Luma Text to Image - ComfyUI Native Node Documentation

class LumaImageGenerationNode(ComfyNodeABC):
    """
    Generates images synchronously based on prompt and aspect ratio.
    """

    RETURN_TYPES = (IO.IMAGE,)
    DESCRIPTION = cleandoc(__doc__ or "")  # Handle potential None value
    FUNCTION = "api_call"
    API_NODE = True
    CATEGORY = "api node/image/Luma"

    @classmethod
    def INPUT_TYPES(s):
        return {
            "required": {
                "prompt": (
                    IO.STRING,
                    {
                        "multiline": True,
                        "default": "",
                        "tooltip": "Prompt for the image generation",
                    },
                ),
                "model": ([model.value for model in LumaImageModel],),
                "aspect_ratio": (
                    [ratio.value for ratio in LumaAspectRatio],
                    {
                        "default": LumaAspectRatio.ratio_16_9,
                    },
                ),
                "seed": (
                    IO.INT,
                    {
                        "default": 0,
                        "min": 0,
                        "max": 0xFFFFFFFFFFFFFFFF,
                        "control_after_generate": True,
                        "tooltip": "Seed to determine if node should re-run; actual results are nondeterministic regardless of seed.",
                    },
                ),
                "style_image_weight": (
                    IO.FLOAT,
                    {
                        "default": 1.0,
                        "min": 0.0,
                        "max": 1.0,
                        "step": 0.01,
                        "tooltip": "Weight of style image. Ignored if no style_image provided.",
                    },
                ),
            },
            "optional": {
                "image_luma_ref": (
                    LumaIO.LUMA_REF,
                    {
                        "tooltip": "Luma Reference node connection to influence generation with input images; up to 4 images can be considered."
                    },
                ),
                "style_image": (
                    IO.IMAGE,
                    {"tooltip": "Style reference image; only 1 image will be used."},
                ),
                "character_image": (
                    IO.IMAGE,
                    {
                        "tooltip": "Character reference images; can be a batch of multiple, up to 4 images can be considered."
                    },
                ),
            },
            "hidden": {
                "auth_token": "AUTH_TOKEN_COMFY_ORG",
            },
        }

    def api_call(
        self,
        prompt: str,
        model: str,
        aspect_ratio: str,
        seed,
        style_image_weight: float,
        image_luma_ref: LumaReferenceChain = None,
        style_image: torch.Tensor = None,
        character_image: torch.Tensor = None,
        auth_token=None,
        **kwargs,
    ):
        # handle image_luma_ref
        api_image_ref = None
        if image_luma_ref is not None:
            api_image_ref = self._convert_luma_refs(
                image_luma_ref, max_refs=4, auth_token=auth_token
            )
        # handle style_luma_ref
        api_style_ref = None
        if style_image is not None:
            api_style_ref = self._convert_style_image(
                style_image, weight=style_image_weight, auth_token=auth_token
            )
        # handle character_ref images
        character_ref = None
        if character_image is not None:
            download_urls = upload_images_to_comfyapi(
                character_image, max_images=4, auth_token=auth_token
            )
            character_ref = LumaCharacterRef(
                identity0=LumaImageIdentity(images=download_urls)
            )

        operation = SynchronousOperation(
            endpoint=ApiEndpoint(
                path="/proxy/luma/generations/image",
                method=HttpMethod.POST,
                request_model=LumaImageGenerationRequest,
                response_model=LumaGeneration,
            ),
            request=LumaImageGenerationRequest(
                prompt=prompt,
                model=model,
                aspect_ratio=aspect_ratio,
                image_ref=api_image_ref,
                style_ref=api_style_ref,
                character_ref=character_ref,
            ),
            auth_token=auth_token,
        )
        response_api: LumaGeneration = operation.execute()

        operation = PollingOperation(
            poll_endpoint=ApiEndpoint(
                path=f"/proxy/luma/generations/{response_api.id}",
                method=HttpMethod.GET,
                request_model=EmptyRequest,
                response_model=LumaGeneration,
            ),
            completed_statuses=[LumaState.completed],
            failed_statuses=[LumaState.failed],
            status_extractor=lambda x: x.state,
            auth_token=auth_token,
        )
        response_poll = operation.execute()

        img_response = requests.get(response_poll.assets.image)
        img = process_image_response(img_response)
        return (img,)

    def _convert_luma_refs(
        self, luma_ref: LumaReferenceChain, max_refs: int, auth_token=None
    ):
        luma_urls = []
        ref_count = 0
        for ref in luma_ref.refs:
            download_urls = upload_images_to_comfyapi(
                ref.image, max_images=1, auth_token=auth_token
            )
            luma_urls.append(download_urls[0])
            ref_count += 1
            if ref_count >= max_refs:
                break
        return luma_ref.create_api_model(download_urls=luma_urls, max_refs=max_refs)

    def _convert_style_image(
        self, style_image: torch.Tensor, weight: float, auth_token=None
    ):
        chain = LumaReferenceChain(
            first_ref=LumaReference(image=style_image, weight=weight)
        )
        return self._convert_luma_refs(chain, max_refs=1, auth_token=auth_token)