ComfyUI Repository Generator

This project provides a set of scripts to generate a wrapper repository for the Flymy.ai service.

Features

• Workflow Configuration: Generate resolution.json, a configuration file for repository creation.
• Repository Generation: Create a repository structure, including files for testing, pipeline setup, and inference.

Installation

To get started, install the package using one of the following methods:

If you use pip:

pip install -e .

If you use poetry:

poetry install

Usage

Step 0: Create the input directory

Create a playground/ directory at the same level as pyproject.toml (the root of the current repository) and place your workflow_api.json there.

To obtain your pipeline's workflow_api.json, enable developer mode in ComfyUI and then download it via the ComfyUI interface.

Step 1: Generate the Workflow Configuration

Run the following command to generate the playground/resolution.json file, which serves as the configuration for repository generation:

python flymy_comfyui_repo_gen/scripts/main.py process_workflow_api

You'll be prompted for input:

1. Enter workflow_api.json path [some default will be here]:
   If you completed Step 0 as described, just press ENTER.

2. Enter output json path [playground/resolution.json]:
   If you completed Step 0 as described, just press ENTER.

3. Do any fields from node "SOME_NODE_ID" need to be added? [y]:
   Press ENTER (or type y) to include the node configuration from workflow_api.json into the input schema.

4. Include 'SOME_NODE_ID.inputs.SOME_FIELD_NAME' field into the input schema? [y]:
   Press ENTER (or type y) to include the specified field from workflow_api.json into the input schema.

5. Verify results: [ "SOME_NODE_ID.inputs.SOME_FIELD_NAME" ] [Y/n]:
   If the result is correct, press ENTER. If not, type n and retry Step 1.

6. Does your ComfyUI pipeline contain any repositories not previously specified? [y/n]:
   Input y if you need to add custom ComfyUI nodes, otherwise n.

7. Install mode (git (g), ComfyUI-Manager (cm)) [git (g)]:
   Choose the installation mode for the node:

   • Type cm if the node is installed via ComfyUI-Manager.
   • Type g if the node is installed from a git repository.

git	ComfyUI-Manager
Auth token (invisible input, required for private repos) []: Enter the auth token for git access.	Enter the title of the node that corresponds to ComfyUI-Manager (checkout https://github.com/ltdrdata/ComfyUI-Manager/blob/main/custom-node-list.json): Enter the ComfyUI-Manager name of the custom node repository.
Repository url: Enter the HTTP/HTTPS URL (NOT SSH!) for the custom node repository. Check out our ComfyUI custom nodes guide.

Step 2: Generate the Repository

After creating the configuration file, generate the repository structure:

python flymy_comfyui_repo_gen/scripts/main.py generate_repository

You will be prompted for:

1. Enter config JSON path [playground/resolution.json]:
   Provide the path to the resolution.json generated in Step 1 (press ENTER if defaults are correct).

2. Enter output directory [playground/result]:
   Specify the output directory for the generated repository (press ENTER for default).

3. Enter repository name [flymy_comfyui_repo]:
   Specify the project name (press ENTER for default).

The generated repository will contain the following files:

• infer.py: A test script to simulate model warm-up and evaluate the pipeline's performance.

• model.py: Defines the base model, handles pipeline setup, and executes inference.

• Types.py: Contains the generated types used for inference operations.

Contribution

Feel free to open issues or submit pull requests to improve this project.

Happy generating!