compat: improve support for comfyui

[mostlygeek]

ComfyUI does not have a good swap experience currently. The main reason is a websocket the frontend opens to the backend that is counted as an inflight request which currently prevents swapping. This is what I would like:

• add a new model.websocket.ignore configuration parameter
• add a new /comfyui endpoint for automatic configuration

The new model.websocket looks like this:

models:
  mycomfy: 
    # websocket: holds websocket options for the model 
    # optional  - default empty settings
    websocket: 
      # ignore: the websocket connection is ignored when considering a swap request
      # default - false, websockets are not ignored
      # when true, a websocket HTTP request will be ignored when consider a swap in or
      # swap out for the model.
      ignore: true   

When websocket.ignore = true a websocket connection will not be considered when swapping. This works for more than just comfyui such as for LLM inference that support websockets (see #754).

The new /comfyui endpoint is a wrapper over the above functionality with a few built in requirements for easy integration:

• automatically looks for a model named "ls_comfyui".
• forces model.websocket.ignore = true
• any future comfyui workarounds will be applied via this endpoint

Ref: #1000

[mostlygeek]

After some a configuration to hold various compatibility tweaks and tweaks is better. For now we can start with ignoring websockets but future models may need other workarounds for

models:
  mymodel:
    # workarounds: holds workarounds for better swapping compatibility with llama-swap
    #  - optional, default blank
    workarounds: 
      # ignoreWebSocket: do not consider connected websocket connections for any swap in/out activity
      # - optional, default false
      ignoreWebsockets: true

[mostlygeek]

Configuration example for the new /comfyui/ endpoint:

models:
  # comfyui_auto is a special model name that works with /comfyui. llama-swap will include 
  # compatibility workarounds so comfyui swaps out smoothly when the UI is idle but another 
  # model is requested
  comfyui_auto:
    checkEndpoint: /
    cmdStop: docker stop comfyui-auto
    cmd: >
      docker run --rm
      --name comfyui-auto --runtime=nvidia
      --gpus '"device=2,3"' -p ${PORT}:8188
      -v /path/to/comfyui/storage-cache/dot-cache:/root/.cache
      -v /path/to/comfyui/storage-cache/dot-config:/root/.config
      -v /path/to/comfyui/storage-nodes/dot-local:/root/.local
      -v /path/to/comfyui/storage-nodes/custom_nodes:/root/ComfyUI/custom_nodes
      -v /path/to/comfyui/storage-models/models:/root/ComfyUI/models
      -v /path/to/comfyui/storage-models/hf-hub:/root/.cache/huggingface/hub
      -v /path/to/comfyui/storage-models/torch-hub:/root/.cache/torch/hub
      -v /path/to/comfyui/storage-user/input:/root/ComfyUI/input
      -v /path/to/comfyui/storage-user/output:/root/ComfyUI/output
      -v /path/to/comfyui/storage-user/user-profile:/root/ComfyUI/user
      -v /path/to/comfyui/storage-user/user-scripts:/root/user-scripts
      -e CLI_ARGS=""
      yanwk/comfyui-boot:cu130-slim-v2

The /comfyui endpoint only works when a model is named comfyui_auto. To replicate the special functionality with a different model name and the /upstream/{my-other-comfyui-model}/ you can do this:

upstream:
   ignorePaths:
    # ComfyUI specific: ignore /ws and /api/jobs to prevent swaps
    # when the UI is open
    - ^\/ws$|^\/api\/jobs$

models:
  # comfyui_auto is a special model name that works with /comfyui. llama-swap will include 
  # compatibility workarounds so comfyui swaps out smoothly when the UI is idle but another 
  # model is requested
  my-other-comfyui-model:
    checkEndpoint: /
    cmdStop: docker stop comfyui-auto
    cmd: >
      docker run --rm
      --name comfyui-auto --runtime=nvidia
      --gpus '"device=2,3"' -p ${PORT}:8188
      -v /path/to/comfyui/storage-cache/dot-cache:/root/.cache
      -v /path/to/comfyui/storage-cache/dot-config:/root/.config
      -v /path/to/comfyui/storage-nodes/dot-local:/root/.local
      -v /path/to/comfyui/storage-nodes/custom_nodes:/root/ComfyUI/custom_nodes
      -v /path/to/comfyui/storage-models/models:/root/ComfyUI/models
      -v /path/to/comfyui/storage-models/hf-hub:/root/.cache/huggingface/hub
      -v /path/to/comfyui/storage-models/torch-hub:/root/.cache/torch/hub
      -v /path/to/comfyui/storage-user/input:/root/ComfyUI/input
      -v /path/to/comfyui/storage-user/output:/root/ComfyUI/output
      -v /path/to/comfyui/storage-user/user-profile:/root/ComfyUI/user
      -v /path/to/comfyui/storage-user/user-scripts:/root/user-scripts
      -e CLI_ARGS=""
      yanwk/comfyui-boot:cu130-slim-v2