Add support for VIDEO as a built-in type

[guill]

This PR adds support for VIDEO as first-class types. In order to avoid unnecessary costs, VIDEO outputs must implement the VideoInput ABC, but their implementation details can vary. Included are two implementations of this type which can be returned by other nodes:

• VideoFromFile - Created with either a path on disk (as a string) or a io.BytesIO containing the contents of a file in a supported format (like .mp4). This implementation won't actually load the video unless necessary. It will also avoid re-encoding when saving if possible.
• VideoFromComponents - Created from an image tensor and an optional audio tensor.

Currently, only h264 encoded videos in .mp4 containers are supported for saving, but the plan is to add additional encodings/containers in the near future (particularly .webm).

This PR also adds ImageInput and AudioInput types which can be used for static analysis and type-hinting. Nothing has changed about how these types themselves are represented.

[huchenlei]

Can you attach a workflow which uses all newly added nodes to help testing.

[comfyanonymous]

putting the node types in the comfy folder was a mistake, can you move this out of there and then I'll move the rest.

the comfy folder shouldn't contain anything related to the higher level nodes.

[guill]

video_test_workflow.json

Attached is a workflow that tests all nodes.

Here's an example node that loads a video from a URL to show usage with io.BytesIO:

class LoadVideoFromURL(ComfyNodeABC):
    @classmethod
    def INPUT_TYPES(cls):
        return {
            "required": {
                "url": ("STRING", {"default": "https://mintlify.s3.us-west-1.amazonaws.com/dripart/images/concepts/node/node_appearance.mp4", "tooltip": "The path to the video file."}),
            }
        }
    RETURN_TYPES = (IO.VIDEO,)
    FUNCTION = "load"

    CATEGORY = "image/video"
    DESCRIPTION = "Loads a video from a URL."

    def load(self, url):
        # Download the URL
        import requests
        import io
        response = requests.get(url)
        if response.status_code != 200:
            raise ValueError(f"Failed to download video from {url}. Status code: {response.status_code}")
        # Create a BytesIO object from the response content
        video_data = response.content
        video_io = io.BytesIO(video_data)

        return (VideoFromFile(video_io),)

[guill]

This was just fixing an incorrect type annotation.

[guill]

I'll also call out this comment on the refactoring commit:

I've created a new comfy_api top-level module. The intention is that
anything within this folder would be covered by semver-style versioning
that would allow custom nodes to rely on them not introducing breaking
changes.

I think it's important that custom nodes feel free to use the VideoInput type and add their own implementations and that we carefully consider any changes that might be backward-incompatible before we make them.