Merge branch 'comfyanonymous:master' into master

comfy/k_diffusion/sa_solver.py

@@ -0,0 +1,121 @@
+# SA-Solver: Stochastic Adams Solver (NeurIPS 2023, arXiv:2309.05019)
+# Conference: https://proceedings.neurips.cc/paper_files/paper/2023/file/f4a6806490d31216a3ba667eb240c897-Paper-Conference.pdf
+# Codebase ref: https://github.com/scxue/SA-Solver
+
+import math
+from typing import Union, Callable
+import torch
+
+
+def compute_exponential_coeffs(s: torch.Tensor, t: torch.Tensor, solver_order: int, tau_t: float) -> torch.Tensor:
+    """Compute (1 + tau^2) * integral of exp((1 + tau^2) * x) * x^p dx from s to t with exp((1 + tau^2) * t) factored out, using integration by parts.
+
+    Integral of exp((1 + tau^2) * x) * x^p dx
+        = product_terms[p] - (p / (1 + tau^2)) * integral of exp((1 + tau^2) * x) * x^(p-1) dx,
+    with base case p=0 where integral equals product_terms[0].
+
+    where
+        product_terms[p] = x^p * exp((1 + tau^2) * x) / (1 + tau^2).
+
+    Construct a recursive coefficient matrix following the above recursive relation to compute all integral terms up to p = (solver_order - 1).
+    Return coefficients used by the SA-Solver in data prediction mode.
+
+    Args:
+        s: Start time s.
+        t: End time t.
+        solver_order: Current order of the solver.
+        tau_t: Stochastic strength parameter in the SDE.
+
+    Returns:
+        Exponential coefficients used in data prediction, with exp((1 + tau^2) * t) factored out, ordered from p=0 to p=solver_order−1, shape (solver_order,).
+    """
+    tau_mul = 1 + tau_t ** 2
+    h = t - s
+    p = torch.arange(solver_order, dtype=s.dtype, device=s.device)
+
+    # product_terms after factoring out exp((1 + tau^2) * t)
+    # Includes (1 + tau^2) factor from outside the integral
+    product_terms_factored = (t ** p - s ** p * (-tau_mul * h).exp())
+
+    # Lower triangular recursive coefficient matrix
+    # Accumulates recursive coefficients based on p / (1 + tau^2)
+    recursive_depth_mat = p.unsqueeze(1) - p.unsqueeze(0)
+    log_factorial = (p + 1).lgamma()
+    recursive_coeff_mat = log_factorial.unsqueeze(1) - log_factorial.unsqueeze(0)
+    if tau_t > 0:
+        recursive_coeff_mat = recursive_coeff_mat - (recursive_depth_mat * math.log(tau_mul))
+    signs = torch.where(recursive_depth_mat % 2 == 0, 1.0, -1.0)
+    recursive_coeff_mat = (recursive_coeff_mat.exp() * signs).tril()
+
+    return recursive_coeff_mat @ product_terms_factored
+
+
+def compute_simple_stochastic_adams_b_coeffs(sigma_next: torch.Tensor, curr_lambdas: torch.Tensor, lambda_s: torch.Tensor, lambda_t: torch.Tensor, tau_t: float, is_corrector_step: bool = False) -> torch.Tensor:
+    """Compute simple order-2 b coefficients from SA-Solver paper (Appendix D. Implementation Details)."""
+    tau_mul = 1 + tau_t ** 2
+    h = lambda_t - lambda_s
+    alpha_t = sigma_next * lambda_t.exp()
+    if is_corrector_step:
+        # Simplified 1-step (order-2) corrector
+        b_1 = alpha_t * (0.5 * tau_mul * h)
+        b_2 = alpha_t * (-h * tau_mul).expm1().neg() - b_1
+    else:
+        # Simplified 2-step predictor
+        b_2 = alpha_t * (0.5 * tau_mul * h ** 2) / (curr_lambdas[-2] - lambda_s)
+        b_1 = alpha_t * (-h * tau_mul).expm1().neg() - b_2
+    return torch.stack([b_2, b_1])
+
+
+def compute_stochastic_adams_b_coeffs(sigma_next: torch.Tensor, curr_lambdas: torch.Tensor, lambda_s: torch.Tensor, lambda_t: torch.Tensor, tau_t: float, simple_order_2: bool = False, is_corrector_step: bool = False) -> torch.Tensor:
+    """Compute b_i coefficients for the SA-Solver (see eqs. 15 and 18).
+
+    The solver order corresponds to the number of input lambdas (half-logSNR points).
+
+    Args:
+        sigma_next: Sigma at end time t.
+        curr_lambdas: Lambda time points used to construct the Lagrange basis, shape (N,).
+        lambda_s: Lambda at start time s.
+        lambda_t: Lambda at end time t.
+        tau_t: Stochastic strength parameter in the SDE.
+        simple_order_2: Whether to enable the simple order-2 scheme.
+        is_corrector_step: Flag for corrector step in simple order-2 mode.
+
+    Returns:
+        b_i coefficients for the SA-Solver, shape (N,), where N is the solver order.
+    """
+    num_timesteps = curr_lambdas.shape[0]
+
+    if simple_order_2 and num_timesteps == 2:
+        return compute_simple_stochastic_adams_b_coeffs(sigma_next, curr_lambdas, lambda_s, lambda_t, tau_t, is_corrector_step)
+
+    # Compute coefficients by solving a linear system from Lagrange basis interpolation
+    exp_integral_coeffs = compute_exponential_coeffs(lambda_s, lambda_t, num_timesteps, tau_t)
+    vandermonde_matrix_T = torch.vander(curr_lambdas, num_timesteps, increasing=True).T
+    lagrange_integrals = torch.linalg.solve(vandermonde_matrix_T, exp_integral_coeffs)
+
+    # (sigma_t * exp(-tau^2 * lambda_t)) * exp((1 + tau^2) * lambda_t)
+    # = sigma_t * exp(lambda_t) = alpha_t
+    # exp((1 + tau^2) * lambda_t) is extracted from the integral
+    alpha_t = sigma_next * lambda_t.exp()
+    return alpha_t * lagrange_integrals
+
+
+def get_tau_interval_func(start_sigma: float, end_sigma: float, eta: float = 1.0) -> Callable[[Union[torch.Tensor, float]], float]:
+    """Return a function that controls the stochasticity of SA-Solver.
+
+    When eta = 0, SA-Solver runs as ODE. The official approach uses
+    time t to determine the SDE interval, while here we use sigma instead.
+
+    See:
+        https://github.com/scxue/SA-Solver/blob/main/README.md
+    """
+
+    def tau_func(sigma: Union[torch.Tensor, float]) -> float:
+        if eta <= 0:
+            return 0.0  # ODE
+
+        if isinstance(sigma, torch.Tensor):
+            sigma = sigma.item()
+        return eta if start_sigma >= sigma >= end_sigma else 0.0
+
+    return tau_func

comfy/k_diffusion/sampling.py

@@ -9,6 +9,7 @@ from tqdm.auto import trange, tqdm
 
 from . import utils
 from . import deis
+from . import sa_solver
 import comfy.model_patcher
 import comfy.model_sampling
 
@@ -1648,3 +1649,113 @@ def sample_seeds_3(model, x, sigmas, extra_args=None, callback=None, disable=Non
             if inject_noise:
                 x = x + sigmas[i + 1] * (noise_coeff_3 * noise_1 + noise_coeff_2 * noise_2 + noise_coeff_1 * noise_3) * s_noise
     return x
+
+
+@torch.no_grad()
+def sample_sa_solver(model, x, sigmas, extra_args=None, callback=None, disable=False, tau_func=None, s_noise=1.0, noise_sampler=None, predictor_order=3, corrector_order=4, use_pece=False, simple_order_2=False):
+    """Stochastic Adams Solver with predictor-corrector method (NeurIPS 2023)."""
+    if len(sigmas) <= 1:
+        return x
+    extra_args = {} if extra_args is None else extra_args
+    seed = extra_args.get("seed", None)
+    noise_sampler = default_noise_sampler(x, seed=seed) if noise_sampler is None else noise_sampler
+    s_in = x.new_ones([x.shape[0]])
+
+    model_sampling = model.inner_model.model_patcher.get_model_object("model_sampling")
+    sigmas = offset_first_sigma_for_snr(sigmas, model_sampling)
+    lambdas = sigma_to_half_log_snr(sigmas, model_sampling=model_sampling)
+
+    if tau_func is None:
+        # Use default interval for stochastic sampling
+        start_sigma = model_sampling.percent_to_sigma(0.2)
+        end_sigma = model_sampling.percent_to_sigma(0.8)
+        tau_func = sa_solver.get_tau_interval_func(start_sigma, end_sigma, eta=1.0)
+
+    max_used_order = max(predictor_order, corrector_order)
+    x_pred = x  # x: current state, x_pred: predicted next state
+
+    h = 0.0
+    tau_t = 0.0
+    noise = 0.0
+    pred_list = []
+
+    # Lower order near the end to improve stability
+    lower_order_to_end = sigmas[-1].item() == 0
+
+    for i in trange(len(sigmas) - 1, disable=disable):
+        # Evaluation
+        denoised = model(x_pred, sigmas[i] * s_in, **extra_args)
+        if callback is not None:
+            callback({"x": x_pred, "i": i, "sigma": sigmas[i], "sigma_hat": sigmas[i], "denoised": denoised})
+        pred_list.append(denoised)
+        pred_list = pred_list[-max_used_order:]
+
+        predictor_order_used = min(predictor_order, len(pred_list))
+        if i == 0 or (sigmas[i + 1] == 0 and not use_pece):
+            corrector_order_used = 0
+        else:
+            corrector_order_used = min(corrector_order, len(pred_list))
+
+        if lower_order_to_end:
+            predictor_order_used = min(predictor_order_used, len(sigmas) - 2 - i)
+            corrector_order_used = min(corrector_order_used, len(sigmas) - 1 - i)
+
+        # Corrector
+        if corrector_order_used == 0:
+            # Update by the predicted state
+            x = x_pred
+        else:
+            curr_lambdas = lambdas[i - corrector_order_used + 1:i + 1]
+            b_coeffs = sa_solver.compute_stochastic_adams_b_coeffs(
+                sigmas[i],
+                curr_lambdas,
+                lambdas[i - 1],
+                lambdas[i],
+                tau_t,
+                simple_order_2,
+                is_corrector_step=True,
+            )
+            pred_mat = torch.stack(pred_list[-corrector_order_used:], dim=1)    # (B, K, ...)
+            corr_res = torch.tensordot(pred_mat, b_coeffs, dims=([1], [0]))  # (B, ...)
+            x = sigmas[i] / sigmas[i - 1] * (-(tau_t ** 2) * h).exp() * x + corr_res
+
+            if tau_t > 0 and s_noise > 0:
+                # The noise from the previous predictor step
+                x = x + noise
+
+            if use_pece:
+                # Evaluate the corrected state
+                denoised = model(x, sigmas[i] * s_in, **extra_args)
+                pred_list[-1] = denoised
+
+        # Predictor
+        if sigmas[i + 1] == 0:
+            # Denoising step
+            x = denoised
+        else:
+            tau_t = tau_func(sigmas[i + 1])
+            curr_lambdas = lambdas[i - predictor_order_used + 1:i + 1]
+            b_coeffs = sa_solver.compute_stochastic_adams_b_coeffs(
+                sigmas[i + 1],
+                curr_lambdas,
+                lambdas[i],
+                lambdas[i + 1],
+                tau_t,
+                simple_order_2,
+                is_corrector_step=False,
+            )
+            pred_mat = torch.stack(pred_list[-predictor_order_used:], dim=1)    # (B, K, ...)
+            pred_res = torch.tensordot(pred_mat, b_coeffs, dims=([1], [0]))  # (B, ...)
+            h = lambdas[i + 1] - lambdas[i]
+            x_pred = sigmas[i + 1] / sigmas[i] * (-(tau_t ** 2) * h).exp() * x + pred_res
+
+            if tau_t > 0 and s_noise > 0:
+                noise = noise_sampler(sigmas[i], sigmas[i + 1]) * sigmas[i + 1] * (-2 * tau_t ** 2 * h).expm1().neg().sqrt() * s_noise
+                x_pred = x_pred + noise
+    return x
+
+
+@torch.no_grad()
+def sample_sa_solver_pece(model, x, sigmas, extra_args=None, callback=None, disable=False, tau_func=None, s_noise=1.0, noise_sampler=None, predictor_order=3, corrector_order=4, simple_order_2=False):
+    """Stochastic Adams Solver with PECE (Predict–Evaluate–Correct–Evaluate) mode (NeurIPS 2023)."""
+    return sample_sa_solver(model, x, sigmas, extra_args=extra_args, callback=callback, disable=disable, tau_func=tau_func, s_noise=s_noise, noise_sampler=noise_sampler, predictor_order=predictor_order, corrector_order=corrector_order, use_pece=True, simple_order_2=simple_order_2)

comfy/ldm/chroma/model.py

@@ -254,13 +254,12 @@ class Chroma(nn.Module):
 
     def forward(self, x, timestep, context, guidance, control=None, transformer_options={}, **kwargs):
         bs, c, h, w = x.shape
-        patch_size = 2
-        x = comfy.ldm.common_dit.pad_to_patch_size(x, (patch_size, patch_size))
+        x = comfy.ldm.common_dit.pad_to_patch_size(x, (self.patch_size, self.patch_size))
 
-        img = rearrange(x, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=patch_size, pw=patch_size)
+        img = rearrange(x, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=self.patch_size, pw=self.patch_size)
 
-        h_len = ((h + (patch_size // 2)) // patch_size)
-        w_len = ((w + (patch_size // 2)) // patch_size)
+        h_len = ((h + (self.patch_size // 2)) // self.patch_size)
+        w_len = ((w + (self.patch_size // 2)) // self.patch_size)
         img_ids = torch.zeros((h_len, w_len, 3), device=x.device, dtype=x.dtype)
         img_ids[:, :, 1] = img_ids[:, :, 1] + torch.linspace(0, h_len - 1, steps=h_len, device=x.device, dtype=x.dtype).unsqueeze(1)
         img_ids[:, :, 2] = img_ids[:, :, 2] + torch.linspace(0, w_len - 1, steps=w_len, device=x.device, dtype=x.dtype).unsqueeze(0)
@@ -268,4 +267,4 @@ class Chroma(nn.Module):
 
         txt_ids = torch.zeros((bs, context.shape[1], 3), device=x.device, dtype=x.dtype)
         out = self.forward_orig(img, img_ids, context, txt_ids, timestep, guidance, control, transformer_options, attn_mask=kwargs.get("attention_mask", None))
-        return rearrange(out, "b (h w) (c ph pw) -> b c (h ph) (w pw)", h=h_len, w=w_len, ph=2, pw=2)[:,:,:h,:w]
+        return rearrange(out, "b (h w) (c ph pw) -> b c (h ph) (w pw)", h=h_len, w=w_len, ph=self.patch_size, pw=self.patch_size)[:,:,:h,:w]

comfy/samplers.py

@@ -720,7 +720,7 @@ KSAMPLER_NAMES = ["euler", "euler_cfg_pp", "euler_ancestral", "euler_ancestral_c
                   "lms", "dpm_fast", "dpm_adaptive", "dpmpp_2s_ancestral", "dpmpp_2s_ancestral_cfg_pp", "dpmpp_sde", "dpmpp_sde_gpu",
                   "dpmpp_2m", "dpmpp_2m_cfg_pp", "dpmpp_2m_sde", "dpmpp_2m_sde_gpu", "dpmpp_3m_sde", "dpmpp_3m_sde_gpu", "ddpm", "lcm",
                   "ipndm", "ipndm_v", "deis", "res_multistep", "res_multistep_cfg_pp", "res_multistep_ancestral", "res_multistep_ancestral_cfg_pp",
-                  "gradient_estimation", "gradient_estimation_cfg_pp", "er_sde", "seeds_2", "seeds_3"]
+                  "gradient_estimation", "gradient_estimation_cfg_pp", "er_sde", "seeds_2", "seeds_3", "sa_solver", "sa_solver_pece"]
 
 class KSAMPLER(Sampler):
     def __init__(self, sampler_function, extra_options={}, inpaint_options={}):

comfy_api/input/video_types.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 from abc import ABC, abstractmethod
-from typing import Optional
+from typing import Optional, Union
+import io
 from comfy_api.util import VideoContainer, VideoCodec, VideoComponents
 
 class VideoInput(ABC):
@@ -31,6 +32,22 @@ class VideoInput(ABC):
         """
         pass
 
+    def get_stream_source(self) -> Union[str, io.BytesIO]:
+        """
+        Get a streamable source for the video. This allows processing without
+        loading the entire video into memory.
+
+        Returns:
+            Either a file path (str) or a BytesIO object that can be opened with av.
+
+        Default implementation creates a BytesIO buffer, but subclasses should
+        override this for better performance when possible.
+        """
+        buffer = io.BytesIO()
+        self.save_to(buffer)
+        buffer.seek(0)
+        return buffer
+
     # Provide a default implementation, but subclasses can provide optimized versions
     # if possible.
     def get_dimensions(self) -> tuple[int, int]:

comfy_api/input_impl/video_types.py

@@ -64,6 +64,15 @@ class VideoFromFile(VideoInput):
         """
         self.__file = file
 
+    def get_stream_source(self) -> str | io.BytesIO:
+        """
+        Return the underlying file source for efficient streaming.
+        This avoids unnecessary memory copies when the source is already a file path.
+        """
+        if isinstance(self.__file, io.BytesIO):
+            self.__file.seek(0)
+        return self.__file
+
     def get_dimensions(self) -> tuple[int, int]:
         """
         Returns the dimensions of the video input.

comfy_api_nodes/apis/__init__.py

@@ -1,6 +1,6 @@
 # generated by datamodel-codegen:
 #   filename:  filtered-openapi.yaml
-#   timestamp: 2025-05-19T21:38:55+00:00
+#   timestamp: 2025-07-06T09:47:31+00:00
 
 from __future__ import annotations
 
@@ -1355,6 +1355,158 @@ class ModelResponseProperties(BaseModel):
     )
 
 
+class Keyframes(BaseModel):
+    image_url: Optional[str] = None
+
+
+class MoonvalleyPromptResponse(BaseModel):
+    error: Optional[Dict[str, Any]] = None
+    frame_conditioning: Optional[Dict[str, Any]] = None
+    id: Optional[str] = None
+    inference_params: Optional[Dict[str, Any]] = None
+    meta: Optional[Dict[str, Any]] = None
+    model_params: Optional[Dict[str, Any]] = None
+    output_url: Optional[str] = None
+    prompt_text: Optional[str] = None
+    status: Optional[str] = None
+
+
+class MoonvalleyTextToVideoInferenceParams(BaseModel):
+    add_quality_guidance: Optional[bool] = Field(
+        True, description='Whether to add quality guidance'
+    )
+    caching_coefficient: Optional[float] = Field(
+        0.3, description='Caching coefficient for optimization'
+    )
+    caching_cooldown: Optional[int] = Field(
+        3, description='Number of caching cooldown steps'
+    )
+    caching_warmup: Optional[int] = Field(
+        3, description='Number of caching warmup steps'
+    )
+    clip_value: Optional[float] = Field(
+        3, description='CLIP value for generation control'
+    )
+    conditioning_frame_index: Optional[int] = Field(
+        0, description='Index of the conditioning frame'
+    )
+    cooldown_steps: Optional[int] = Field(
+        None, description='Number of cooldown steps (calculated based on num_frames)'
+    )
+    fps: Optional[int] = Field(
+        24, description='Frames per second of the generated video'
+    )
+    guidance_scale: Optional[float] = Field(
+        12.5, description='Guidance scale for generation control'
+    )
+    height: Optional[int] = Field(
+        1080, description='Height of the generated video in pixels'
+    )
+    negative_prompt: Optional[str] = Field(None, description='Negative prompt text')
+    num_frames: Optional[int] = Field(64, description='Number of frames to generate')
+    seed: Optional[int] = Field(
+        None, description='Random seed for generation (default: random)'
+    )
+    shift_value: Optional[float] = Field(
+        3, description='Shift value for generation control'
+    )
+    steps: Optional[int] = Field(80, description='Number of denoising steps')
+    use_guidance_schedule: Optional[bool] = Field(
+        True, description='Whether to use guidance scheduling'
+    )
+    use_negative_prompts: Optional[bool] = Field(
+        False, description='Whether to use negative prompts'
+    )
+    use_timestep_transform: Optional[bool] = Field(
+        True, description='Whether to use timestep transformation'
+    )
+    warmup_steps: Optional[int] = Field(
+        None, description='Number of warmup steps (calculated based on num_frames)'
+    )
+    width: Optional[int] = Field(
+        1920, description='Width of the generated video in pixels'
+    )
+
+
+class MoonvalleyTextToVideoRequest(BaseModel):
+    image_url: Optional[str] = None
+    inference_params: Optional[MoonvalleyTextToVideoInferenceParams] = None
+    prompt_text: Optional[str] = None
+    webhook_url: Optional[str] = None
+
+
+class MoonvalleyUploadFileRequest(BaseModel):
+    file: Optional[StrictBytes] = None
+
+
+class MoonvalleyUploadFileResponse(BaseModel):
+    access_url: Optional[str] = None
+
+
+class MoonvalleyVideoToVideoInferenceParams(BaseModel):
+    add_quality_guidance: Optional[bool] = Field(
+        True, description='Whether to add quality guidance'
+    )
+    caching_coefficient: Optional[float] = Field(
+        0.3, description='Caching coefficient for optimization'
+    )
+    caching_cooldown: Optional[int] = Field(
+        3, description='Number of caching cooldown steps'
+    )
+    caching_warmup: Optional[int] = Field(
+        3, description='Number of caching warmup steps'
+    )
+    clip_value: Optional[float] = Field(
+        3, description='CLIP value for generation control'
+    )
+    conditioning_frame_index: Optional[int] = Field(
+        0, description='Index of the conditioning frame'
+    )
+    cooldown_steps: Optional[int] = Field(
+        None, description='Number of cooldown steps (calculated based on num_frames)'
+    )
+    guidance_scale: Optional[float] = Field(
+        12.5, description='Guidance scale for generation control'
+    )
+    negative_prompt: Optional[str] = Field(None, description='Negative prompt text')
+    seed: Optional[int] = Field(
+        None, description='Random seed for generation (default: random)'
+    )
+    shift_value: Optional[float] = Field(
+        3, description='Shift value for generation control'
+    )
+    steps: Optional[int] = Field(80, description='Number of denoising steps')
+    use_guidance_schedule: Optional[bool] = Field(
+        True, description='Whether to use guidance scheduling'
+    )
+    use_negative_prompts: Optional[bool] = Field(
+        False, description='Whether to use negative prompts'
+    )
+    use_timestep_transform: Optional[bool] = Field(
+        True, description='Whether to use timestep transformation'
+    )
+    warmup_steps: Optional[int] = Field(
+        None, description='Number of warmup steps (calculated based on num_frames)'
+    )
+
+
+class ControlType(str, Enum):
+    motion_control = 'motion_control'
+    pose_control = 'pose_control'
+
+
+class MoonvalleyVideoToVideoRequest(BaseModel):
+    control_type: ControlType = Field(
+        ..., description='Supported types for video control'
+    )
+    inference_params: Optional[MoonvalleyVideoToVideoInferenceParams] = None
+    prompt_text: str = Field(..., description='Describes the video to generate')
+    video_url: str = Field(..., description='Url to control video')
+    webhook_url: Optional[str] = Field(
+        None, description='Optional webhook URL for notifications'
+    )
+
+
 class Moderation(str, Enum):
     low = 'low'
     auto = 'auto'
@@ -3107,6 +3259,23 @@ class LumaUpscaleVideoGenerationRequest(BaseModel):
     resolution: Optional[LumaVideoModelOutputResolution] = None
 
 
+class MoonvalleyImageToVideoRequest(MoonvalleyTextToVideoRequest):
+    keyframes: Optional[Dict[str, Keyframes]] = None
+
+
+class MoonvalleyResizeVideoRequest(MoonvalleyVideoToVideoRequest):
+    frame_position: Optional[List[int]] = Field(None, max_length=2, min_length=2)
+    frame_resolution: Optional[List[int]] = Field(None, max_length=2, min_length=2)
+    scale: Optional[List[int]] = Field(None, max_length=2, min_length=2)
+
+
+class MoonvalleyTextToImageRequest(BaseModel):
+    image_url: Optional[str] = None
+    inference_params: Optional[MoonvalleyTextToVideoInferenceParams] = None
+    prompt_text: Optional[str] = None
+    webhook_url: Optional[str] = None
+
+
 class OutputContent(RootModel[Union[OutputTextContent, OutputAudioContent]]):
     root: Union[OutputTextContent, OutputAudioContent]
 

comfy_api_nodes/nodes_moonvalley.py

@@ -0,0 +1,639 @@
+import logging
+from typing import Any, Callable, Optional, TypeVar
+import random
+import torch
+from comfy_api_nodes.util.validation_utils import get_image_dimensions, validate_image_dimensions, validate_video_dimensions
+
+
+from comfy_api_nodes.apis import (
+    MoonvalleyTextToVideoRequest,
+    MoonvalleyTextToVideoInferenceParams,
+    MoonvalleyVideoToVideoInferenceParams,
+    MoonvalleyVideoToVideoRequest,
+    MoonvalleyPromptResponse
+)
+from comfy_api_nodes.apis.client import (
+    ApiEndpoint,
+    HttpMethod,
+    SynchronousOperation,
+    PollingOperation,
+    EmptyRequest,
+)
+from comfy_api_nodes.apinode_utils import (
+    download_url_to_video_output,
+    upload_images_to_comfyapi,
+    upload_video_to_comfyapi,
+)
+from comfy_api_nodes.mapper_utils import model_field_to_node_input
+
+from comfy_api.input.video_types import VideoInput
+from comfy.comfy_types.node_typing import IO
+from comfy_api.input_impl import VideoFromFile
+import av
+import io
+
+API_UPLOADS_ENDPOINT = "/proxy/moonvalley/uploads"
+API_PROMPTS_ENDPOINT = "/proxy/moonvalley/prompts"
+API_VIDEO2VIDEO_ENDPOINT = "/proxy/moonvalley/prompts/video-to-video"
+API_TXT2VIDEO_ENDPOINT = "/proxy/moonvalley/prompts/text-to-video"
+API_IMG2VIDEO_ENDPOINT = "/proxy/moonvalley/prompts/image-to-video"
+
+MIN_WIDTH = 300
+MIN_HEIGHT = 300
+
+MAX_WIDTH = 10000
+MAX_HEIGHT = 10000
+
+MIN_VID_WIDTH = 300
+MIN_VID_HEIGHT = 300
+
+MAX_VID_WIDTH = 10000
+MAX_VID_HEIGHT = 10000
+
+MAX_VIDEO_SIZE = 1024 * 1024 * 1024  # 1 GB max for in-memory video processing
+
+MOONVALLEY_MAREY_MAX_PROMPT_LENGTH = 5000
+R = TypeVar("R")
+class MoonvalleyApiError(Exception):
+    """Base exception for Moonvalley API errors."""
+    pass
+
+def is_valid_task_creation_response(response: MoonvalleyPromptResponse) -> bool:
+    """Verifies that the initial response contains a task ID."""
+    return bool(response.id)
+
+def validate_task_creation_response(response) -> None:
+    if not is_valid_task_creation_response(response):
+        error_msg = f"Moonvalley Marey API: Initial request failed. Code: {response.code}, Message: {response.message}, Data: {response}"
+        logging.error(error_msg)
+        raise MoonvalleyApiError(error_msg)
+
+def get_video_from_response(response):
+    video = response.output_url
+    logging.info(
+        "Moonvalley Marey API: Task %s succeeded. Video URL: %s", response.id, video
+    )
+    return video
+
+
+def get_video_url_from_response(response) -> Optional[str]:
+    """Returns the first video url from the Moonvalley video generation task result.
+    Will not raise an error if the response is not valid.
+    """
+    if response:
+        return str(get_video_from_response(response))
+    else:
+        return None
+
+
+def poll_until_finished(
+    auth_kwargs: dict[str, str],
+    api_endpoint: ApiEndpoint[Any, R],
+    result_url_extractor: Optional[Callable[[R], str]] = None,
+    node_id: Optional[str] = None,
+) -> R:
+    """Polls the Moonvalley API endpoint until the task reaches a terminal state, then returns the response."""
+    return PollingOperation(
+        poll_endpoint=api_endpoint,
+        completed_statuses=[
+            "completed",
+        ],
+        max_poll_attempts=240,  # 64 minutes with 16s interval
+        poll_interval=16.0,
+        failed_statuses=["error"],
+        status_extractor=lambda response: (
+            response.status
+            if response and response.status
+            else None
+        ),
+        auth_kwargs=auth_kwargs,
+        result_url_extractor=result_url_extractor,
+        node_id=node_id,
+    ).execute()
+
+def validate_prompts(prompt:str, negative_prompt: str, max_length=MOONVALLEY_MAREY_MAX_PROMPT_LENGTH):
+    """Verifies that the prompt isn't empty and that neither prompt is too long."""
+    if not prompt:
+        raise ValueError("Positive prompt is empty")
+    if len(prompt) > max_length:
+        raise ValueError(f"Positive prompt is too long: {len(prompt)} characters")
+    if negative_prompt and len(negative_prompt) > max_length:
+        raise ValueError(
+            f"Negative prompt is too long: {len(negative_prompt)} characters"
+        )
+    return True
+
+def validate_input_media(width, height, with_frame_conditioning, num_frames_in=None):
+        # inference validation
+        # T = num_frames
+        # in all cases, the following must be true: T divisible by 16 and H,W by 8. in addition...
+        # with image conditioning: H*W must be divisible by 8192
+        # without image conditioning: T divisible by 32
+    if num_frames_in and not num_frames_in % 16 == 0 :
+        return False, (
+            "The input video total frame count must be divisible by 16!"
+        )
+
+    if height % 8 != 0 or width % 8 != 0:
+        return False, (
+            f"Height ({height}) and width ({width}) must be " "divisible by 8"
+        )
+
+    if with_frame_conditioning:
+        if (height * width) % 8192 != 0:
+            return False, (
+                f"Height * width ({height * width}) must be "
+                "divisible by 8192 for frame conditioning"
+            )
+    else:
+        if num_frames_in and not num_frames_in % 32 == 0 :
+            return False, (
+                "The input video total frame count must be divisible by 32!"
+            )
+
+
+def validate_input_image(image: torch.Tensor, with_frame_conditioning: bool=False) -> None:
+    """
+    Validates the input image adheres to the expectations of the API:
+    - The image resolution should not be less than 300*300px
+    - The aspect ratio of the image should be between 1:2.5 ~ 2.5:1
+
+    """
+    height, width = get_image_dimensions(image)
+    validate_input_media(width, height, with_frame_conditioning )
+    validate_image_dimensions(image, min_width=300, min_height=300, max_height=MAX_HEIGHT, max_width=MAX_WIDTH)
+
+def validate_input_video(video: VideoInput, num_frames_out: int, with_frame_conditioning: bool=False):
+    try:
+        width, height = video.get_dimensions()
+    except Exception as e:
+        logging.error("Error getting dimensions of video: %s", e)
+        raise ValueError(f"Cannot get video dimensions: {e}") from e
+
+    validate_input_media(width, height, with_frame_conditioning)
+    validate_video_dimensions(video, min_width=MIN_VID_WIDTH, min_height=MIN_VID_HEIGHT, max_width=MAX_VID_WIDTH, max_height=MAX_VID_HEIGHT)
+
+    trimmed_video = validate_input_video_length(video, num_frames_out)
+    return trimmed_video
+
+
+def validate_input_video_length(video: VideoInput, num_frames: int):
+
+    if video.get_duration() > 60:
+        raise MoonvalleyApiError("Input Video lenth should be less than 1min. Please trim.")
+
+    if num_frames == 128:
+       if video.get_duration() < 5:
+           raise MoonvalleyApiError("Input Video length is less than 5s. Please use a video longer than or equal to 5s.")
+       if video.get_duration() > 5:
+        #    trim video to 5s
+        video = trim_video(video, 5)
+    if num_frames == 256:
+        if video.get_duration() < 10:
+            raise MoonvalleyApiError("Input Video length is less than 10s. Please use a video longer than or equal to 10s.")
+        if video.get_duration() > 10:
+            # trim video to 10s
+            video = trim_video(video, 10)
+    return video
+
+def trim_video(video: VideoInput, duration_sec: float) -> VideoInput:
+    """
+    Returns a new VideoInput object trimmed from the beginning to the specified duration,
+    using av to avoid loading entire video into memory.
+
+    Args:
+        video: Input video to trim
+        duration_sec: Duration in seconds to keep from the beginning
+
+    Returns:
+        VideoFromFile object that owns the output buffer
+    """
+    output_buffer = io.BytesIO()
+
+    input_container = None
+    output_container = None
+
+    try:
+        # Get the stream source - this avoids loading entire video into memory
+        # when the source is already a file path
+        input_source = video.get_stream_source()
+
+        # Open containers
+        input_container = av.open(input_source, mode='r')
+        output_container = av.open(output_buffer, mode='w', format='mp4')
+
+        # Set up output streams for re-encoding
+        video_stream = None
+        audio_stream = None
+
+        for stream in input_container.streams:
+            logging.info(f"Found stream: type={stream.type}, class={type(stream)}")
+            if isinstance(stream, av.VideoStream):
+                # Create output video stream with same parameters
+                video_stream = output_container.add_stream('h264', rate=stream.average_rate)
+                video_stream.width = stream.width
+                video_stream.height = stream.height
+                video_stream.pix_fmt = 'yuv420p'
+                logging.info(f"Added video stream: {stream.width}x{stream.height} @ {stream.average_rate}fps")
+            elif isinstance(stream, av.AudioStream):
+                # Create output audio stream with same parameters
+                audio_stream = output_container.add_stream('aac', rate=stream.sample_rate)
+                audio_stream.sample_rate = stream.sample_rate
+                audio_stream.layout = stream.layout
+                logging.info(f"Added audio stream: {stream.sample_rate}Hz, {stream.channels} channels")
+
+        # Calculate target frame count that's divisible by 32
+        fps = input_container.streams.video[0].average_rate
+        estimated_frames = int(duration_sec * fps)
+        target_frames = (estimated_frames // 32) * 32  # Round down to nearest multiple of 32
+
+        if target_frames == 0:
+            raise ValueError("Video too short: need at least 32 frames for Moonvalley")
+
+        frame_count = 0
+        audio_frame_count = 0
+
+        # Decode and re-encode video frames
+        if video_stream:
+            for frame in input_container.decode(video=0):
+                if frame_count >= target_frames:
+                    break
+
+                # Re-encode frame
+                for packet in video_stream.encode(frame):
+                    output_container.mux(packet)
+                frame_count += 1
+
+            # Flush encoder
+            for packet in video_stream.encode():
+                output_container.mux(packet)
+
+            logging.info(f"Encoded {frame_count} video frames (target: {target_frames})")
+
+        # Decode and re-encode audio frames
+        if audio_stream:
+            input_container.seek(0)  # Reset to beginning for audio
+            for frame in input_container.decode(audio=0):
+                if frame.time >= duration_sec:
+                    break
+
+                # Re-encode frame
+                for packet in audio_stream.encode(frame):
+                    output_container.mux(packet)
+                audio_frame_count += 1
+
+            # Flush encoder
+            for packet in audio_stream.encode():
+                output_container.mux(packet)
+
+            logging.info(f"Encoded {audio_frame_count} audio frames")
+
+        # Close containers
+        output_container.close()
+        input_container.close()
+
+
+        # Return as VideoFromFile using the buffer
+        output_buffer.seek(0)
+        return VideoFromFile(output_buffer)
+
+    except Exception as e:
+        # Clean up on error
+        if input_container is not None:
+            input_container.close()
+        if output_container is not None:
+            output_container.close()
+        raise RuntimeError(f"Failed to trim video: {str(e)}") from e
+
+# --- BaseMoonvalleyVideoNode ---
+class BaseMoonvalleyVideoNode:
+    def parseWidthHeightFromRes(self, resolution: str):
+        # Accepts a string like "16:9 (1920 x 1080)" and returns width, height as a dict
+        res_map = {
+            "16:9 (1920 x 1080)": {"width": 1920, "height": 1080},
+            "9:16 (1080 x 1920)": {"width": 1080, "height": 1920},
+            "1:1 (1152 x 1152)": {"width": 1152, "height": 1152},
+            "4:3 (1440 x 1080)": {"width": 1440, "height": 1080},
+            "3:4 (1080 x 1440)": {"width": 1080, "height": 1440},
+            "21:9 (2560 x 1080)": {"width": 2560, "height": 1080},
+        }
+        if resolution in res_map:
+            return res_map[resolution]
+        else:
+            # Default to 1920x1080 if unknown
+            return {"width": 1920, "height": 1080}
+
+    def parseControlParameter(self, value):
+        control_map = {
+            "Motion Transfer": "motion_control",
+            "Canny": "canny_control",
+            "Pose Transfer": "pose_control",
+            "Depth": "depth_control"
+        }
+        if value in control_map:
+            return control_map[value]
+        else:
+            return control_map["Motion Transfer"]
+
+    def get_response(
+        self, task_id: str, auth_kwargs: dict[str, str], node_id: Optional[str] = None
+    ) -> MoonvalleyPromptResponse:
+        return poll_until_finished(
+            auth_kwargs,
+            ApiEndpoint(
+                path=f"{API_PROMPTS_ENDPOINT}/{task_id}",
+                method=HttpMethod.GET,
+                request_model=EmptyRequest,
+                response_model=MoonvalleyPromptResponse,
+            ),
+            result_url_extractor=get_video_url_from_response,
+            node_id=node_id,
+        )
+
+    @classmethod
+    def INPUT_TYPES(cls):
+        return {
+            "required": {
+                "prompt": model_field_to_node_input(
+                    IO.STRING, MoonvalleyTextToVideoRequest, "prompt_text",
+                    multiline=True
+                ),
+                "negative_prompt": model_field_to_node_input(
+                    IO.STRING,
+                    MoonvalleyTextToVideoInferenceParams,
+                    "negative_prompt",
+                    multiline=True,
+                    default="gopro, bright, contrast, static, overexposed, bright, vignette, artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, flare, saturation, distorted, warped, wide angle, contrast, saturated, vibrant, glowing, cross dissolve, texture, videogame, saturation, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, blown out, horrible, blurry, worst quality, bad, transition, dissolve, cross-dissolve, melt, fade in, fade out, wobbly, weird, low quality, plastic, stock footage, video camera, boring, static",
+                ),
+
+                "resolution": (IO.COMBO, {
+                        "options": ["16:9 (1920 x 1080)",
+                                    "9:16 (1080 x 1920)",
+                                    "1:1 (1152 x 1152)",
+                                    "4:3 (1440 x 1080)",
+                                    "3:4 (1080 x 1440)",
+                                    "21:9 (2560 x 1080)"],
+                        "default": "16:9 (1920 x 1080)",
+                        "tooltip": "Resolution of the output video",
+                    }),
+                # "length": (IO.COMBO,{"options":['5s','10s'], "default": '5s'}),
+                "prompt_adherence": model_field_to_node_input(IO.FLOAT,MoonvalleyTextToVideoInferenceParams,"guidance_scale",default=7.0, step=1, min=1, max=20),
+                "seed": model_field_to_node_input(IO.INT,MoonvalleyTextToVideoInferenceParams, "seed", default=random.randint(0, 2**32 - 1), min=0, max=4294967295, step=1, display="number", tooltip="Random seed value", control_after_generate=True),
+                "steps": model_field_to_node_input(IO.INT, MoonvalleyTextToVideoInferenceParams, "steps", default=100, min=1, max=100),
+            },
+            "hidden": {
+                "auth_token": "AUTH_TOKEN_COMFY_ORG",
+                "comfy_api_key": "API_KEY_COMFY_ORG",
+                "unique_id": "UNIQUE_ID",
+            },
+            "optional": {
+                "image": model_field_to_node_input(
+                    IO.IMAGE,
+                    MoonvalleyTextToVideoRequest,
+                    "image_url",
+                    tooltip="The reference image used to generate the video",
+                ),
+            }
+        }
+
+    RETURN_TYPES = ("STRING",)
+    FUNCTION = "generate"
+    CATEGORY = "api node/video/Moonvalley Marey"
+    API_NODE = True
+
+    def generate(self, **kwargs):
+        return None
+
+# --- MoonvalleyImg2VideoNode ---
+class MoonvalleyImg2VideoNode(BaseMoonvalleyVideoNode):
+
+    @classmethod
+    def INPUT_TYPES(cls):
+        return super().INPUT_TYPES()
+
+    RETURN_TYPES = ("VIDEO",)
+    RETURN_NAMES = ("video",)
+    DESCRIPTION = "Moonvalley Marey Image to Video Node"
+
+    def generate(self, prompt, negative_prompt, unique_id: Optional[str] = None, **kwargs):
+        image = kwargs.get("image", None)
+        if (image is None):
+            raise MoonvalleyApiError("image is required")
+        total_frames = get_total_frames_from_length()
+
+        validate_input_image(image,True)
+        validate_prompts(prompt, negative_prompt, MOONVALLEY_MAREY_MAX_PROMPT_LENGTH)
+        width_height = self.parseWidthHeightFromRes(kwargs.get("resolution"))
+
+        inference_params=MoonvalleyTextToVideoInferenceParams(
+                    negative_prompt=negative_prompt,
+                    steps=kwargs.get("steps"),
+                    seed=kwargs.get("seed"),
+                    guidance_scale=kwargs.get("prompt_adherence"),
+                    num_frames=total_frames,
+                    width=width_height.get("width"),
+                    height=width_height.get("height"),
+                    use_negative_prompts=True
+                )
+        """Upload image to comfy backend to have a URL available for further processing"""
+        # Get MIME type from tensor - assuming PNG format for image tensors
+        mime_type = "image/png"
+
+        image_url = upload_images_to_comfyapi(image, max_images=1, auth_kwargs=kwargs, mime_type=mime_type)[0]
+
+        request = MoonvalleyTextToVideoRequest(
+                image_url=image_url,
+                prompt_text=prompt,
+                inference_params=inference_params
+            )
+        initial_operation = SynchronousOperation(
+            endpoint=ApiEndpoint(path=API_IMG2VIDEO_ENDPOINT,
+                                 method=HttpMethod.POST,
+                                 request_model=MoonvalleyTextToVideoRequest,
+                                 response_model=MoonvalleyPromptResponse
+                                 ),
+            request=request,
+            auth_kwargs=kwargs,
+        )
+        task_creation_response = initial_operation.execute()
+        validate_task_creation_response(task_creation_response)
+        task_id = task_creation_response.id
+
+        final_response = self.get_response(
+            task_id, auth_kwargs=kwargs, node_id=unique_id
+        )
+        video = download_url_to_video_output(final_response.output_url)
+        return (video, )
+
+# --- MoonvalleyVid2VidNode ---
+class MoonvalleyVideo2VideoNode(BaseMoonvalleyVideoNode):
+    def __init__(self):
+        super().__init__()
+
+    @classmethod
+    def INPUT_TYPES(cls):
+        input_types = super().INPUT_TYPES()
+        for param in ["resolution", "image"]:
+            if param in input_types["required"]:
+                del input_types["required"][param]
+            if param in input_types["optional"]:
+                del input_types["optional"][param]
+        input_types["optional"] = {
+                "video": (IO.VIDEO, {"default": "", "multiline": False, "tooltip": "The reference video used to generate the output video. Input a 5s video for 128 frames and a 10s video for 256 frames. Longer videos will be trimmed automatically."}),
+                "control_type": (
+                    ["Motion Transfer", "Pose Transfer"],
+                    {"default": "Motion Transfer"},
+                ),
+                "motion_intensity": (
+                    "INT",
+                    {
+                        "default": 100,
+                        "step": 1,
+                        "min": 0,
+                        "max": 100,
+                        "tooltip": "Only used if control_type is 'Motion Transfer'",
+                    },
+                )
+            }
+
+        return input_types
+
+    RETURN_TYPES = ("VIDEO",)
+    RETURN_NAMES = ("video",)
+
+    def generate(self, prompt, negative_prompt, unique_id: Optional[str] = None, **kwargs):
+        video = kwargs.get("video")
+        num_frames = get_total_frames_from_length()
+
+        if not video :
+            raise MoonvalleyApiError("video is required")
+
+
+        """Validate video input"""
+        video_url=""
+        if video:
+            validated_video = validate_input_video(video, num_frames, False)
+            video_url = upload_video_to_comfyapi(validated_video, auth_kwargs=kwargs)
+
+        control_type = kwargs.get("control_type")
+        motion_intensity = kwargs.get("motion_intensity")
+
+        """Validate prompts and inference input"""
+        validate_prompts(prompt, negative_prompt)
+        inference_params=MoonvalleyVideoToVideoInferenceParams(
+            negative_prompt=negative_prompt,
+            steps=kwargs.get("steps"),
+            seed=kwargs.get("seed"),
+            guidance_scale=kwargs.get("prompt_adherence"),
+            control_params={'motion_intensity': motion_intensity}
+        )
+
+        control = self.parseControlParameter(control_type)
+
+        request = MoonvalleyVideoToVideoRequest(
+                control_type=control,
+                video_url=video_url,
+                prompt_text=prompt,
+                inference_params=inference_params
+            )
+
+        initial_operation = SynchronousOperation(
+            endpoint=ApiEndpoint(path=API_VIDEO2VIDEO_ENDPOINT,
+                                 method=HttpMethod.POST,
+                                 request_model=MoonvalleyVideoToVideoRequest,
+                                 response_model=MoonvalleyPromptResponse
+                                 ),
+            request=request,
+            auth_kwargs=kwargs,
+        )
+        task_creation_response = initial_operation.execute()
+        validate_task_creation_response(task_creation_response)
+        task_id = task_creation_response.id
+
+        final_response = self.get_response(
+            task_id, auth_kwargs=kwargs, node_id=unique_id
+        )
+
+        video = download_url_to_video_output(final_response.output_url)
+
+        return (video, )
+
+# --- MoonvalleyTxt2VideoNode ---
+class MoonvalleyTxt2VideoNode(BaseMoonvalleyVideoNode):
+    def __init__(self):
+        super().__init__()
+
+    RETURN_TYPES = ("VIDEO",)
+    RETURN_NAMES = ("video",)
+
+    @classmethod
+    def INPUT_TYPES(cls):
+        input_types = super().INPUT_TYPES()
+        # Remove image-specific parameters
+        for param in ["image"]:
+            if param in input_types["optional"]:
+                del input_types["optional"][param]
+        return input_types
+
+    def generate(self, prompt, negative_prompt, unique_id: Optional[str] = None, **kwargs):
+        validate_prompts(prompt, negative_prompt, MOONVALLEY_MAREY_MAX_PROMPT_LENGTH)
+        width_height = self.parseWidthHeightFromRes(kwargs.get("resolution"))
+        num_frames = get_total_frames_from_length()
+
+        inference_params=MoonvalleyTextToVideoInferenceParams(
+                    negative_prompt=negative_prompt,
+                    steps=kwargs.get("steps"),
+                    seed=kwargs.get("seed"),
+                    guidance_scale=kwargs.get("prompt_adherence"),
+                    num_frames=num_frames,
+                    width=width_height.get("width"),
+                    height=width_height.get("height"),
+                )
+        request = MoonvalleyTextToVideoRequest(
+                prompt_text=prompt,
+                inference_params=inference_params
+            )
+
+        initial_operation = SynchronousOperation(
+            endpoint=ApiEndpoint(path=API_TXT2VIDEO_ENDPOINT,
+                                 method=HttpMethod.POST,
+                                 request_model=MoonvalleyTextToVideoRequest,
+                                 response_model=MoonvalleyPromptResponse
+                                 ),
+            request=request,
+            auth_kwargs=kwargs,
+        )
+        task_creation_response = initial_operation.execute()
+        validate_task_creation_response(task_creation_response)
+        task_id = task_creation_response.id
+
+        final_response = self.get_response(
+            task_id, auth_kwargs=kwargs, node_id=unique_id
+        )
+
+        video = download_url_to_video_output(final_response.output_url)
+        return (video, )
+
+
+
+NODE_CLASS_MAPPINGS = {
+    "MoonvalleyImg2VideoNode": MoonvalleyImg2VideoNode,
+    "MoonvalleyTxt2VideoNode": MoonvalleyTxt2VideoNode,
+    # "MoonvalleyVideo2VideoNode": MoonvalleyVideo2VideoNode,
+}
+
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+    "MoonvalleyImg2VideoNode": "Moonvalley Marey Image to Video",
+    "MoonvalleyTxt2VideoNode": "Moonvalley Marey Text to Video",
+    # "MoonvalleyVideo2VideoNode": "Moonvalley Marey Video to Video",
+}
+
+def get_total_frames_from_length(length="5s"):
+    # if length == '5s':
+    #     return 128
+    # elif length == '10s':
+    #     return 256
+    return 128
+    # else:
+    #     raise MoonvalleyApiError("length is required")

comfy_extras/nodes_custom_sampler.py

@@ -2,6 +2,7 @@ import math
 import comfy.samplers
 import comfy.sample
 from comfy.k_diffusion import sampling as k_diffusion_sampling
+from comfy.k_diffusion import sa_solver
 from comfy.comfy_types import IO, ComfyNodeABC, InputTypeDict
 import latent_preview
 import torch
@@ -521,6 +522,49 @@ class SamplerER_SDE(ComfyNodeABC):
         return (sampler,)
 
 
+class SamplerSASolver(ComfyNodeABC):
+    @classmethod
+    def INPUT_TYPES(cls) -> InputTypeDict:
+        return {
+            "required": {
+                "model": (IO.MODEL, {}),
+                "eta": (IO.FLOAT, {"default": 1.0, "min": 0.0, "max": 10.0, "step": 0.01, "round": False},),
+                "sde_start_percent": (IO.FLOAT, {"default": 0.2, "min": 0.0, "max": 1.0, "step": 0.001},),
+                "sde_end_percent": (IO.FLOAT, {"default": 0.8, "min": 0.0, "max": 1.0, "step": 0.001},),
+                "s_noise": (IO.FLOAT, {"default": 1.0, "min": 0.0, "max": 100.0, "step": 0.01, "round": False},),
+                "predictor_order": (IO.INT, {"default": 3, "min": 1, "max": 6}),
+                "corrector_order": (IO.INT, {"default": 4, "min": 0, "max": 6}),
+                "use_pece": (IO.BOOLEAN, {}),
+                "simple_order_2": (IO.BOOLEAN, {}),
+            }
+        }
+
+    RETURN_TYPES = (IO.SAMPLER,)
+    CATEGORY = "sampling/custom_sampling/samplers"
+
+    FUNCTION = "get_sampler"
+
+    def get_sampler(self, model, eta, sde_start_percent, sde_end_percent, s_noise, predictor_order, corrector_order, use_pece, simple_order_2):
+        model_sampling = model.get_model_object("model_sampling")
+        start_sigma = model_sampling.percent_to_sigma(sde_start_percent)
+        end_sigma = model_sampling.percent_to_sigma(sde_end_percent)
+        tau_func = sa_solver.get_tau_interval_func(start_sigma, end_sigma, eta=eta)
+
+        sampler_name = "sa_solver"
+        sampler = comfy.samplers.ksampler(
+            sampler_name,
+            {
+                "tau_func": tau_func,
+                "s_noise": s_noise,
+                "predictor_order": predictor_order,
+                "corrector_order": corrector_order,
+                "use_pece": use_pece,
+                "simple_order_2": simple_order_2,
+            },
+        )
+        return (sampler,)
+
+
 class Noise_EmptyNoise:
     def __init__(self):
         self.seed = 0
@@ -829,6 +873,7 @@ NODE_CLASS_MAPPINGS = {
     "SamplerDPMPP_2S_Ancestral": SamplerDPMPP_2S_Ancestral,
     "SamplerDPMAdaptative": SamplerDPMAdaptative,
     "SamplerER_SDE": SamplerER_SDE,
+    "SamplerSASolver": SamplerSASolver,
     "SplitSigmas": SplitSigmas,
     "SplitSigmasDenoise": SplitSigmasDenoise,
     "FlipSigmas": FlipSigmas,

comfyui_version.py

@@ -1,3 +1,3 @@
 # This file is automatically generated by the build process when version is
 # updated in pyproject.toml.
-__version__ = "0.3.43"
+__version__ = "0.3.44"

nodes.py

@@ -2310,6 +2310,7 @@ def init_builtin_api_nodes():
         "nodes_pika.py",
         "nodes_runway.py",
         "nodes_tripo.py",
+        "nodes_moonvalley.py",
         "nodes_rodin.py",
         "nodes_gemini.py",
     ]

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "ComfyUI"
-version = "0.3.43"
+version = "0.3.44"
 readme = "README.md"
 license = { file = "LICENSE" }
 requires-python = ">=3.9"

requirements.txt

@@ -1,5 +1,5 @@
 comfyui-frontend-package==1.23.4
-comfyui-workflow-templates==0.1.34
+comfyui-workflow-templates==0.1.35
 comfyui-embedded-docs==0.2.4
 torch
 torchsde