ComfyUI Warper Nodes

ComfyUI Warper Nodes

A collection of general-purpose nodes for ComfyUI.

Current Nodes

• DWPose Scaler (Warper): Provides advanced scaling and adjustment options for DWPose keypoints. (Category: Warper Tools)

• Flow Visualizer (Warper): Calculates optical flow between image sequences using the RAFT model and provides visual representations (color-coded flow images) for both forward and backward flow. Ideal for understanding motion in your generations. (Category: Warper Tools/Flow)

  • Model Requirement: This node requires a RAFT model checkpoint.

• Smart Video Batcher (Warper): Splits image sequences into smaller batches with special padding for the final batch. (Category: Warper Tools/Looping)

• Get Batch By Index (Warper): Selects a single image batch from a list of batches by its index. (Category: Warper Tools/Looping)

• Smart Overlapping Batcher (Warper): Creates overlapping batches from an image sequence to help maintain temporal consistency. (Category: Warper Tools/Looping)

• Mouth Mask from Pose (Warper): Generates a circular mask around the mouth area based on facial keypoints. (Category: Warper Tools/DWMasking)

• Facial Part Mask from Pose (Warper): Creates masks for various facial parts (entire face, mouth, eyes) using different shapes (convex hull, ellipse, etc.) based on facial keypoints. (Category: Warper Tools/DWMasking)

Prompt Loading Nodes (Category: Warper Tools/Prompts)

• Load Prompts from Directory (Warper): Loads text prompts from .txt files in a specified directory with options for batch processing.

  • Inputs:
    • directory_path: Path to directory containing .txt files
    • skip_first_files: Number of files to skip from the beginning (useful with incremental integer nodes for sequential access)
    • load_cap: Maximum number of prompts to load at once
  • Outputs:
    • prompts: List of loaded prompt texts
    • total_loaded: Number of prompts successfully loaded
    • file_names: List of filenames that were loaded
  • Use Case: Perfect for batch processing multiple prompts or iterating through a prompt library with an incremental counter node.

• Load Single Prompt by Index (Warper): Loads a single prompt from a directory by its index position.

  • Inputs:
    • directory_path: Path to directory containing .txt files
    • file_index: Index of the file to load (0-based)
  • Outputs:
    • prompt: The loaded prompt text
    • file_name: Name of the loaded file
    • total_files: Total number of .txt files in the directory
    • is_valid: Boolean indicating if the load was successful
  • Use Case: Ideal for workflows that need to process prompts one at a time with precise index control.

• Combine Prompts (Warper): Combines multiple prompts with customizable separators and formatting.

  • Inputs:
    • prompt1-4: Up to 4 prompt inputs (all optional)
    • separator: Text to insert between prompts (default: ", ")
    • prefix: Text to add before the combined result
    • suffix: Text to add after the combined result
  • Outputs:
    • combined_prompt: The final combined prompt string
  • Use Case: Useful for building complex prompts from multiple sources or adding consistent formatting.

Text File Iterator Nodes (Category: Warper Tools/Text)

• Text File Iterator (Warper): Loads text files from a directory with different iteration modes.

  • Inputs:
    • directory_path: Path to directory containing .txt files
    • index: Index number (behavior depends on mode)
    • mode: Iteration mode selection
      • fixed: Always loads the file at the specified index
      • incremental: Cycles through files sequentially, remembering position between runs
      • randomized: Randomly selects a file using the index as a seed for reproducible results
  • Outputs:
    • text_content: The content of the loaded text file
  • Use Case: Perfect for iterating through prompt files with different access patterns - fixed for specific prompts, incremental for sequential workflows, or randomized for variety.

• Text File Iterator with Info (Warper): Enhanced version that provides additional file information.

  • Inputs: Same as Text File Iterator
  • Outputs:
    • text_content: The content of the loaded text file
    • file_name: Name of the loaded file
    • file_index: Index of the loaded file in the directory
    • total_files: Total number of .txt files in the directory
  • Use Case: Ideal when you need to track which file is being used or display progress information.

Installation

1. Navigate to your ComfyUI custom_nodes directory.

2. Clone this repository:

   git clone https://github.com/AIWarper/ComfyUI-WarperNodes
   

3. Install RAFT Model (Required for Flow Visualizer Node):

   • Download the RAFT model checkpoint: Tartan-C-T-TSKH-spring540x960-M.pth
     • Download Link: Google Drive Folder (Look for the specified .pth file within this folder.)
   • Create a raft folder inside your ComfyUI models directory if it doesn't already exist: ComfyUI/models/raft/
   • Place the downloaded Tartan-C-T-TSKH-spring540x960-M.pth file into this ComfyUI/models/raft/ directory.

4. Install any Python package dependencies (if listed in requirements.txt):

   cd ComfyUI-WarperNodes
   pip install -r requirements.txt
   

   (If you use ComfyUI's portable version, you might need to use its embedded Python, e.g., path/to/ComfyUI/python_embeded/python.exe -m pip install -r requirements.txt. If a requirements.txt file is not present, this step may not be needed.)

5. Restart ComfyUI.

Usage Notes

• Ensure the RAFT model is correctly placed as described above for the Flow Visualizer (Warper) node to function.
• The nodes will appear under the "Warper Tools" category in the ComfyUI "Add Node" menu.

TODO:

• Add example images/gifs for other nodes.
• More detailed usage instructions for each node.