multiple text attribute dropdowns (e.g., hair lengths, colors, styles, outfits, lighting) into a single, cohesive comma-separated output block—serving it simultaneously as a raw string and an encoded conditioning prompt.
|
|
|
|
---
|
|
|
|
## ✨ Features
|
|
|
|
* **📦 100% Self-Contained Architecture:** No messy global asset paths. The custom node scans its own local `lists/` subfolder, ensuring absolute portability. If you copy the node folder to another machine, it works instantly.
|
|
* **🧠 Dual Prompt Injection Engine:** Outputs both a raw `STRING` (for prompt text stacking/debugging) and a pre-tokenized `CONDITIONING` block (to wire directly into KSampler slots, bypassing separate Text Encode nodes).
|
|
* **💻 Desktop Companion Utility (`Node Builder Assistant.exe`):** A zero-dependency, hardware-accelerated Windows UI tool that lets you visually build, update, and manage your custom node structures offline without writing a single line of Python.
|
|
* **🧼 Automatic Formatting & Cleaning:** The node backend strips out double quotes (`"`), trailing backslashes (`\`), empty rows, and handles string concatenation automatically, preventing messy trailing commas in your generation prompts.
|
|
|
|
---
|
|
|
|
## 📂 Repository Architecture
|
|
|
|
When properly deployed, your project folder maintains this modular layout:
|
|
|
|
```text
|
|
ComfyUI-Dynamic-Dropdowns/
|
|
├── Node Builder Assistant.exe # Standalone node generator app
|
|
├── __init__.py # The active ComfyUI custom node script
|
|
├── README.md # Project documentation
|
|
├── .gitignore # Dev exclusion rules
|
|
├── lists/ # Put your choice text assets here!
|
|
│ ├── hair_length.txt
|
|
│ ├── hair_color.txt
|
|
│ └── hairstyle.txt
|
|
└── generator/ # Companion App source folder (for updates)
|
|
├── app.py # PyWebView window engine wrapper
|
|
├── generator.html # Desktop interface panel
|
|
├── requirements.txt # Dev build dependencies
|
|
└── build.py # Isolated one-click .exe compiler
|
|
```
|
|
---
|
|
|
|
## ⚙️ Installation Guide
|
|
|
|
### Standard User Quick Start
|
|
1. Clone this repository directly into your ComfyUI nodes directory:
|
|
```bash
|
|
git clone [https://github.com/0velia/ComfyUI-Dynamic-Dropdowns.git](https://github.com/0velia/ComfyUI-Dynamic-Dropdowns.git)
|
|
```
|
|
|
|
Launch or restart ComfyUI. A 🎭 Dynamic Dropdowns (Demo Mode) node will immediately be available on your canvas to show you how the pipeline routes text choices.
|
|
|
|
🛠️ Customizing and Building Your Own Nodes
|
|
Once you see how the demo works, you can build your own fully customized menu layouts:
|
|
|
|
Open the repository folder and double-click Node Builder Assistant.exe.
|
|
|
|
Design your custom dropdown panels and match them to your own text files inside the lists/ folder.
|
|
|
|
Click Build Custom Node, download the new code, and save it as __init__.py to overwrite the demo placeholder!
|
|
|
|
---
|
|
|
|
## 🚀 How to Use
|
|
|
|
### 1. Generating or Updating Nodes via the Companion App
|
|
If you want to configure what dropdown columns display inside ComfyUI:
|
|
1. Double-click **`Node Builder Assistant.exe`** right in the root directory.
|
|
2. Click **+ Add Dropdown Menu** for each menu attribute slot you wish to manage.
|
|
3. Assign the text label name (e.g., `hair_color`) and match it exactly to the `.txt` file name located inside your `lists/` folder.
|
|
4. Click **Build Custom Node** to generate the node logic.
|
|
5. Click **Download Code**, save the file, and name it exactly **`__init__.py`**, replacing the existing script in the root folder.
|
|
|
|
### 2. Canvas Workflow Mapping (Daisy-Chaining)
|
|
Thanks to the optional input stacking architecture, you can link multiple dynamic dropdown nodes sequentially to keep your canvas beautifully organized:
|
|
|
|
1. **Add Node 1 (e.g., Character Attributes):** Leave `input_string` empty. Select your core details (e.g., hair, clothing).
|
|
2. **Add Node 2 (e.g., Environment & Framing):** Connect the `PROMPT_STRING` output from Node 1 directly into the `input_string` slot of Node 2.
|
|
3. **Connect Pipeline Engines:** Wire your checkpoint's `CLIP` output directly into Node 2's `clip` slot.
|
|
4. **Final Delivery:** Route Node 2's `CONDITIONING` dot directly into your KSampler. The backend seamlessly combines the outputs (`character traits, framing choices, lighting styles`) into a single comma-separated string without duplicate punctuation artifacts, even if inputs are left unlinked.
|
|
|
|
---
|
|
|
|
## 🛠️ Developer / Compilation Setup
|
|
If you want to modify the companion app layout styles or tweak the core source generation engine:
|
|
1. Make your code alterations inside `generator/generator.html` or `generator/app.py`.
|
|
2. Open your command terminal inside the `/generator` folder and execute the one-click compilation automator script:
|
|
python build.py
|
|
|
|
This script dynamically sets up an isolated Python virtual environment, installs the compression assets, strips module bloat, compiles the new optimized binary, moves it directly to the root path as `Node Builder Assistant.exe`, and purges temporary cache files automatically.
|