nodes/HM-RunningHub-ComfyUI_RH_APICall
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Supports direct processing of video output and can simultaneously handle workflows with both images and videos.

Browse Source
This commit is contained in:
unknown
2025-02-06 23:09:46 +08:00
parent 6338b51b76
commit dd76306148
3 changed files with 584 additions and 273 deletions
Download Patch File Download Diff File Expand all files Collapse all files
RH_ExecuteNode.py
+111 -126
View File
@@ -1,11 +1,11 @@
import requests
import time
import json
from PIL import Image, ImageOps
from PIL import Image
from io import BytesIO
import numpy as np
import torch
import os # 确保导入 os 库
import os
class ExecuteNode:
def __init__(self):
@@ -16,30 +16,26 @@ class ExecuteNode:
return {
"required": {
"apiConfig": ("STRUCT",), # 设置节点的输入
"nodeInfoList": ("ARRAY", {"default": []}), # NodeInfoList节点的输出
},
"optional": {
"nodeInfoList": ("ARRAY", {"default": []}), # NodeInfoList节点的输出, 设置为可选
"run_timeout": ("INT", {"default": 600}), # 最大运行超时时间（秒）
"query_interval": ("INT", {"default": 10}), # 查询间隔时间（秒）
},
}
RETURN_TYPES = ("IMAGE",) # 支持单张和多张图片输出
RETURN_NAMES = ("images",) # 定义返回名称，与 ComfyUI 的预期匹配
RETURN_TYPES = ("IMAGE", "VIDEO") # 支持单张和多张图片输出，也支持视频
RETURN_NAMES = ("images", "videos") # 定义返回名称，与 ComfyUI 的预期匹配
CATEGORY = "RunningHub"
FUNCTION = "process" # 指向 process 方法
def process(self, apiConfig, nodeInfoList, run_timeout=600, query_interval=2):
def process(self, apiConfig, nodeInfoList=None, run_timeout=600, query_interval=2):
# Ensure query_interval is not less than 2
if query_interval < 2:
print("Query interval is too low, setting to minimum value of 2 seconds.")
query_interval = 2
"""
该节点通过调用 RunningHub API 创建任务并返回生成的图片链接。
"""
# 打印请求数据，方便调试
print(f"API request data: {apiConfig}")
print(f"Node info list: {nodeInfoList}")
print(f"Run timeout: {run_timeout} seconds")
@@ -49,19 +45,17 @@ class ExecuteNode:
account_status = self.check_account_status(apiConfig["apiKey"], apiConfig["base_url"])
if int(account_status["currentTaskCounts"]) > 0:
print("There are tasks running, waiting for them to finish.")
# 等待最多 run_timeout 秒，如果任务未完成，则超时
start_time = time.time()
while account_status["currentTaskCounts"] > 0 and time.time() - start_time < run_timeout:
time.sleep(query_interval) # 每 query_interval 秒查询一次
time.sleep(query_interval)
account_status = self.check_account_status(apiConfig["apiKey"], apiConfig["base_url"])
if int(account_status["currentTaskCounts"]) > 0:
raise Exception(f"Timeout: There are still running tasks after {run_timeout} seconds.")
# Print nodeInfoList for debugging
print(f"ExecuteNode NodeInfoList: {nodeInfoList}")
# 2. 创建任务
task_creation_result = self.create_task(apiConfig, nodeInfoList, apiConfig["base_url"])
# 2. 创建任务，如果 nodeInfoList 为空则不传递
task_creation_result = self.create_task(apiConfig, nodeInfoList or [], apiConfig["base_url"])
if task_creation_result["code"] != 0:
raise Exception(f"Task creation failed: {task_creation_result['msg']}")
@@ -73,14 +67,13 @@ class ExecuteNode:
task_start_time = time.time()
while task_status != "success":
print(f"Task still running, checking again in {query_interval} seconds...")
time.sleep(query_interval) # 每 query_interval 秒检查一次任务状态
time.sleep(query_interval)
task_status_result = self.check_task_status(task_id, apiConfig["apiKey"], apiConfig["base_url"])
print(f"Task info, taskId: {task_id}, status: {task_status_result}")
if isinstance(task_status_result, dict):
task_status = task_status_result.get("taskStatus", "unknown") # 从结果中获取任务状态
task_status = task_status_result.get("taskStatus", "unknown")
elif isinstance(task_status_result, list):
# 假设任务完成后返回的数据列表表示任务成功
task_status = "success"
else:
task_status = "unknown"
@@ -89,13 +82,105 @@ class ExecuteNode:
print(f"Task failed or completed with status: {task_status}")
break
# 检查是否超过 run_timeout
if time.time() - task_start_time > run_timeout:
raise Exception(f"Timeout: Task {task_id} did not complete within {run_timeout} seconds.")
# 4. 任务完成，处理输出
return self.process_task_output(task_id, apiConfig["apiKey"], apiConfig["base_url"])
def process_task_output(self, task_id, api_key, base_url):
"""
处理任务输出，返回文件链接。
返回图像和视频，若无图像则返回空图像，若无视频则返回空视频。
"""
task_status_result = self.check_task_status(task_id, api_key, base_url)
print("Task Status Result:", json.dumps(task_status_result, indent=4, ensure_ascii=False))
image_urls = []
video_urls = []
# 检查任务是否返回了图像
if isinstance(task_status_result, dict):
file_url = task_status_result.get("fileUrl")
file_type = task_status_result.get("fileType")
if file_url and file_type.lower() in ["png", "jpg", "jpeg"]:
image_urls.append(file_url)
elif file_url and file_type.lower() in ["mp4", "avi", "mov"]:
video_urls.append(file_url)
elif isinstance(task_status_result, list):
for output in task_status_result:
if isinstance(output, dict):
file_url = output.get("fileUrl")
file_type = output.get("fileType")
if file_url and file_type.lower() in ["png", "jpg", "jpeg"]:
image_urls.append(file_url)
elif file_url and file_type.lower() in ["mp4", "avi", "mov"]:
video_urls.append(file_url)
if not image_urls:
image_urls = None # No images found, set to None
if not video_urls:
video_urls = None # No videos found, set to None
# Download and process images if available
image_data_list = []
if image_urls:
for url in image_urls:
print("Downloading image from URL:", url)
image_tensor = self.download_image(url)
image_data_list.append(image_tensor)
# 下载视频并保存在本地
video_data_list = []
if video_urls:
for url in video_urls:
print("Downloading video from URL:", url)
video_path = self.download_video(url)
video_data_list.append(video_path)
print(f"Returning {len(image_data_list)} images and {len(video_data_list)} videos.")
return (image_data_list, video_data_list)
def download_image(self, image_url):
"""
从 URL 下载图像并转换为适合预览或保存的 torch.Tensor 格式。
"""
response = requests.get(image_url)
if response.status_code == 200:
img = Image.open(BytesIO(response.content)).convert("RGB")
img_array = np.array(img).astype(np.float32) / 255.0
img_tensor = torch.from_numpy(img_array)
return img_tensor
else:
raise Exception(f"Failed to download image: {image_url}")
def download_video(self, video_url):
"""
从 URL 下载视频并保存到本地。
"""
response = requests.get(video_url, stream=True)
if response.status_code == 200:
# 确保保存视频的目录存在
output_dir = "output"
if not os.path.exists(output_dir):
os.makedirs(output_dir)
# 获取视频文件名并保存
video_filename = f"RH_output_video_{str(int(time.time()))}.mp4"
video_path = os.path.join(output_dir, video_filename)
with open(video_path, "wb") as f:
for chunk in response.iter_content(chunk_size=1024):
if chunk:
f.write(chunk)
print(f"Video saved to {video_path}")
return video_path
else:
raise Exception(f"Failed to download video: {video_url}")
def check_account_status(self, api_key, base_url):
"""
查询账户状态，检查是否可以提交新任务
@@ -113,7 +198,6 @@ class ExecuteNode:
result = response.json()
if result["code"] != 0:
raise Exception(f"Failed to get account status: {result['msg']}")
# 检查并确保 currentTaskCounts 是整数
try:
current_task_counts = int(result["data"]["currentTaskCounts"])
except ValueError:
@@ -134,14 +218,7 @@ class ExecuteNode:
data = {
"workflowId": apiConfig["workflowId"],
"apiKey": apiConfig["apiKey"],
"nodeInfoList": [
{
"nodeId": int(nodeInfo["nodeId"]), # 确保 nodeId 为整数类型
"fieldName": nodeInfo["fieldName"],
"fieldValue": nodeInfo["fieldValue"],
}
for nodeInfo in nodeInfoList
],
"nodeInfoList": nodeInfoList, # 如果nodeInfoList为空，传递空列表
}
response = requests.post(url, json=data, headers=headers)
@@ -163,7 +240,6 @@ class ExecuteNode:
response = requests.post(url, json=data, headers=headers)
# 打印响应以便调试
print("Response Status Code:", response.status_code)
try:
response_json = response.json()
@@ -176,114 +252,23 @@ class ExecuteNode:
result = response.json()
# 检查是否有数据结果
if result.get("data") and isinstance(result["data"], list):
if len(result["data"]) > 0:
return result["data"] # 返回整个列表，表示任务成功并且有结果
return result["data"]
else:
# data 是空列表，可能是任务还在运行，或出错
if result.get("code") != 0:
msg = result.get("msg")
if msg != "APIKEY_TASK_IS_RUNNING":
return {"error": msg} # 出现了错误，返回错误信息
return {"error": msg}
else:
return {"taskStatus": "RUNNING"} # 任务仍在运行
return {"taskStatus": "RUNNING"} # 如果没有其他信息，认为任务在运行中
return {"taskStatus": "RUNNING"}
return {"taskStatus": "RUNNING"}
# 如果没有 data 字段或 data 不是列表，检查任务状态
if result.get("code") != 0:
msg = result.get("msg")
if msg != "APIKEY_TASK_IS_RUNNING":
return {"error": msg} # 返回错误信息
return {"error": msg}
else:
return {"taskStatus": "RUNNING"} # 任务仍在运行
return {"taskStatus": "RUNNING"}
return {"taskStatus": "UNKNOWN_ERROR"}
def process_task_output(self, task_id, api_key, base_url):
"""
处理任务输出，返回文件链接。
"""
task_status_result = self.check_task_status(task_id, api_key, base_url)
# 记录任务状态结果以了解其结构
print("Task Status Result:", json.dumps(task_status_result, indent=4, ensure_ascii=False))
image_urls = []
# 确保 task_status_result 是字典或列表
if isinstance(task_status_result, dict):
# 检查 fileUrl 和 fileType
file_url = task_status_result.get("fileUrl")
file_type = task_status_result.get("fileType")
if file_url and file_type.lower() in ["png", "jpg", "jpeg"]:
image_urls.append(file_url)
elif isinstance(task_status_result, list):
for output in task_status_result:
if isinstance(output, dict):
file_url = output.get("fileUrl")
file_type = output.get("fileType")
if file_url and file_type.lower() in ["png", "jpg", "jpeg"]:
image_urls.append(file_url)
if not image_urls:
raise Exception("No valid image output found.")
# 下载并处理所有图像
image_data_list = []
for url in image_urls:
print("Downloading image from URL:", url) # 记录图像 URL
image_tensor = self.download_image(url) # 下载并处理图像
print("Image downloaded and processed successfully.")
# 打印张量信息，避免在 uint8 上调用 mean()
print(f"Image tensor shape: {image_tensor.shape}")
mean_val = image_tensor.mean()
print(f"Image tensor min: {image_tensor.min()}, max: {image_tensor.max()}, mean: {mean_val}")
# 确保返回的是 torch.Tensor
if not isinstance(image_tensor, torch.Tensor):
raise TypeError(f"Expected torch.Tensor, got {type(image_tensor)}")
image_data_list.append(image_tensor)
print(f"Returning {len(image_data_list)} images.")
return (image_data_list,) # 返回一个包含图像列表的元组，与 RETURN_TYPES 和 RETURN_NAMES 匹配
def download_image(self, image_url):
"""
从 URL 下载图像并转换为适合预览或保存的 torch.Tensor 格式。
"""
response = requests.get(image_url)
if response.status_code == 200:
img = Image.open(BytesIO(response.content)).convert("RGB")
# 转换为 numpy 数组并调整数据类型和范围
img_array = np.array(img).astype(np.float32) / 255.0 # 归一化到 [0, 1]
# **保持形状为 [H, W, C]，不进行 permute 操作**
img_tensor = torch.from_numpy(img_array) # 形状 [H, W, C]
# 打印图像尺寸
print(f"Downloaded image dimensions: {img_tensor.shape}") # 例如 (高度, 宽度, 3)
return img_tensor
else:
raise Exception(f"Failed to download image: {image_url}")
def download_video(self, video_url):
"""
从 URL 下载视频。
根据 ComfyUI 的要求实现此方法。
"""
response = requests.get(video_url, stream=True)
if response.status_code == 200:
# 示例：将视频保存到临时位置并返回路径或数据
video_content = response.content
# 您可能需要根据 ComfyUI 的要求处理视频数据
# 目前，返回原始字节
return video_content
else:
raise Exception(f"Failed to download video: {video_url}")
examples/hunyuanTextToVideo.json
+253 -147
View File
@@ -1,61 +1,7 @@
{
"last_node_id": 42,
"last_link_id": 32,
"last_node_id": 44,
"last_link_id": 34,
"nodes": [
{
"id": 17,
"type": "RH_NodeInfoListNode",
"pos": [
68,
1440
],
"size": {
"0": 330,
"1": 106
},
"flags": {},
"order": 2,
"mode": 0,
"inputs": [
{
"name": "previousNodeInfoList",
"type": "ARRAY",
"link": null,
"label": "previousNodeInfoList",
"shape": 7
},
{
"name": "fieldValue",
"type": "STRING",
"link": 32,
"widget": {
"name": "fieldValue"
},
"label": "fieldValue",
"slot_index": 1
}
],
"outputs": [
{
"name": "ARRAY",
"type": "ARRAY",
"links": [
23
],
"slot_index": 0,
"shape": 3,
"label": "ARRAY"
}
],
"properties": {
"Node name for S&R": "RH_NodeInfoListNode"
},
"widgets_values": [
90,
"text",
"土星环"
]
},
{
"id": 42,
"type": "String",
@@ -88,96 +34,12 @@
"wh32，一个采光很好的房间，中午，女子穿着蕾丝睡衣在床上躺着，正面对着镜头漏出了微笑"
]
},
{
"id": 30,
"type": "RH_ExecuteNode",
"pos": [
526,
1477
],
"size": {
"0": 315,
"1": 102
},
"flags": {},
"order": 3,
"mode": 0,
"inputs": [
{
"name": "apiConfig",
"type": "STRUCT",
"link": 20,
"label": "apiConfig"
},
{
"name": "nodeInfoList",
"type": "ARRAY",
"link": 23,
"label": "nodeInfoList"
}
],
"outputs": [
{
"name": "images",
"type": "IMAGE",
"links": [
30
],
"slot_index": 0,
"shape": 3,
"label": "images"
}
],
"properties": {
"Node name for S&R": "RH_ExecuteNode"
},
"widgets_values": [
600,
2
]
},
{
"id": 9,
"type": "RH_SettingsNode",
"pos": [
53,
1227
],
"size": [
502.2504859687508,
137.38932101438627
],
"flags": {},
"order": 1,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "STRUCT",
"type": "STRUCT",
"links": [
20
],
"slot_index": 0,
"shape": 3,
"label": "STRUCT"
}
],
"properties": {
"Node name for S&R": "RH_SettingsNode"
},
"widgets_values": [
"https://www.runninghub.cn",
"ed37fbdd79c34a7ca612aedbe5cea13e",
"1871595400514633730"
]
},
{
"id": 39,
"type": "VHS_VideoCombine",
"pos": [
887,
1224
1395,
1231
],
"size": [
214.7587890625,
@@ -239,16 +101,244 @@
"hidden": false,
"paused": false,
"params": {
"filename": "AnimateDiff_00002.mp4",
"filename": "AnimateDiff_00005.mp4",
"subfolder": "",
"type": "output",
"format": "video/h264-mp4",
"frame_rate": 24,
"workflow": "AnimateDiff_00002.png"
"workflow": "AnimateDiff_00005.png"
},
"muted": false
}
}
},
{
"id": 9,
"type": "RH_SettingsNode",
"pos": [
53,
1227
],
"size": {
"0": 502.25048828125,
"1": 137.3893280029297
},
"flags": {},
"order": 1,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "STRUCT",
"type": "STRUCT",
"links": [
20
],
"slot_index": 0,
"shape": 3,
"label": "STRUCT"
}
],
"properties": {
"Node name for S&R": "RH_SettingsNode"
},
"widgets_values": [
"https://www.runninghub.cn",
"ed37fbdd79c34a7ca612aedbe5cea13e",
"1871595400514633730"
]
},
{
"id": 30,
"type": "RH_ExecuteNode",
"pos": [
550,
1468
],
"size": {
"0": 315,
"1": 102
},
"flags": {},
"order": 3,
"mode": 0,
"inputs": [
{
"name": "apiConfig",
"type": "STRUCT",
"link": 20,
"label": "apiConfig"
},
{
"name": "nodeInfoList",
"type": "ARRAY",
"link": 23,
"label": "nodeInfoList"
}
],
"outputs": [
{
"name": "images",
"type": "IMAGE",
"links": [
30
],
"slot_index": 0,
"shape": 3,
"label": "images"
},
{
"name": "videos",
"type": "VIDEO",
"links": [
33
],
"shape": 3,
"label": "videos",
"slot_index": 1
}
],
"properties": {
"Node name for S&R": "RH_ExecuteNode"
},
"widgets_values": [
600,
5
]
},
{
"id": 17,
"type": "RH_NodeInfoListNode",
"pos": [
82,
1435
],
"size": {
"0": 330,
"1": 106
},
"flags": {},
"order": 2,
"mode": 0,
"inputs": [
{
"name": "previousNodeInfoList",
"type": "ARRAY",
"link": null,
"label": "previousNodeInfoList",
"shape": 7
},
{
"name": "fieldValue",
"type": "STRING",
"link": 32,
"widget": {
"name": "fieldValue"
},
"label": "fieldValue",
"slot_index": 1
}
],
"outputs": [
{
"name": "ARRAY",
"type": "ARRAY",
"links": [
23
],
"slot_index": 0,
"shape": 3,
"label": "ARRAY"
}
],
"properties": {
"Node name for S&R": "RH_NodeInfoListNode"
},
"widgets_values": [
90,
"text",
"土星环"
]
},
{
"id": 43,
"type": "easy showAnythingLazy",
"pos": [
943,
1509
],
"size": {
"0": 210,
"1": 26
},
"flags": {},
"order": 5,
"mode": 0,
"inputs": [
{
"name": "anything",
"type": "*",
"link": 33,
"label": "anything"
}
],
"outputs": [
{
"name": "output",
"type": "*",
"links": [
34
],
"shape": 3,
"label": "output",
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "easy showAnythingLazy"
}
},
{
"id": 44,
"type": "ShowText|pysssss",
"pos": [
928,
1639
],
"size": [
315,
76
],
"flags": {},
"order": 6,
"mode": 0,
"inputs": [
{
"name": "text",
"type": "STRING",
"link": 34,
"widget": {
"name": "text"
},
"label": "文本"
}
],
"outputs": [
{
"name": "STRING",
"type": "STRING",
"links": null,
"shape": 6,
"label": "字符串"
}
],
"properties": {
"Node name for S&R": "ShowText|pysssss"
},
"widgets_values": [
"",
"output\\RH_output_video_1738854232.mp4"
]
}
],
"links": [
@@ -283,16 +373,32 @@
17,
1,
"STRING"
],
[
33,
30,
1,
43,
0,
"*"
],
[
34,
43,
0,
44,
0,
"STRING"
]
],
"groups": [],
"config": {},
"extra": {
"ds": {
"scale": 0.8769226950000008,
"scale": 0.9646149645000006,
"offset": [
589.8700344728978,
-719.2667377838786
197.9158053009368,
-912.3150194555913
]
},
"ue_links": [],
examples/rh_video_save_to_local_direct.json
+220
View File
@@ -0,0 +1,220 @@
{
"last_node_id": 53,
"last_link_id": 42,
"nodes": [
{
"id": 9,
"type": "RH_SettingsNode",
"pos": [
66,
1357
],
"size": {
"0": 502.25048828125,
"1": 137.3893280029297
},
"flags": {},
"order": 0,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "STRUCT",
"type": "STRUCT",
"links": [
33
],
"slot_index": 0,
"shape": 3,
"label": "STRUCT"
}
],
"properties": {
"Node name for S&R": "RH_SettingsNode"
},
"widgets_values": [
"https://www.runninghub.cn",
"ed37fbdd79c34a7ca612aedbe5cea13e",
"1887510793653780481"
]
},
{
"id": 43,
"type": "RH_ExecuteNode",
"pos": [
635,
1367
],
"size": {
"0": 315,
"1": 102
},
"flags": {},
"order": 1,
"mode": 0,
"inputs": [
{
"name": "apiConfig",
"type": "STRUCT",
"link": 33,
"label": "apiConfig"
},
{
"name": "nodeInfoList",
"type": "ARRAY",
"link": null,
"label": "nodeInfoList",
"slot_index": 1
}
],
"outputs": [
{
"name": "images",
"type": "IMAGE",
"links": [],
"shape": 3,
"label": "images",
"slot_index": 0
},
{
"name": "videos",
"type": "VIDEO",
"links": [
36
],
"shape": 3,
"label": "videos",
"slot_index": 1
}
],
"properties": {
"Node name for S&R": "RH_ExecuteNode"
},
"widgets_values": [
600,
10
]
},
{
"id": 44,
"type": "easy showAnythingLazy",
"pos": [
992,
1373
],
"size": {
"0": 210,
"1": 26
},
"flags": {},
"order": 2,
"mode": 0,
"inputs": [
{
"name": "anything",
"type": "*",
"link": 36,
"label": "anything"
}
],
"outputs": [
{
"name": "output",
"type": "*",
"links": [
39
],
"shape": 3,
"label": "output",
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "easy showAnythingLazy"
}
},
{
"id": 47,
"type": "ShowText|pysssss",
"pos": [
584,
1572
],
"size": {
"0": 347.5731201171875,
"1": 242.42494201660156
},
"flags": {},
"order": 3,
"mode": 0,
"inputs": [
{
"name": "text",
"type": "STRING",
"link": 39,
"widget": {
"name": "text"
},
"label": "文本"
}
],
"outputs": [
{
"name": "STRING",
"type": "STRING",
"links": null,
"shape": 6,
"label": "字符串"
}
],
"properties": {
"Node name for S&R": "ShowText|pysssss"
},
"widgets_values": [
"",
"output\\RH_output_video_1738853689.mp4"
]
}
],
"links": [
[
33,
9,
0,
43,
0,
"STRUCT"
],
[
36,
43,
1,
44,
0,
"*"
],
[
39,
44,
0,
47,
0,
"STRING"
]
],
"groups": [],
"config": {},
"extra": {
"ds": {
"scale": 1.0610764609500007,
"offset": [
143.41590852037982,
-895.2557214802658
]
},
"ue_links": [],
"VHS_latentpreview": false,
"VHS_latentpreviewrate": 0
},
"version": 0.4
}