nodes/Xyc2016-Comfyui_Fd_Nodes
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add FD_SeedreamImage

Browse Source
This commit is contained in:
xuyucai
2026-03-02 11:19:51 +08:00
parent a7b976a388
commit 4e7c4eabd2
2 changed files with 176 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/Comfyui_Fd_Nodes/config.py
+9 -1
View File
@@ -15,4 +15,12 @@ FD_FLUX2KLEIN_URL = os.getenv("FD_FLUX2KLEIN_URL")
FD_FLUX2KLEIN_USERNAME = os.getenv("FD_FLUX2KLEIN_USERNAME")
FD_FLUX2KLEIN_PASSWORD = os.getenv("FD_FLUX2KLEIN_PASSWORD")
FD_OSS_URL_PATH_PREFIX_FLUX = os.getenv("FD_OSS_URL_PATH_PREFIX_FLUX", "devops/comfyui/segment_img")
FD_GEN_IMAGE_NOTIFICATION_WEBHOOK_URL = os.getenv("FD_GEN_IMAGE_NOTIFICATION_WEBHOOK_URL")
FD_GEN_IMAGE_NOTIFICATION_WEBHOOK_URL = os.getenv("FD_GEN_IMAGE_NOTIFICATION_WEBHOOK_URL")
FD_LITELLM_BASE_URL = os.getenv("FD_LITELLM_BASE_URL")
FD_LITELLM_API_KEY = os.getenv("FD_LITELLM_API_KEY")
FD_OSS_URL_PATH_PREFIX_BEFORE_GEN = os.getenv("FD_OSS_URL_PATH_PREFIX_BEFORE_GEN", "devops/comfyui/segment_img")
assert FD_LITELLM_BASE_URL, "FD_LITELLM_BASE_URL is not set"
assert FD_LITELLM_API_KEY, "FD_LITELLM_API_KEY is not set"
src/Comfyui_Fd_Nodes/nodes.py
+167
View File
@@ -22,11 +22,14 @@ from .config import (
FD_FLUX2KLEIN_URL,
FD_FLUX2KLEIN_USERNAME,
FD_GEN_IMAGE_NOTIFICATION_WEBHOOK_URL,
FD_LITELLM_BASE_URL,
FD_LITELLM_API_KEY,
FD_OSS_ACCESS_KEY_ID,
FD_OSS_ACCESS_KEY_SECRET,
FD_OSS_BUCKET_NAME,
FD_OSS_ENDPOINT,
FD_OSS_URL_PATH_PREFIX_FLUX,
FD_OSS_URL_PATH_PREFIX_BEFORE_GEN,
FD_OSS_URL_PREFIX,
)
from .old_fd_nodes import FD_imgToText_Doubao, FD_Upload
@@ -364,6 +367,168 @@ class FD_Flux2KleinGenImage(ComfyNodeABC):
return (output_image,)
class FD_SeedreamImage(ComfyNodeABC):
"""
Node to generate images using Seedream 5.0 Lite model.
"""
def __init__(self):
auth = oss2.Auth(FD_OSS_ACCESS_KEY_ID, FD_OSS_ACCESS_KEY_SECRET)
self.bucket = oss2.Bucket(
auth=auth,
bucket_name=FD_OSS_BUCKET_NAME,
endpoint=FD_OSS_ENDPOINT,
connect_timeout=30
)
self.oss_url_prefix = FD_OSS_URL_PREFIX
@classmethod
def INPUT_TYPES(cls) -> InputTypeDict:
return {
"required": {
"prompt": (
IO.STRING,
{
"multiline": True,
"default": "",
"tooltip": "Text prompt for generation",
},
),
"model": (
["doubao-seedream-5.0-lite"],
{
"default": "doubao-seedream-5.0-lite",
"tooltip": "Model to use for generation",
},
),
"size": (
["1K", "2K"],
{
"default": "2K",
"tooltip": "Output image size",
},
),
},
"optional": {
"images": (
IO.IMAGE,
{
"default": None,
"tooltip": "Optional image(s) to use as context. Multiple images supported.",
},
),
"output_format": (
["png", "jpg"],
{
"default": "png",
"tooltip": "Output image format",
},
),
},
}
RETURN_TYPES = (IO.IMAGE, )
FUNCTION = "api_call"
CATEGORY = "image/generation"
DESCRIPTION = "Generate images using Seedream 5.0 Lite model."
API_NODE = True
def api_call(
self,
prompt: str,
model: str,
size: str,
images: Optional[IO.IMAGE] = None,
output_format: str = "png",
**kwargs,
):
body = {
"model": model,
"prompt": prompt,
"sequential_image_generation": "disabled",
"size": size,
"output_format": output_format,
"watermark": False,
}
# Upload images to OSS if provided
if images is not None:
batch_size = images.shape[0]
image_url_list = []
for i in range(batch_size):
single_image = images[i : i + 1]
# original_image = single_image.squeeze()
# scaled_image = downscale_image_tensor(single_image, total_pixels=2048 * 2048).squeeze()
# logger.info(
# "FD_SeedreamImage Image %s resolution: original=%s scaled=%s",
# i,
# tuple(original_image.shape),
# tuple(scaled_image.shape),
# )
scaled_image = single_image.squeeze() # 现在是测试阶段，先不要downscale TODO: 到时候考虑改一下
image_np = (scaled_image.numpy() * 255).astype(np.uint8)
img = Image.fromarray(image_np)
img_byte_arr = BytesIO()
img.save(img_byte_arr, format="PNG")
img_byte_arr = img_byte_arr.getvalue()
file_oss_path = f"{FD_OSS_URL_PATH_PREFIX_BEFORE_GEN}/{bytes_calculate_hex_md5(img_byte_arr)}.png"
self.bucket.put_object(file_oss_path, img_byte_arr)
print(f"upload {file_oss_path}")
oss_file_url = f"{self.oss_url_prefix}{file_oss_path}"
image_url_list.append(oss_file_url)
body['image'] = image_url_list
if FD_GEN_IMAGE_NOTIFICATION_WEBHOOK_URL:
try:
print("Sending seedream webhook message...")
webhook_send(FD_GEN_IMAGE_NOTIFICATION_WEBHOOK_URL, {
"seedream_request": body,
})
except Exception:
pass
logger.info(f"Calling Seedream API with {body}")
# Call API
headers = {
"Authorization": f"Bearer {FD_LITELLM_API_KEY}",
"Content-Type": "application/json",
}
response = requests.post(
url=f"{FD_LITELLM_BASE_URL}/v1/images/generations",
headers=headers,
json=body,
timeout=300,
)
response.raise_for_status()
if response.status_code != 200:
raise Exception(f"Failed to call API: {response.content}")
result = response.json()
logger.info(f"Seedream API response: {result}")
if FD_GEN_IMAGE_NOTIFICATION_WEBHOOK_URL:
try:
print("Sending seedream webhook message...")
webhook_send(FD_GEN_IMAGE_NOTIFICATION_WEBHOOK_URL, {
"seedream_full": {
"request": body,
"response": result,
}
})
except Exception:
pass
# Get result image URL
result_url = result["data"][0]["url"]
image_content = requests.get(result_url).content
image_bytesio = BytesIO(image_content)
output_image = bytesio_to_image_tensor(image_bytesio)
return (output_image,)
class Example:
"""
A example node
@@ -479,6 +644,7 @@ NODE_CLASS_MAPPINGS = {
"FD_imgToText_Doubao": FD_imgToText_Doubao,
"FD_GeminiImage": FD_GeminiImage,
"FD_Flux2KleinGenImage": FD_Flux2KleinGenImage,
"FD_SeedreamImage": FD_SeedreamImage,
}
# A dictionary that contains the friendly/humanly readable titles for the nodes
@@ -488,4 +654,5 @@ NODE_DISPLAY_NAME_MAPPINGS = {
"FD_imgToText_Doubao": "FD Image to Text (Doubao)",
"FD_GeminiImage": "FD Gemini Image",
"FD_Flux2KleinGenImage": "FD Flux2Klein Gen Image",
"FD_SeedreamImage": "FD Seedream Image",
}