mirror of
https://github.com/Comfy-Org/ComfyUI.git
synced 2026-08-18 23:38:29 +08:00
Merge remote-tracking branch 'origin/master' into pr-15148
This commit is contained in:
Claude
@@ -231,6 +231,25 @@ class IdeogramV3Request(BaseModel):
|
||||
)
|
||||
|
||||
|
||||
class IdeogramPImageRequest(BaseModel):
|
||||
prompt: str = Field(
|
||||
...,
|
||||
description="The text prompt, or an Ideogram 4.0 structured JSON caption "
|
||||
"(used verbatim when prompt_upsampling is 'OFF').",
|
||||
)
|
||||
quality: str | None = Field(
|
||||
None, description="Generation tier: 'VERY_LOW', 'LOW', 'MEDIUM' or 'HIGH'."
|
||||
)
|
||||
resolution: str | None = Field(None, description="Output size class: '1K' or '2K'.")
|
||||
aspect_ratio: str | None = Field(
|
||||
None, description="Aspect ratio in WxH format", examples=['16x9']
|
||||
)
|
||||
prompt_upsampling: str | None = Field(
|
||||
None, description="Prompt expansion: 'AUTO', 'ON' or 'OFF'."
|
||||
)
|
||||
seed: int | None = Field(None, ge=0, le=2147483647)
|
||||
|
||||
|
||||
class IdeogramV4Request(BaseModel):
|
||||
text_prompt: str | None = Field(
|
||||
None,
|
||||
|
||||
@@ -118,3 +118,81 @@ class MinimaxVideoGenerationResponse(BaseModel):
|
||||
task_id: str = Field(
|
||||
..., description='The task ID for the asynchronous video generation task.'
|
||||
)
|
||||
|
||||
|
||||
class Hailuo03TextContent(BaseModel):
|
||||
type: str = Field("text")
|
||||
text: str = Field(...)
|
||||
|
||||
|
||||
class Hailuo03ImageContentUrl(BaseModel):
|
||||
url: str = Field(...)
|
||||
|
||||
|
||||
class Hailuo03ImageContent(BaseModel):
|
||||
type: str = Field("image_url")
|
||||
image_url: Hailuo03ImageContentUrl = Field(...)
|
||||
role: str = Field(...)
|
||||
|
||||
|
||||
class Hailuo03VideoContentUrl(BaseModel):
|
||||
url: str = Field(...)
|
||||
|
||||
|
||||
class Hailuo03VideoContent(BaseModel):
|
||||
type: str = Field("video_url")
|
||||
video_url: Hailuo03VideoContentUrl = Field(...)
|
||||
role: str = Field("reference_video")
|
||||
|
||||
|
||||
class Hailuo03AudioContentUrl(BaseModel):
|
||||
url: str = Field(...)
|
||||
|
||||
|
||||
class Hailuo03AudioContent(BaseModel):
|
||||
type: str = Field("audio_url")
|
||||
audio_url: Hailuo03AudioContentUrl = Field(...)
|
||||
role: str = Field("reference_audio")
|
||||
|
||||
|
||||
class Hailuo03TaskCreationRequest(BaseModel):
|
||||
model: str = Field(...)
|
||||
content: list[Hailuo03TextContent | Hailuo03ImageContent | Hailuo03VideoContent | Hailuo03AudioContent] = Field(
|
||||
..., min_length=1
|
||||
)
|
||||
resolution: str = Field(...)
|
||||
duration: int = Field(..., ge=5, le=15)
|
||||
ratio: str | None = Field(None)
|
||||
seed: int | None = Field(None, ge=0, le=4294967295)
|
||||
aigc_watermark: bool | None = Field(None)
|
||||
|
||||
|
||||
class Hailuo03TaskCreationResponse(BaseModel):
|
||||
task_id: str = Field(...)
|
||||
|
||||
|
||||
class Hailuo03TaskError(BaseModel):
|
||||
code: int | str | None = Field(None)
|
||||
message: str | None = Field(None)
|
||||
|
||||
|
||||
class Hailuo03TaskContent(BaseModel):
|
||||
url: str | None = Field(None)
|
||||
|
||||
|
||||
class Hailuo03TaskUsage(BaseModel):
|
||||
total_seconds: float = Field(0)
|
||||
input_seconds: float = Field(0)
|
||||
output_seconds: float = Field(0)
|
||||
|
||||
|
||||
class Hailuo03Task(BaseModel):
|
||||
id: str = Field(...)
|
||||
status: str = Field(...)
|
||||
error: Hailuo03TaskError | None = Field(None)
|
||||
content: Hailuo03TaskContent | None = Field(None)
|
||||
usage: Hailuo03TaskUsage | None = Field(None)
|
||||
|
||||
|
||||
class Hailuo03TaskQueryResponse(BaseModel):
|
||||
task: Hailuo03Task = Field(...)
|
||||
|
||||
@@ -289,7 +289,7 @@ class BriaRemoveVideoBackground(IO.ComfyNode):
|
||||
],
|
||||
is_api_node=True,
|
||||
price_badge=IO.PriceBadge(
|
||||
expr="""{"type":"usd","usd":0.0042,"format":{"suffix":"/second"}}""",
|
||||
expr="""{"type":"usd","usd":0.005,"format":{"suffix":"/second"}}""",
|
||||
),
|
||||
)
|
||||
|
||||
@@ -533,7 +533,7 @@ class BriaTransparentVideoBackground(IO.ComfyNode):
|
||||
],
|
||||
is_api_node=True,
|
||||
price_badge=IO.PriceBadge(
|
||||
expr="""{"type":"usd","usd":0.0042,"format":{"suffix":"/second"}}""",
|
||||
expr="""{"type":"usd","usd":0.005,"format":{"suffix":"/second"}}""",
|
||||
),
|
||||
)
|
||||
|
||||
|
||||
@@ -6,6 +6,7 @@ import numpy as np
|
||||
import torch
|
||||
from comfy_api_nodes.apis.ideogram import (
|
||||
IdeogramGenerateResponse,
|
||||
IdeogramPImageRequest,
|
||||
IdeogramV3Request,
|
||||
IdeogramV3EditRequest,
|
||||
IdeogramV4Request,
|
||||
@@ -524,12 +525,155 @@ class IdeogramV4(IO.ComfyNode):
|
||||
return IO.NodeOutput(await download_and_process_images(image_urls))
|
||||
|
||||
|
||||
class IdeogramPImage(IO.ComfyNode):
|
||||
|
||||
@classmethod
|
||||
def define_schema(cls):
|
||||
return IO.Schema(
|
||||
node_id="IdeogramPImage",
|
||||
display_name="Ideogram P-Image",
|
||||
category="partner/image/Ideogram",
|
||||
description="Generates images using P-Image, Ideogram's fast text-to-image model. "
|
||||
"Strong typography and photorealism; "
|
||||
"supports Ideogram 4.0 structured JSON captions for exact text, "
|
||||
"colors and layout.",
|
||||
inputs=[
|
||||
IO.String.Input(
|
||||
"prompt",
|
||||
multiline=True,
|
||||
default="",
|
||||
tooltip="Text prompt. Also accepts an Ideogram 4.0 structured JSON caption "
|
||||
"(exact colors as #RRGGBB hexes, exact text strings, bounding-box "
|
||||
"layout) — set prompt_upsampling to OFF to use it verbatim.",
|
||||
),
|
||||
IO.Combo.Input(
|
||||
"quality",
|
||||
options=["VERY_LOW", "LOW", "MEDIUM", "HIGH"],
|
||||
default="MEDIUM",
|
||||
tooltip="Speed/price/quality tier. MEDIUM is the everyday default; HIGH for "
|
||||
"complex prompts, fine detail and difficult text; VERY_LOW/LOW for "
|
||||
"drafts at scale. Difficult text renders poorly below MEDIUM.",
|
||||
),
|
||||
IO.Combo.Input(
|
||||
"resolution",
|
||||
options=["1K", "2K"],
|
||||
default="1K",
|
||||
tooltip="Output size class (exact pixels follow the aspect ratio, e.g. "
|
||||
"16:9 gives 1280x720 at 1K and 2560x1440 at 2K). "
|
||||
"Prefer HIGH + 2K for crisp typography.",
|
||||
),
|
||||
IO.Combo.Input(
|
||||
"aspect_ratio",
|
||||
options=list(V3_RATIO_MAP.keys()),
|
||||
default="1:1",
|
||||
tooltip="The aspect ratio for image generation.",
|
||||
),
|
||||
IO.Combo.Input(
|
||||
"prompt_upsampling",
|
||||
options=["AUTO", "ON", "OFF"],
|
||||
default="AUTO",
|
||||
tooltip="Expands short prompts into a detailed structured caption before "
|
||||
"generation (the rewritten prompt is returned as final_prompt). "
|
||||
"Set OFF when supplying your own JSON caption or exact wording.",
|
||||
),
|
||||
IO.Int.Input(
|
||||
"seed",
|
||||
default=42,
|
||||
min=0,
|
||||
max=2147483647,
|
||||
step=1,
|
||||
control_after_generate=True,
|
||||
display_mode=IO.NumberDisplay.number,
|
||||
optional=True,
|
||||
tooltip="Seed for reproducible generation. With prompt_upsampling OFF, "
|
||||
"the same seed and settings return the same image; with ON/AUTO "
|
||||
"the prompt rewrite varies per run — reproduce a result by reusing "
|
||||
"its final_prompt output with prompt_upsampling OFF and the same "
|
||||
"seed.",
|
||||
),
|
||||
],
|
||||
outputs=[
|
||||
IO.Image.Output(),
|
||||
IO.String.Output(
|
||||
"final_prompt",
|
||||
tooltip="The prompt the image was actually generated from (the rewritten "
|
||||
"structured caption when prompt_upsampling ran, else your prompt). "
|
||||
"Feed it back with prompt_upsampling OFF and the same seed to "
|
||||
"reproduce this image.",
|
||||
),
|
||||
],
|
||||
hidden=[
|
||||
IO.Hidden.auth_token_comfy_org,
|
||||
IO.Hidden.api_key_comfy_org,
|
||||
IO.Hidden.unique_id,
|
||||
],
|
||||
is_api_node=True,
|
||||
price_badge=IO.PriceBadge(
|
||||
depends_on=IO.PriceBadgeDepends(widgets=["quality", "resolution"]),
|
||||
expr="""
|
||||
(
|
||||
$q := widgets.quality;
|
||||
$is2k := $contains(widgets.resolution, "2k");
|
||||
$usd :=
|
||||
$contains($q, "very_low") ? ($is2k ? 0.00858 : 0.00429) :
|
||||
$contains($q, "high") ? ($is2k ? 0.0429 : 0.02145) :
|
||||
$contains($q, "medium") ? ($is2k ? 0.0286 : 0.0143) :
|
||||
($is2k ? 0.02145 : 0.010725);
|
||||
{"type": "usd", "usd": $usd}
|
||||
)
|
||||
""",
|
||||
),
|
||||
)
|
||||
|
||||
@classmethod
|
||||
async def execute(
|
||||
cls,
|
||||
prompt: str,
|
||||
quality: str = "MEDIUM",
|
||||
resolution: str = "1K",
|
||||
aspect_ratio: str = "1:1",
|
||||
prompt_upsampling: str = "AUTO",
|
||||
seed: int = 42,
|
||||
):
|
||||
validate_string(prompt, strip_whitespace=True, min_length=1)
|
||||
request = IdeogramPImageRequest(
|
||||
prompt=prompt,
|
||||
quality=quality,
|
||||
resolution=resolution,
|
||||
aspect_ratio=V3_RATIO_MAP[aspect_ratio],
|
||||
prompt_upsampling=prompt_upsampling,
|
||||
seed=seed,
|
||||
)
|
||||
response = await sync_op(
|
||||
cls,
|
||||
ApiEndpoint(path="/proxy/ideogram/text-to-image/p-image-ideogram", method="POST"),
|
||||
response_model=IdeogramGenerateResponse,
|
||||
data=request,
|
||||
max_retries=1,
|
||||
)
|
||||
if not response.data:
|
||||
raise Exception("No images were generated in the response")
|
||||
image_urls = [image_data.url for image_data in response.data if image_data.url]
|
||||
if not image_urls:
|
||||
if any(image_data.is_image_safe is False for image_data in response.data):
|
||||
raise Exception(
|
||||
"The generation was blocked by Ideogram's content safety filter. "
|
||||
"Adjust the prompt and try again."
|
||||
)
|
||||
raise Exception("No image URLs were generated in the response")
|
||||
return IO.NodeOutput(
|
||||
await download_and_process_images(image_urls),
|
||||
response.data[0].prompt or prompt,
|
||||
)
|
||||
|
||||
|
||||
class IdeogramExtension(ComfyExtension):
|
||||
@override
|
||||
async def get_node_list(self) -> list[type[IO.ComfyNode]]:
|
||||
return [
|
||||
IdeogramV3,
|
||||
IdeogramV4,
|
||||
IdeogramPImage,
|
||||
]
|
||||
|
||||
|
||||
|
||||
@@ -5,6 +5,16 @@ from typing_extensions import override
|
||||
|
||||
from comfy_api.latest import IO, ComfyExtension
|
||||
from comfy_api_nodes.apis.minimax import (
|
||||
Hailuo03AudioContent,
|
||||
Hailuo03AudioContentUrl,
|
||||
Hailuo03ImageContent,
|
||||
Hailuo03ImageContentUrl,
|
||||
Hailuo03TaskCreationRequest,
|
||||
Hailuo03TaskCreationResponse,
|
||||
Hailuo03TaskQueryResponse,
|
||||
Hailuo03TextContent,
|
||||
Hailuo03VideoContent,
|
||||
Hailuo03VideoContentUrl,
|
||||
MinimaxFileRetrieveResponse,
|
||||
MiniMaxModel,
|
||||
MinimaxTaskResultResponse,
|
||||
@@ -17,7 +27,11 @@ from comfy_api_nodes.util import (
|
||||
download_url_to_video_output,
|
||||
poll_op,
|
||||
sync_op,
|
||||
upload_audio_to_comfyapi,
|
||||
upload_images_to_comfyapi,
|
||||
upload_video_to_comfyapi,
|
||||
validate_image_aspect_ratio,
|
||||
validate_image_dimensions,
|
||||
validate_string,
|
||||
)
|
||||
|
||||
@@ -293,9 +307,9 @@ class MinimaxHailuoVideoNode(IO.ComfyNode):
|
||||
def define_schema(cls) -> IO.Schema:
|
||||
return IO.Schema(
|
||||
node_id="MinimaxHailuoVideoNode",
|
||||
display_name="MiniMax Hailuo Video",
|
||||
display_name="MiniMax Hailuo 02 Video",
|
||||
category="partner/video/MiniMax",
|
||||
description="Generates videos from prompt, with optional start frame using the new MiniMax Hailuo-02 model.",
|
||||
description="Generates videos from prompt, with optional start frame using the MiniMax Hailuo-02 model.",
|
||||
inputs=[
|
||||
IO.String.Input(
|
||||
"prompt_text",
|
||||
@@ -437,6 +451,491 @@ class MinimaxHailuoVideoNode(IO.ComfyNode):
|
||||
return IO.NodeOutput(await download_url_to_video_output(file_url))
|
||||
|
||||
|
||||
HAILUO_03_CREATE_ENDPOINT = "/proxy/minimax/v2/video_generation"
|
||||
HAILUO_03_QUERY_ENDPOINT = "/proxy/minimax/v2/query/video_generation" # + /{task_id}
|
||||
HAILUO_03_MODELS = {"MiniMax H3": "MiniMax-H3"}
|
||||
HAILUO_03_FAILED_STATUSES = ["failed", "cancelled", "expired"]
|
||||
|
||||
|
||||
def _hailuo03_model_inputs(include_ratio: bool = True, allow_adaptive: bool = True):
|
||||
inputs = [
|
||||
IO.String.Input(
|
||||
"prompt",
|
||||
multiline=True,
|
||||
default="",
|
||||
tooltip="Text prompt for video generation.",
|
||||
),
|
||||
IO.Combo.Input(
|
||||
"resolution",
|
||||
options=["2K"],
|
||||
tooltip="Resolution of the output video.",
|
||||
),
|
||||
]
|
||||
if include_ratio:
|
||||
ratio_options = ["16:9", "4:3", "1:1", "3:4", "9:16", "21:9"]
|
||||
if allow_adaptive:
|
||||
ratio_options.insert(0, "adaptive")
|
||||
inputs.append(
|
||||
IO.Combo.Input(
|
||||
"ratio",
|
||||
options=ratio_options,
|
||||
default=ratio_options[0],
|
||||
tooltip="Aspect ratio of the output video.",
|
||||
)
|
||||
)
|
||||
inputs.append(
|
||||
IO.Int.Input(
|
||||
"duration",
|
||||
default=5,
|
||||
min=5,
|
||||
max=15,
|
||||
step=1,
|
||||
tooltip="Duration of the output video in seconds (5-15).",
|
||||
display_mode=IO.NumberDisplay.slider,
|
||||
)
|
||||
)
|
||||
return inputs
|
||||
|
||||
|
||||
async def _hailuo03_run_task(
|
||||
cls: type[IO.ComfyNode],
|
||||
*,
|
||||
model_id: str,
|
||||
content: list,
|
||||
resolution: str,
|
||||
duration: int,
|
||||
ratio: str | None,
|
||||
seed: int,
|
||||
watermark: bool,
|
||||
) -> IO.NodeOutput:
|
||||
response = await sync_op(
|
||||
cls,
|
||||
ApiEndpoint(path=HAILUO_03_CREATE_ENDPOINT, method="POST"),
|
||||
response_model=Hailuo03TaskCreationResponse,
|
||||
data=Hailuo03TaskCreationRequest(
|
||||
model=model_id,
|
||||
content=content,
|
||||
resolution=resolution,
|
||||
duration=duration,
|
||||
ratio=ratio,
|
||||
seed=seed,
|
||||
aigc_watermark=watermark,
|
||||
),
|
||||
)
|
||||
task_result = await poll_op(
|
||||
cls,
|
||||
ApiEndpoint(path=f"{HAILUO_03_QUERY_ENDPOINT}/{response.task_id}"),
|
||||
response_model=Hailuo03TaskQueryResponse,
|
||||
status_extractor=lambda r: r.task.status,
|
||||
failed_statuses=HAILUO_03_FAILED_STATUSES,
|
||||
poll_interval=15,
|
||||
)
|
||||
video_url = task_result.task.content.url if task_result.task.content else None
|
||||
if not video_url:
|
||||
raise Exception(f"No video URL in the response: {task_result.model_dump()}")
|
||||
return IO.NodeOutput(await download_url_to_video_output(video_url))
|
||||
|
||||
|
||||
class MinimaxHailuo03TextToVideoNode(IO.ComfyNode):
|
||||
@classmethod
|
||||
def define_schema(cls):
|
||||
return IO.Schema(
|
||||
node_id="MinimaxHailuo03TextToVideoNode",
|
||||
display_name="MiniMax H3 Text to Video",
|
||||
category="partner/video/MiniMax",
|
||||
description="Generate video from a text prompt using the MiniMax H3 model.",
|
||||
inputs=[
|
||||
IO.DynamicCombo.Input(
|
||||
"model",
|
||||
options=[IO.DynamicCombo.Option("MiniMax H3", _hailuo03_model_inputs(allow_adaptive=False))],
|
||||
tooltip="Model to use for video generation.",
|
||||
),
|
||||
IO.Int.Input(
|
||||
"seed",
|
||||
default=42,
|
||||
min=0,
|
||||
max=4294967295,
|
||||
step=1,
|
||||
display_mode=IO.NumberDisplay.number,
|
||||
control_after_generate=True,
|
||||
tooltip="Random seed. The same request with the same seed gives similar, "
|
||||
"but not guaranteed identical, results.",
|
||||
),
|
||||
IO.Boolean.Input(
|
||||
"watermark",
|
||||
default=False,
|
||||
tooltip="Whether to add an AIGC watermark to the video.",
|
||||
advanced=True,
|
||||
),
|
||||
],
|
||||
outputs=[
|
||||
IO.Video.Output(),
|
||||
],
|
||||
hidden=[
|
||||
IO.Hidden.auth_token_comfy_org,
|
||||
IO.Hidden.api_key_comfy_org,
|
||||
IO.Hidden.unique_id,
|
||||
],
|
||||
is_api_node=True,
|
||||
price_badge=IO.PriceBadge(
|
||||
depends_on=IO.PriceBadgeDepends(widgets=["model.duration"]),
|
||||
expr="""
|
||||
(
|
||||
$dur := $lookup(widgets, "model.duration");
|
||||
{"type": "usd", "usd": $dur * 0.1859}
|
||||
)
|
||||
""",
|
||||
),
|
||||
)
|
||||
|
||||
@classmethod
|
||||
async def execute(
|
||||
cls,
|
||||
model: dict,
|
||||
seed: int,
|
||||
watermark: bool,
|
||||
) -> IO.NodeOutput:
|
||||
validate_string(model["prompt"], strip_whitespace=True, min_length=1)
|
||||
return await _hailuo03_run_task(
|
||||
cls,
|
||||
model_id=HAILUO_03_MODELS[model["model"]],
|
||||
content=[Hailuo03TextContent(text=model["prompt"])],
|
||||
resolution=model["resolution"],
|
||||
duration=model["duration"],
|
||||
ratio=model["ratio"],
|
||||
seed=seed,
|
||||
watermark=watermark,
|
||||
)
|
||||
|
||||
|
||||
class MinimaxHailuo03FirstLastFrameNode(IO.ComfyNode):
|
||||
@classmethod
|
||||
def define_schema(cls):
|
||||
return IO.Schema(
|
||||
node_id="MinimaxHailuo03FirstLastFrameNode",
|
||||
display_name="MiniMax H3 First-Last-Frame to Video",
|
||||
category="partner/video/MiniMax",
|
||||
description="Generate video from a first frame image and an optional last frame image "
|
||||
"using the MiniMax H3 model. The aspect ratio of the video follows the supplied images.",
|
||||
inputs=[
|
||||
IO.DynamicCombo.Input(
|
||||
"model",
|
||||
options=[IO.DynamicCombo.Option("MiniMax H3", _hailuo03_model_inputs(include_ratio=False))],
|
||||
tooltip="Model to use for video generation.",
|
||||
),
|
||||
IO.Image.Input(
|
||||
"first_frame",
|
||||
tooltip="First frame image for the video.",
|
||||
),
|
||||
IO.Image.Input(
|
||||
"last_frame",
|
||||
tooltip="Optional last frame image for the video.",
|
||||
optional=True,
|
||||
),
|
||||
IO.Int.Input(
|
||||
"seed",
|
||||
default=42,
|
||||
min=0,
|
||||
max=4294967295,
|
||||
step=1,
|
||||
display_mode=IO.NumberDisplay.number,
|
||||
control_after_generate=True,
|
||||
tooltip="Random seed. The same request with the same seed gives similar, "
|
||||
"but not guaranteed identical, results.",
|
||||
),
|
||||
IO.Boolean.Input(
|
||||
"watermark",
|
||||
default=False,
|
||||
tooltip="Whether to add an AIGC watermark to the video.",
|
||||
advanced=True,
|
||||
),
|
||||
],
|
||||
outputs=[
|
||||
IO.Video.Output(),
|
||||
],
|
||||
hidden=[
|
||||
IO.Hidden.auth_token_comfy_org,
|
||||
IO.Hidden.api_key_comfy_org,
|
||||
IO.Hidden.unique_id,
|
||||
],
|
||||
is_api_node=True,
|
||||
price_badge=IO.PriceBadge(
|
||||
depends_on=IO.PriceBadgeDepends(widgets=["model.duration"]),
|
||||
expr="""
|
||||
(
|
||||
$dur := $lookup(widgets, "model.duration");
|
||||
{"type": "usd", "usd": $dur * 0.1859}
|
||||
)
|
||||
""",
|
||||
),
|
||||
)
|
||||
|
||||
@classmethod
|
||||
async def execute(
|
||||
cls,
|
||||
model: dict,
|
||||
first_frame: torch.Tensor,
|
||||
seed: int,
|
||||
watermark: bool,
|
||||
last_frame: torch.Tensor | None = None,
|
||||
) -> IO.NodeOutput:
|
||||
validate_string(model["prompt"], strip_whitespace=True, min_length=1)
|
||||
for frame in (first_frame, last_frame):
|
||||
if frame is not None:
|
||||
validate_image_aspect_ratio(frame, (2, 5), (5, 2), strict=False) # 0.4 to 2.5
|
||||
validate_image_dimensions(frame, min_width=256, min_height=256)
|
||||
|
||||
content: list = [
|
||||
Hailuo03TextContent(text=model["prompt"]),
|
||||
Hailuo03ImageContent(
|
||||
image_url=Hailuo03ImageContentUrl(
|
||||
url=(
|
||||
await upload_images_to_comfyapi(
|
||||
cls, first_frame, max_images=1, wait_label="Uploading first frame"
|
||||
)
|
||||
)[0],
|
||||
),
|
||||
role="first_frame",
|
||||
),
|
||||
]
|
||||
if last_frame is not None:
|
||||
content.append(
|
||||
Hailuo03ImageContent(
|
||||
image_url=Hailuo03ImageContentUrl(
|
||||
url=(
|
||||
await upload_images_to_comfyapi(
|
||||
cls, last_frame, max_images=1, wait_label="Uploading last frame"
|
||||
)
|
||||
)[0],
|
||||
),
|
||||
role="last_frame",
|
||||
)
|
||||
)
|
||||
return await _hailuo03_run_task(
|
||||
cls,
|
||||
model_id=HAILUO_03_MODELS[model["model"]],
|
||||
content=content,
|
||||
resolution=model["resolution"],
|
||||
duration=model["duration"],
|
||||
ratio=None,
|
||||
seed=seed,
|
||||
watermark=watermark,
|
||||
)
|
||||
|
||||
|
||||
class MinimaxHailuo03ReferenceNode(IO.ComfyNode):
|
||||
@classmethod
|
||||
def define_schema(cls):
|
||||
return IO.Schema(
|
||||
node_id="MinimaxHailuo03ReferenceNode",
|
||||
display_name="MiniMax H3 Reference to Video",
|
||||
category="partner/video/MiniMax",
|
||||
description="Generate video conditioned on reference images, videos, and audio using the "
|
||||
"MiniMax H3 model. Refer to the references in the prompt by their order: "
|
||||
"'Image 1', 'Image 2', 'Video 1', 'Audio 1', and so on.",
|
||||
inputs=[
|
||||
IO.DynamicCombo.Input(
|
||||
"model",
|
||||
options=[
|
||||
IO.DynamicCombo.Option(
|
||||
"MiniMax H3",
|
||||
[
|
||||
*_hailuo03_model_inputs(),
|
||||
IO.Autogrow.Input(
|
||||
"reference_images",
|
||||
template=IO.Autogrow.TemplateNames(
|
||||
IO.Image.Input("reference_image"),
|
||||
names=[
|
||||
"image_1",
|
||||
"image_2",
|
||||
"image_3",
|
||||
"image_4",
|
||||
"image_5",
|
||||
"image_6",
|
||||
"image_7",
|
||||
"image_8",
|
||||
"image_9",
|
||||
],
|
||||
min=0,
|
||||
),
|
||||
tooltip="Subject or style reference images, referred to in the prompt "
|
||||
"as 'Image 1'..'Image 9' in connection order. Up to 9 images.",
|
||||
),
|
||||
IO.Autogrow.Input(
|
||||
"reference_videos",
|
||||
template=IO.Autogrow.TemplateNames(
|
||||
IO.Video.Input("reference_video"),
|
||||
names=["video_1", "video_2", "video_3"],
|
||||
min=0,
|
||||
),
|
||||
tooltip="Motion or scene reference videos, referred to in the prompt "
|
||||
"as 'Video 1'..'Video 3' in connection order. Up to 3 videos, "
|
||||
"2-15 seconds each, 15 seconds in total.",
|
||||
),
|
||||
IO.Autogrow.Input(
|
||||
"reference_audios",
|
||||
template=IO.Autogrow.TemplateNames(
|
||||
IO.Audio.Input("reference_audio"),
|
||||
names=["audio_1", "audio_2", "audio_3"],
|
||||
min=0,
|
||||
),
|
||||
tooltip="Audio references, referred to in the prompt as "
|
||||
"'Audio 1'..'Audio 3' in connection order. Up to 3 clips, "
|
||||
"2-15 seconds each, 15 seconds in total. Cannot be used without "
|
||||
"a reference image or video.",
|
||||
),
|
||||
],
|
||||
)
|
||||
],
|
||||
tooltip="Model to use for video generation.",
|
||||
),
|
||||
IO.Int.Input(
|
||||
"seed",
|
||||
default=42,
|
||||
min=0,
|
||||
max=4294967295,
|
||||
step=1,
|
||||
display_mode=IO.NumberDisplay.number,
|
||||
control_after_generate=True,
|
||||
tooltip="Random seed. The same request with the same seed gives similar, "
|
||||
"but not guaranteed identical, results.",
|
||||
),
|
||||
IO.Boolean.Input(
|
||||
"watermark",
|
||||
default=False,
|
||||
tooltip="Whether to add an AIGC watermark to the video.",
|
||||
advanced=True,
|
||||
),
|
||||
],
|
||||
outputs=[
|
||||
IO.Video.Output(),
|
||||
],
|
||||
hidden=[
|
||||
IO.Hidden.auth_token_comfy_org,
|
||||
IO.Hidden.api_key_comfy_org,
|
||||
IO.Hidden.unique_id,
|
||||
],
|
||||
is_api_node=True,
|
||||
price_badge=IO.PriceBadge(
|
||||
depends_on=IO.PriceBadgeDepends(
|
||||
widgets=["model.duration"],
|
||||
input_groups=["model.reference_images", "model.reference_videos"],
|
||||
),
|
||||
expr="""
|
||||
(
|
||||
$dur := $lookup(widgets, "model.duration");
|
||||
$imgsRaw := $lookup(inputGroups, "model.reference_images");
|
||||
$imgs := $imgsRaw ? $imgsRaw : 0;
|
||||
$vidsRaw := $lookup(inputGroups, "model.reference_videos");
|
||||
$vids := $vidsRaw ? $vidsRaw : 0;
|
||||
$base := $dur * 0.1859 + ($imgs > 5 ? ($imgs - 5) * 0.0572 : 0);
|
||||
$vids > 0
|
||||
? {"type": "range_usd", "min_usd": $base + $vids * 2 * 0.1859,
|
||||
"max_usd": $base + 15 * 0.1859, "format": {"approximate": true}}
|
||||
: {"type": "usd", "usd": $base}
|
||||
)
|
||||
""",
|
||||
),
|
||||
)
|
||||
|
||||
@classmethod
|
||||
async def execute(
|
||||
cls,
|
||||
model: dict,
|
||||
seed: int,
|
||||
watermark: bool,
|
||||
) -> IO.NodeOutput:
|
||||
validate_string(model["prompt"], strip_whitespace=True, min_length=1)
|
||||
|
||||
reference_images = model.get("reference_images", {})
|
||||
reference_videos = model.get("reference_videos", {})
|
||||
reference_audios = model.get("reference_audios", {})
|
||||
if not reference_images and not reference_videos:
|
||||
raise ValueError("At least one reference image or video is required.")
|
||||
|
||||
for image in reference_images.values():
|
||||
validate_image_aspect_ratio(image, (2, 5), (5, 2), strict=False) # 0.4 to 2.5
|
||||
validate_image_dimensions(image, min_width=256, min_height=256)
|
||||
|
||||
total_video_duration = 0.0
|
||||
for i, video in enumerate(reference_videos.values(), 1):
|
||||
try:
|
||||
fps = float(video.get_frame_rate())
|
||||
except Exception:
|
||||
fps = 0.0
|
||||
if fps and not (23.9 <= fps <= 60.5):
|
||||
raise ValueError(f"Reference video {i} is {fps:.2f} FPS. Supported range is 23.976-60 FPS.")
|
||||
try:
|
||||
dur = video.get_duration()
|
||||
except Exception:
|
||||
continue
|
||||
if dur < 1.8:
|
||||
raise ValueError(f"Reference video {i} is too short: {dur:.1f}s. Minimum duration is 2 seconds.")
|
||||
total_video_duration += dur
|
||||
if total_video_duration > 15.1:
|
||||
raise ValueError(
|
||||
f"Total reference video duration is {total_video_duration:.1f}s. Maximum is 15 seconds."
|
||||
)
|
||||
|
||||
total_audio_duration = 0.0
|
||||
for i, audio in enumerate(reference_audios.values(), 1):
|
||||
dur = int(audio["waveform"].shape[-1]) / int(audio["sample_rate"])
|
||||
if dur < 1.8:
|
||||
raise ValueError(f"Reference audio {i} is too short: {dur:.1f}s. Minimum duration is 2 seconds.")
|
||||
total_audio_duration += dur
|
||||
if total_audio_duration > 15.1:
|
||||
raise ValueError(
|
||||
f"Total reference audio duration is {total_audio_duration:.1f}s. Maximum is 15 seconds."
|
||||
)
|
||||
|
||||
content: list = [Hailuo03TextContent(text=model["prompt"])]
|
||||
for i, image in enumerate(reference_images.values(), 1):
|
||||
content.append(
|
||||
Hailuo03ImageContent(
|
||||
image_url=Hailuo03ImageContentUrl(
|
||||
url=(
|
||||
await upload_images_to_comfyapi(
|
||||
cls, image, max_images=1, wait_label=f"Uploading image {i}"
|
||||
)
|
||||
)[0],
|
||||
),
|
||||
role="reference_image",
|
||||
)
|
||||
)
|
||||
for i, video in enumerate(reference_videos.values(), 1):
|
||||
content.append(
|
||||
Hailuo03VideoContent(
|
||||
video_url=Hailuo03VideoContentUrl(
|
||||
url=await upload_video_to_comfyapi(cls, video, wait_label=f"Uploading video {i}"),
|
||||
),
|
||||
)
|
||||
)
|
||||
for audio in reference_audios.values():
|
||||
content.append(
|
||||
Hailuo03AudioContent(
|
||||
audio_url=Hailuo03AudioContentUrl(
|
||||
url=await upload_audio_to_comfyapi(
|
||||
cls,
|
||||
audio,
|
||||
container_format="mp3",
|
||||
codec_name="libmp3lame",
|
||||
mime_type="audio/mpeg",
|
||||
),
|
||||
),
|
||||
)
|
||||
)
|
||||
return await _hailuo03_run_task(
|
||||
cls,
|
||||
model_id=HAILUO_03_MODELS[model["model"]],
|
||||
content=content,
|
||||
resolution=model["resolution"],
|
||||
duration=model["duration"],
|
||||
ratio=model["ratio"],
|
||||
seed=seed,
|
||||
watermark=watermark,
|
||||
)
|
||||
|
||||
|
||||
class MinimaxExtension(ComfyExtension):
|
||||
@override
|
||||
async def get_node_list(self) -> list[type[IO.ComfyNode]]:
|
||||
@@ -445,6 +944,9 @@ class MinimaxExtension(ComfyExtension):
|
||||
MinimaxImageToVideoNode,
|
||||
# MinimaxSubjectToVideoNode,
|
||||
MinimaxHailuoVideoNode,
|
||||
MinimaxHailuo03TextToVideoNode,
|
||||
MinimaxHailuo03FirstLastFrameNode,
|
||||
MinimaxHailuo03ReferenceNode,
|
||||
]
|
||||
|
||||
|
||||
|
||||
@@ -1,5 +1,5 @@
|
||||
comfyui-frontend-package==1.47.11
|
||||
comfyui-workflow-templates==0.11.19
|
||||
comfyui-workflow-templates==0.11.20
|
||||
comfyui-embedded-docs==0.5.9
|
||||
torch
|
||||
torchsde
|
||||
|
||||
Reference in New Issue
Block a user