nodes/TashaSkyUp-EternalKernelPytorchNodes
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
Files
5b8c2bfebab3f39ccacdbf0b95dfce7d6fc5590a
TashaSkyUp-EternalKernelPyt…/video.py
T

3322 lines
126 KiB
Python
Raw Blame History

### info for code completion AI ###
"""
all of these classes are plugins for comfyui and follow the same pattern
all of the images are torch tensors and it is unknown and unimportant if they are on the cpu or gpu
all image inputs are (B,W,H,C)
avoid numpy and PIL as much as possible
"""
import shutil
import fractions
import os
from abc import ABC, ABCMeta, abstractmethod
from os.path import splitext, join
from typing import List, Tuple, Union
import cv2
import torch
import torchaudio
# from cv2 import VideoWriter, VideoWriter_fourcc, VideoCapture
# from line_profiler import profile
from torch import dtype
# from custom_nodes.EternalKernelLiteGraphNodes.image import torch_image_show
from torchvision.models.optical_flow import raft_large, Raft_Large_Weights, raft_small, Raft_Small_Weights
import os
import torch
from torchvision.utils import draw_keypoints
from skimage.metrics import structural_similarity as ssim
import cv2
import numpy as np
import cv2
import subprocess
import imageio_ffmpeg as ffmpeg
# from comfy_extras.nodes_video_model import SVD_img2vid_Conditioning
from . import NODE_CLASS_MAPPINGS, NODE_DISPLAY_NAME_MAPPINGS
# NODE_CLASS_MAPPINGS = {} # this is the dictionary that will be used to register the nodes
# setup directories
root_p = os.path.dirname(os.path.dirname(os.path.dirname(os.path.realpath(__file__))))
# input_dir = os.path.join(root_p, 'input', 'video')
# output_dir = os.path.join(root_p, 'output', 'video')
video_dir = os.path.join(root_p, 'video')
if not os.path.exists(video_dir):
os.makedirs(video_dir)
# video_folders = lambda s: sorted(
# [name for name in os.listdir(s.input_dir) if os.path.isdir(os.path.join(s.input_dir, name))])
# turn the above in the a full function
def video_folders(s):
ret = sorted(
[name for name in os.listdir(s.input_dir) if os.path.isdir(os.path.join(s.input_dir, name))])
return ret
def video_files(s):
# any files in in_video_folders ending in .mp4
ret = sorted([name for name in os.listdir(s.input_dir) if
os.path.isfile(os.path.join(s.input_dir, name)) and name.endswith('.mp4')])
return ret
def run_ffmpeg_command(args):
"""
Executes a command using the ffmpeg executable.
Parameters:
args (list): List of command-line arguments for ffmpeg.
"""
if isinstance(args, str):
args = args.split()
cmd = [ffmpeg.get_ffmpeg_exe()] + args
subprocess.run(cmd)
# start with an ABC to define the common widget interface
class ABCWidgetMetaclass(ABCMeta):
"""A metaclass that automatically registers classes."""
def __init__(cls, name, bases, attrs):
if (ABC in bases) or (ABCVideoWidget in bases) or ("ABC" in name):
pass
else:
NODE_CLASS_MAPPINGS[str(name)] = cls
cls.CATEGORY = "ETK/video"
super().__init__(name, bases, attrs)
class ABCVideoWidget(ABC, metaclass=ABCWidgetMetaclass):
"""Abstract base class for simple construction"""
input_dir = video_dir
output_dir = video_dir
CATEGORY = "video"
RETURN_TYPES = ("VIDEO",)
FUNCTION = "handler"
@classmethod
@abstractmethod
def INPUT_TYPES(cls):
"""All subclasses must provide a INPUT_TYPES classmethod"""
pass
@abstractmethod
def handler(self, one: object, two: object):
"""All subclasses must provide a handler method"""
pass
class ABCVideoFileToABCVideoFolder(ABCVideoWidget, metaclass=ABCWidgetMetaclass):
"""Abstract base class for simple construction"""
@classmethod
def INPUT_TYPES(cls):
return {"required":
{
"video_in": (video_files(cls),),
"folder_out": (video_folders(cls),),
},
"optional":
{"text":
("STRING", {"multiline": False}),
}
}
@abstractmethod
def handler(self, one, two, text):
return (one, two,)
class ABCABCVideoFolderToImage(ABCVideoWidget, metaclass=ABCWidgetMetaclass):
"""Abstract base class for simple construction"""
RETURN_TYPES = ("IMAGE",)
@classmethod
def INPUT_TYPES(cls):
return {"required":
{
"folder_in": ("STRING", {"default": "full_path_to_folder"}),
"idx_slice_start": ("INT", {"min": 0, "max": 99999, "step": 1, "default": 0}),
"idx_slice_stop": ("INT", {"min": 0, "max": 99999, "step": 1, "default": 1}),
"slice_idx_step": ("INT", {"min": 0, "max": 9999, "step": 1, "default": 1}),
"idx": ("INT", {"min": 0, "max": 99999, "step": 1, "default": 0}),
},
"optional":
{"text":
("STRING", {"multiline": False}),
}
}
@abstractmethod
def handler(self, one, two, text):
return (one,)
class ABCABCVideoFileToImage(ABCVideoWidget, metaclass=ABCWidgetMetaclass):
"""Abstract base class for simple construction"""
RETURN_TYPES = ("IMAGE",)
@classmethod
def INPUT_TYPES(cls):
return {"required":
{
"video_in": ("STRING", {"default": "full_path_to_video"}),
"idx_slice_start": ("INT", {"min": 0, "max": 99999, "step": 1, "default": 0}),
"idx_slice_stop": ("INT", {"min": 0, "max": 99999, "step": 1, "default": 1}),
"slice_idx_step": ("INT", {"min": 0, "max": 9999, "step": 1, "default": 1}),
"idx": ("INT", {"min": 0, "max": 99999, "step": 1, "default": 0}),
"func_only": ([False, True], {"default": False}),
},
}
@abstractmethod
def handler(self, one, two, text):
return (one,)
## end of abstract base classes ##
class VideoToFramesFolder(ABCVideoFileToABCVideoFolder, metaclass=ABCWidgetMetaclass):
"""use cv2 to open the video and use it as a source for image stacks"""
RETURN_TYPES = ("IMAGE",)
def __init__(self):
super().__init__()
self.old = self.INPUT_TYPES
# self.INPUT_TYPES = lambda s: self.old().update({})
def handler(self, video_in, folder_out, text):
import time
start = time.time()
video_in_path = os.path.join(self.input_dir, video_in)
video_out_path = os.path.join(self.output_dir, folder_out)
vidcap = cv2.VideoCapture(video_in_path)
success, image = vidcap.read()
count = 0
while success:
cv2.imwrite(os.path.join(video_out_path, "frame%d.jpg" % count), image)
success, image = vidcap.read()
count += 1
vidcap.release()
end = time.time()
print("time elapsed: ", end - start)
return (folder_out,)
class VideoToFramesFolderFFMPEG(ABCVideoFileToABCVideoFolder, metaclass=ABCWidgetMetaclass):
"""use ffmpeg to open the video and use it as a source for image stacks, use PNG or jpeg with quality 100"""
@classmethod
def INPUT_TYPES(cls):
return {"required":
{
"video_in": ("STRING", {"default": "full_path_to_folder"}),
"folder_out": ("STRING", {"default": "full_path_to_folder"}),
"format": (["PNG", "JPEG"], {"default": "PNG"}),
},
}
RETURN_TYPES = ("STRING",)
RETURN_NAMES = ("folder_out",)
def __init__(self):
super().__init__()
self.old = self.INPUT_TYPES
# self.INPUT_TYPES = lambda s: self.old().update({})
def handler(self, video_in, folder_out, format):
import time
start = time.time()
video_in_path = os.path.join(self.input_dir, video_in)
video_out_path = os.path.join(self.output_dir, folder_out)
if not os.path.exists(video_out_path):
os.makedirs(video_out_path)
# specifically this calls ffmpeg with the following arguments:
# -i means input file
# -qscale:v means quality scale video, default is 2, which is a good compromise between size and quality
# 1 is the quality scale value, which is a value between 1 and 31, where 1 is the highest quality
start = time.time()
if format == "JPEG":
subprocess.call(
["ffmpeg", "-i", video_in_path, "-qscale:v", "1", os.path.join(video_out_path, "frame%05d.jpg")])
end = time.time()
print("ffmpeg jpg process took {} seconds".format(end - start))
elif format == "PNG":
subprocess.call(["ffmpeg", "-i", video_in_path, os.path.join(video_out_path, "frame%05d.png")])
end = time.time()
print("ffmpeg png process took {} seconds".format(end - start))
else:
raise ValueError("format must be either PNG or JPEG")
return (folder_out,)
class VideoFramesFolderToImageStack(metaclass=ABCWidgetMetaclass):
"""Use cv2 to open the video and save the videos individual frames"""
@classmethod
def INPUT_TYPES(cls):
from addict import Dict
ret = Dict()
ret.required = Dict()
ret.required.folder_in = ("STRING", {"default": "full_path_to_folder"})
ret.required.use_subfolders = ([True, False], {"default": False})
ret.required.by_date_or_name = (["DATE", "NAME"], {"default": "DATE"})
ret.required.idx_slice_start = ("INT", {"min": 0, "max": 99999, "step": 1, "default": 0})
ret.required.idx_slice_stop = ("INT", {"min": 0, "max": 99999, "step": 1, "default": 1})
ret.required.slice_idx_step = ("INT", {"min": 0, "max": 9999, "step": 1, "default": 1})
ret.required.idx = ("INT", {"min": 0, "max": 99999, "step": 1, "default": 0})
ret.required.use_float16 = ([True, False], {"default": False})
return ret
CATEGORY = "video"
RETURN_TYPES = ("IMAGE", "LIST",)
RETURN_NAMES = ("image_stack", "[img_fname]",)
FUNCTION = "handler"
@torch.inference_mode()
def handler(self, folder_in,
use_subfolders,
by_date_or_name,
idx_slice_start,
idx_slice_stop,
slice_idx_step,
idx,
use_float16
):
"""
Return slices of time from the video, sequences of images, defined by the start, stop and step
- torch tensors (T,H,W,C)
"""
import glob
idx = idx * slice_idx_step
folder_in_path = folder_in
image_list = []
image_fname_list = []
p = os.path.abspath(folder_in_path)
# List and sort all PNG or JPG files in the folder
if not use_subfolders and by_date_or_name == "NAME":
file_fps = sorted(
glob.glob(os.path.join(p, '*.png')) + glob.glob(os.path.join(p, '*.jpg')))
elif not use_subfolders and by_date_or_name == "DATE":
file_fps = sorted(
glob.glob(os.path.join(p, '*.png')) + glob.glob(os.path.join(p, '*.jpg')),
key=os.path.getmtime)
elif use_subfolders and by_date_or_name == "NAME":
file_fps = sorted(
glob.glob(os.path.join(p, '**/*.png'), recursive=True) +
glob.glob(os.path.join(p, '**/*.jpg'), recursive=True))
elif use_subfolders and by_date_or_name == "DATE":
file_fps = sorted(
glob.glob(os.path.join(p, '**/*.png'), recursive=True) +
glob.glob(os.path.join(p, '**/*.jpg'), recursive=True),
key=os.path.getmtime)
# Ensure the slice doesn't go beyond the end of the file list
if idx_slice_stop + idx > len(file_fps):
idx_slice_stop = len(file_fps) - idx
slice_size = idx_slice_stop - idx_slice_start
# for i in range(idx_slice_start + idx, idx_slice_stop + idx):
for fn in file_fps[idx_slice_start + idx:idx_slice_start + idx + slice_size]:
image = cv2.imread(fn)
image_fname_list.append(os.path.basename(fn))
if image is not None:
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB) # converting to RGB format
image_list.append(image)
# Convert list of images to tensor (T,H,W,C)
if use_float16:
tensor = torch.stack([torch.tensor(img, dtype=torch.float16) for img in image_list])
else:
tensor = torch.stack([torch.tensor(img, dtype=torch.float32) for img in image_list])
del image_list
tensor = tensor / 255.0
return (tensor, image_fname_list,)
class SmoothStackTemporalByDistance2(ABCABCVideoFolderToImage, metaclass=ABCWidgetMetaclass):
@classmethod
def INPUT_TYPES(cls):
return {"required":
{
"image_stack": ("IMAGE",),
"top_thresh_normed": ("FLOAT", {"min": 0.0, "max": 100000.0, "step": 0.01, "default": 1.0}),
"bottom_thresh_normed": ("FLOAT", {"min": 0.0, "max": 100000.0, "step": 0.01, "default": 1.0}),
"max_iterations": ("INT", {"min": 0, "max": 1000, "step": 1, "default": 1}),
"target_deviation": ("FLOAT", {"min": 0.0, "max": 100000.0, "step": 0.01, "default": 0.0}),
}
}
RETURN_TYPES = ("IMAGE", "STRING")
def handler(self, image_stack, top_thresh_normed=0.9, bottom_thresh_normed=0.10, max_iterations=1,
target_deviation=0.0):
import torch
# Code for printing memory size and available GPU memory stays the same
image_stack = image_stack.to(torch.float16)
info = ""
iteration = 0
while True:
# Compute the first-order differences (gradient) along the time dimension (0)
first_order_diff = torch.gradient(image_stack, dim=0)[0]
# Compute the Euclidean norm of the first-order differences
euclidean_diff_first_order = torch.norm(first_order_diff.view(first_order_diff.shape[0], -1), dim=1).clone()
euclidean_diff_first_order_range = euclidean_diff_first_order.max() - euclidean_diff_first_order.min()
euclidean_diff_first_order_normed = (euclidean_diff_first_order - euclidean_diff_first_order.min()) / \
euclidean_diff_first_order_range
euclidean_diff_first_order_normed = euclidean_diff_first_order_normed[:-1]
start_mean = euclidean_diff_first_order.mean()
start_std = euclidean_diff_first_order.std()
if start_std <= target_deviation or iteration >= max_iterations:
break
# Identify the frames with differences greater than the top threshold, excluding the last index
outlier_indices_top = torch.where(euclidean_diff_first_order_normed > top_thresh_normed)[0]
# Interpolate frames where the differences are greater than the top threshold
interpolated_frames_top = 0.5 * (image_stack[outlier_indices_top] + image_stack[outlier_indices_top + 1])
image_stack[outlier_indices_top + 1] = interpolated_frames_top
# Identify the frames with differences less than the bottom threshold, excluding the last index
outlier_indices_bottom = torch.where(euclidean_diff_first_order_normed < bottom_thresh_normed)[0]
# Create a mask to delete frames where the differences are less than the bottom threshold
mask = torch.ones(image_stack.size(0), dtype=torch.bool)
mask[outlier_indices_bottom + 1] = False
image_stack = image_stack[mask]
# now remove the first and last frames
# image_stack = image_stack[1:-1]
# print the shape and stats for image_stack
# relate how many frames where changed via top threshold and bottom threshold
# also report on the mean and std of the euclidean_diff_first_order
info += f"Number of frames changed via top threshold: {len(outlier_indices_top)}\n " \
f"Number of frames changed via bottom threshold: {len(outlier_indices_bottom)}\n " \
f"Mean of euclidean_diff_first_order: {start_mean}\n " \
f"Std of euclidean_diff_first_order: {start_std}\n " \
f"\n\n"
print(info)
iteration += 1
if image_stack.shape[0] == 0:
print("image_stack is empty, returning None")
return (None, None,)
print(
f"image_stack shape: {image_stack.shape}, image_stack stats: {image_stack.min()}, {image_stack.max()}")
return (image_stack.to(torch.float32), info,)
class SmoothStackTemporalByDistance(ABCABCVideoFolderToImage, metaclass=ABCWidgetMetaclass):
"""
Smooth the temporal dimension of the image stack by calculating the distance between each frame
and the next frame and then averaging the frames that are within a certain distance of each other.
"""
@classmethod
def INPUT_TYPES(cls):
return {"required":
{
"image_stack": ("IMAGE",),
"top_thresh_normed": ("FLOAT", {"min": 0.0, "max": 100000.0, "step": 0.1, "default": 1.0}),
"bottom_thresh_normed": ("FLOAT", {"min": 0.0, "max": 100000.0, "step": 0.1, "default": 1.0}),
"iterations": ("INT", {"min": 0, "max": 1000, "step": 1, "default": 1}),
}
}
RETURN_TYPES = ("IMAGE", "STRING")
def handler(self, image_stack,
top_thresh_normed=0.9, bottom_thresh_normed=0.10, iterations=1):
import torch
# calculate and print the total size in memory in bytes of the tensor
print(
f"Size of tensor in memory in GB: {image_stack.element_size() * image_stack.nelement() / (1024 ** 3)} GB, dtype: {image_stack.dtype}")
# print the available memory on the GPU in GB
print(f"Available memory on GPU in GB: {torch.cuda.get_device_properties(0).total_memory / (1024 ** 3)} GB")
# image_stack = image_stack.clone()
image_stack = image_stack.to(torch.float16)
info = ""
for _ in range(iterations):
# Compute the first-order differences (gradient) along the time dimension (0)
first_order_diff = torch.gradient(image_stack, dim=0)[0]
# Compute the Euclidean norm of the first-order differences
euclidean_diff_first_order = torch.norm(first_order_diff.view(first_order_diff.shape[0], -1), dim=1).clone()
euclidean_diff_first_order_range = euclidean_diff_first_order.max() - euclidean_diff_first_order.min()
euclidean_diff_first_order_normed = (euclidean_diff_first_order - euclidean_diff_first_order.min()) / \
euclidean_diff_first_order_range
euclidean_diff_first_order_normed = euclidean_diff_first_order_normed[:-1]
start_mean = euclidean_diff_first_order.mean()
start_std = euclidean_diff_first_order.std()
# Identify the frames with differences greater than the top threshold, excluding the last index
outlier_indices_top = torch.where(euclidean_diff_first_order_normed > top_thresh_normed)[0]
# Interpolate frames where the differences are greater than the top threshold
interpolated_frames_top = 0.5 * (image_stack[outlier_indices_top] + image_stack[outlier_indices_top + 1])
image_stack[outlier_indices_top + 1] = interpolated_frames_top
# Identify the frames with differences less than the bottom threshold, excluding the last index
outlier_indices_bottom = torch.where(euclidean_diff_first_order_normed < bottom_thresh_normed)[0]
# Create a mask to delete frames where the differences are less than the bottom threshold
mask = torch.ones(image_stack.size(0), dtype=torch.bool)
mask[outlier_indices_bottom + 1] = False
image_stack = image_stack[mask]
# now remove the first and last frames
# image_stack = image_stack[1:-1]
# print the shape and stats for image_stack
# relate how many frames where changed via top threshold and bottom threshold
# also report on the mean and std of the euclidean_diff_first_order
info += f"Number of frames changed via top threshold: {len(outlier_indices_top)}\n " \
f"Number of frames changed via bottom threshold: {len(outlier_indices_bottom)}\n " \
f"Mean of euclidean_diff_first_order: {start_mean}\n " \
f"Std of euclidean_diff_first_order: {start_std}\n " \
f"\n\n"
print(info)
if image_stack.shape[0] == 0:
print("image_stack is empty, returning None")
return (None, None,)
print(
f"image_stack shape: {image_stack.shape}, image_stack stats: {image_stack.min()}, {image_stack.max()}")
return (image_stack.to(torch.float32), info,)
class GetImageStackStatistics(ABCABCVideoFolderToImage, metaclass=ABCWidgetMetaclass):
"""Get statistics of an image stack"""
@classmethod
def INPUT_TYPES(cls):
return {"required":
{
"image_stack": ("IMAGE",),
"method": (["SSIM", "MSE", "REGIONAL"], {"default": "SSIM"}),
},
}
RETURN_TYPES = ("STRING", "LIST", "IMAGE",)
def handler(self, image_stack, method):
import numpy as np
import torch
# if the dtype is not float32
if image_stack.dtype != torch.float32:
image_stack = image_stack.to(torch.float32)
image_stack_BCWH = image_stack.permute(0, 3, 1, 2)
image_stack_BCWH_DIFF = image_stack_BCWH[1:] - image_stack_BCWH[:-1]
# we want to calculate the mean squared error
# of the image stack
# so we need to calculate the difference between each image in the stack and the previous image
# we have that in image_stack_BCWH_DIFF
# now we need to calculate the mean squared error of each image in the stack
# which will give us a tensor of shape (B)
# where B is the number of images in the stack
if method == "MSE":
stack_mse_B = torch.mean(image_stack_BCWH_DIFF ** 2, dim=(1, 2, 3))
# find the max possible value of the mean squared error
elif method == "REGIONAL":
"""
this method resamples the image stack of B,W,H,C to B,8,8,C
then calculates the mean squared error of each region of the image
by further slicing into [b,1,1,C] x 64 regional representations
the greatest mean squared from each region is returned
"""
# the input image stack is defined as "image_stack" and it is a torch tensor, B,H,W,C
B, H, W, C = image_stack.shape
levels = [3, 4, 5, 6, 7, 8, 16, 32, 64]
max_mse_value_in_all_regions_and_frames_for_this_level = 0
level_stats = {k: {"max_mse": None, "max_std_dev": None} for k in levels}
for size in levels:
# need to re order the channels for the interpolation
resampled_stack = image_stack.permute(0, 3, 1, 2)
# Resample the image stack to 8x8 pixels
resampled_stack = torch.nn.functional.interpolate(resampled_stack, size=(size, size), mode='bilinear',
align_corners=False)
# but the channel order back to the original
resampled_stack = resampled_stack.permute(0, 2, 3, 1)
# Reshape the resampled stack to (B, 64, C)
regional_stack = resampled_stack.view(B, size ** 2, C)
# Calculate the mean squared difference between consecutive frames for each region
mse_values = torch.mean((regional_stack[1:] - regional_stack[:-1]) ** 2, dim=2)
# max mse for each frame
max_mse_ea_frame_at_level = torch.max(mse_values, dim=1)[0]
max_dev_ea_frame_at_level = torch.std(mse_values, dim=1)
# Calculate the standard deviation of the mean squared error for each region along the time dimension
max_std_dev_of_all_regions_over_time = torch.max(torch.std(mse_values, dim=0))
# in terms of video
# represents the maximum amount of change in the image stack, basically the maximum amount of motion
max_mse_value_in_all_regions_and_frames_for_this_level = torch.max(mse_values.view(-1))
level_stats[size]["max_r_mse_ea_frame"] = max_mse_ea_frame_at_level
level_stats[size]["max_r_dev_ea_frame"] = max_dev_ea_frame_at_level
level_stats[size]["max_mse"] = max_mse_value_in_all_regions_and_frames_for_this_level
level_stats[size]["max_std_dev"] = max_std_dev_of_all_regions_over_time
all_l_max_mse_ea_frame = [level_stats[k]["max_r_mse_ea_frame"].tolist() for k in levels]
all_l_max_std_ea_frame = [level_stats[k]["max_r_dev_ea_frame"].tolist() for k in levels]
max_l_r_mse_per_batch = [] # Initialize list to store the maximum mean squared error for each batch_num
for batch_num in range(B - 1):
max_value = float('-inf') # Initialize max_value to negative infinity for each batch_num
for level in levels:
mse_for_level_and_batch = level_stats[level]["max_r_mse_ea_frame"][batch_num]
# Update max_value if the current mse_for_level_and_batch is greater
if mse_for_level_and_batch > max_value:
max_value = mse_for_level_and_batch
max_l_r_mse_per_batch.append(
max_value) # Append the maximum value for the current batch_num to the list
max_l_r_dev_per_batch = [] # Initialize list to store the maximum standard deviation for each batch_num
for batch_num in range(B - 1):
max_value = float('-inf')
for level in levels:
dev_for_level_and_batch = level_stats[level]["max_r_dev_ea_frame"][batch_num]
if dev_for_level_and_batch > max_value:
max_value = dev_for_level_and_batch
max_l_r_dev_per_batch.append(max_value)
# all_f_max_mse_ea_frame = [level_stats[k]["max_mse_ea_frame"] for k in levels]
max_mse_value_in_all_regions_and_frames_for_all_levels = torch.max(torch.tensor(all_l_max_mse_ea_frame))
max_std_dev_value_in_all_regions_and_frames_for_all_levels = torch.max(torch.tensor(all_l_max_std_ea_frame))
out_str = f"max difference between any region over time: {max_mse_value_in_all_regions_and_frames_for_this_level}\nmax of how much the amount of motion changes over time: {max_std_dev_of_all_regions_over_time}"
out_list = [max_mse_value_in_all_regions_and_frames_for_all_levels,
max_std_dev_value_in_all_regions_and_frames_for_all_levels,
max_l_r_mse_per_batch,
max_l_r_dev_per_batch
]
# return the resampled image stack
out_image = None
return (out_str, out_list, out_image,)
else:
# not implimented
raise NotImplementedError
image_stack_diff_len = len(stack_mse_B)
required_image_info_width = 512 + image_stack_diff_len
mean = torch.mean(stack_mse_B)
std = torch.std(stack_mse_B)
median = torch.median(stack_mse_B)
max_mse = torch.max(stack_mse_B)
min_mse = torch.min(stack_mse_B)
out_str = f"mean: {mean}\n std: {std}\n median: {median}\n max: {max_mse}\n min: {min_mse}\n len: {image_stack_diff_len}"
# create the histogram
hist_values, _ = np.histogram(stack_mse_B.numpy(), bins=min(100, image_stack_diff_len))
hist_values = hist_values / hist_values.max() # Normalize to [0, 1]
# Create an image tensor with shape (1, 512, required_width, 3) filled with zeros
out_hist_image = torch.zeros((1, 512, required_image_info_width, 3))
# Determine the scaling factors for width and height for the histogram
scale_width = 512 // len(hist_values)
scale_height = out_hist_image.shape[1]
# Fill the image tensor with the histogram
for i, value in enumerate(hist_values):
height = 512 - int(value * scale_height)
start_x = int(i * scale_width)
end_x = int((i + 1) * scale_width)
out_hist_image[0, height:, start_x:end_x - 1, :] = 1 # Color as white
# repeat each element of the stack_mse_B tensor by scale_width
stack_mse_B = stack_mse_B.repeat_interleave(scale_width)
# Create a plot for the difference for each frame pair
plot_offset = 512
for i, value in enumerate(stack_mse_B):
height = 512 - int(value * (scale_height / .5)) # Normalize by the max value
start_x = plot_offset + i
end_x = plot_offset + i + 1
out_hist_image[0, height:, start_x:end_x, :] = (1 * height) / 512 # Color as white
return (out_str, out_hist_image,)
# @profile
class VideoFileToImageStack(ABCABCVideoFileToImage, metaclass=ABCWidgetMetaclass):
"""Use cv2 to open a video and save the videos individual frames"""
RETURN_TYPES = ("IMAGE", "FUNC",)
RETURN_NAMES = ("image_stack", "FUNC(**kwargs)",)
def handler(self, video_in, idx_slice_start, idx_slice_stop, slice_idx_step, idx, func_only=False):
"""
Return slices of time from the video, sequences of images, defined by the start, stop and step
- torch tensors (T,H,W,C)
"""
import sys
import gc
# print memory usage
print(f"Memory usage: {sys.getsizeof(self)} bytes")
if func_only:
return (None, VideoFileToImageStack.handler,)
image_list = []
video_in_fp = video_in
vidcap = cv2.VideoCapture(video_in_fp)
max_frames = int(vidcap.get(cv2.CAP_PROP_FRAME_COUNT))
if not vidcap.isOpened():
print("Error: Unable to open video file")
return
start = (idx * slice_idx_step) + idx_slice_start
stop = (idx * slice_idx_step) + idx_slice_stop
slice_length = stop - start
if (start + slice_length) > max_frames:
raise ValueError(f"idx_slice_stop ({idx_slice_stop}) is greater than the max frames ({max_frames})")
for frame_count in range(start, start + slice_length):
# Seek to frame_count frame
vidcap.set(cv2.CAP_PROP_POS_FRAMES, frame_count)
success, image = vidcap.read()
if not success:
print("Error: Unable to read frame from video file")
break
# image_list.append(image)
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
image_list.append(image)
print(f"Memory usage: {sys.getsizeof(self)} bytes")
vidcap.release()
# Convert list of images to tensor (T,H,W,C)
try:
tensor = torch.stack([torch.tensor(img, dtype=torch.float32) for img in image_list])
except RuntimeError as e:
print("Error: Unable to convert image list to tensor")
raise e
tensor = tensor / 255.0
del image_list
del vidcap
del image
gc.collect()
# print memory usage
print(f"Memory usage: {sys.getsizeof(self)} bytes")
return (tensor, VideoFileToImageStack.handler,)
class ImageStackToVideoFramesFolder(metaclass=ABCWidgetMetaclass):
"""Really just saves the image stack to a folder"""
@classmethod
def INPUT_TYPES(self):
return {"required":
{
"image_stack": ("IMAGE",),
"folder_out": ("STRING", {"multiline": False, "default": "video01"}),
},
}
RETURN_TYPES = ("STRING",)
CATEGORY = "video"
OUTPUT_NODE = True
FUNCTION = "handler"
def handler(self, **kwargs):
import tempfile
import uuid
import shutil
image_stack = kwargs["image_stack"]
folder_out = kwargs["folder_out"]
folder_out_path = video_dir
# full path to the output folder
folder_out_path = os.path.join(video_dir, folder_out)
# Create the output folder if it doesn't exist
if not os.path.exists(folder_out_path):
os.makedirs(folder_out_path)
# move any current files in the output folder to a random folder inside the video folder
random_folder = os.path.join(folder_out_path, str(uuid.uuid4()))
os.makedirs(random_folder)
for file in os.listdir(folder_out_path):
file_path = os.path.join(folder_out_path, file)
if os.path.isfile(file_path):
shutil.move(file_path, random_folder)
# Save each image in the image stack to the output folder
for i in range(image_stack.shape[0]):
image = image_stack[i].numpy()
# make sure the image is in the correct color format
image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
# convert to 0-255
image = image * 255
# use zfill to get a good ordering of the frames
file_name = os.path.join(folder_out_path, "frame" + str(i).zfill(6) + ".jpg")
# use the best compression nearly lossless
cv2.imwrite(file_name, image, [cv2.IMWRITE_JPEG_QUALITY, 99])
image_stack = None
del image_stack
return (folder_out,)
class SmoothStackTemporal(ABCABCVideoFolderToImage, metaclass=ABCWidgetMetaclass):
"""
Smooth the temporal dimension of the image stack by calculating the distance between each frame
and the next frame and then averaging the frames that are within a certain distance of each other.
Supports adaptive mode with target deviation and baseline mode.
Can also perform frame interpolation using RIFE VFI.
"""
rife_kwargs = {
"ckpt_name": "rife49.pth",
"clear_cache_after_n_frames": 1000,
"multiplier": 2,
"fast_mode": False,
"ensemble": False,
"scale_factor": 1.0,
"optional_interpolation_states": None,
}
@classmethod
def INPUT_TYPES(cls):
return {"required":
{
"image_stack": ("IMAGE",),
"top_threshold": ("FLOAT", {"min": 0.0, "max": 100000.0, "step": 0.01, "default": 1.0}),
"bottom_threshold": ("FLOAT", {"min": 0.0, "max": 100000.0, "step": 0.01, "default": 1.0}),
"max_iters": ("INT", {"min": 0, "max": 1000, "step": 1, "default": 1}),
"target_deviation": ("FLOAT", {"min": 0.0, "max": 100000.0, "step": 0.01, "default": 0.0}),
"interp_frames": ("BOOLEAN", {"default": False}),
"mode": (["adaptive", "baseline"],),
}
}
RETURN_TYPES = ("IMAGE", "STRING")
def compute_stats(self, image_stack):
stats_handler = GetImageStackStatistics().handler
res = stats_handler(image_stack=image_stack, method="REGIONAL")
stats_str, stats_list, _ = res
region_diff = torch.tensor(stats_list[2]) # The list of each frame pairs maximum regional difference
max_diff = stats_list[0] # Overall maximum difference
max_dev = stats_list[1] # Overall regional deviation
return stats_str, region_diff, max_diff, max_dev
def compute_baseline_motion(self, region_diff):
values, counts = torch.unique(region_diff, return_counts=True)
baseline_motion = values[torch.argmax(counts)].item()
return baseline_motion
def normalize_diff(self, region_diff):
region_diff_range = region_diff.max() - region_diff.min()
if region_diff_range == 0:
region_diff_range = 1 # Avoid division by zero
region_diff_norm = (region_diff - region_diff.min()) / region_diff_range
return region_diff_norm
def process_top_outliers(self, image_stack, region_diff_norm, top_threshold, rife_vfi):
import torch.nn.functional as F
top_outliers = torch.where(region_diff_norm > top_threshold)[0]
new_image_stack = []
for idx in range(image_stack.shape[0]):
new_image_stack.append(image_stack[idx])
if idx in top_outliers and idx < image_stack.shape[0] - 1: # Ensure we don't go out of bounds
frame_pair = image_stack[idx:idx + 2]
interpolated_frames = rife_vfi(frames=frame_pair, **self.rife_kwargs)[0]
# Calculate the MSE for each interpolated frame with the first and last frames of the pair
mse_first = [F.mse_loss(interpolated_frame, frame_pair[0]) for interpolated_frame in
interpolated_frames]
mse_last = [F.mse_loss(interpolated_frame, frame_pair[1]) for interpolated_frame in interpolated_frames]
# Find the interpolated frame with the smallest difference between mse_first and mse_last
mse_diff = [abs(mse_f - mse_l) for mse_f, mse_l in zip(mse_first, mse_last)]
middle_frame_index = torch.argmin(torch.tensor(mse_diff)).item()
choice = interpolated_frames[middle_frame_index]
new_image_stack.append(choice)
new_image_stack = torch.stack(new_image_stack)
return new_image_stack, len(top_outliers)
def drop_bottom_outliers(self, image_stack, region_diff_norm, bottom_threshold):
bottom_outliers = torch.where(region_diff_norm < bottom_threshold)[0]
if len(bottom_outliers) > 0:
mask = torch.ones(image_stack.size(0), dtype=torch.bool)
mask[bottom_outliers + 1] = False
image_stack = image_stack[mask]
return image_stack, len(bottom_outliers)
def log_info(self, iteration, top_outliers_count, bottom_outliers_count, mean_diff, std_diff, max_diff, max_dev,
stats_str):
info = f"Iteration {iteration}:\n"
info += f"Frames changed by top threshold: {top_outliers_count}\n"
info += f"Frames dropped by bottom threshold: {bottom_outliers_count}\n"
info += f"Mean of region_diff: {mean_diff}\n"
info += f"Std of region_diff: {std_diff}\n"
info += f"Overall max difference: {max_diff}\n"
info += f"Overall max deviation: {max_dev}\n"
info += f"Image stack statistics: {stats_str}\n\n"
return info
def handler(self, image_stack, top_threshold=0.9, bottom_threshold=0.10, max_iters=1, target_deviation=0.0,
interp_frames=False, mode="adaptive"):
import torch
import nodes
image_stack = image_stack.to(torch.float16)
log = ""
iteration = 0
if interp_frames:
rife_vfi = nodes.NODE_CLASS_MAPPINGS["RIFE VFI"]().vfi
if mode == "baseline":
# Compute initial statistics and baseline motion
stats_str, region_diff, max_diff, max_dev = self.compute_stats(image_stack)
baseline_motion = self.compute_baseline_motion(region_diff)
target_deviation = baseline_motion
log += f"Baseline motion (mode of max regional differences): {baseline_motion}\n"
while True:
# Step 1: Compute statistics
stats_str, region_diff, max_diff, max_dev = self.compute_stats(image_stack)
# Step 2: Normalize differences
region_diff_norm = self.normalize_diff(region_diff)
mean_diff = region_diff.mean().item()
std_diff = region_diff.std().item()
# Step 3: Check if the target deviation is met or if max iterations are reached
if std_diff <= target_deviation or iteration >= max_iters:
log += f"std_diff <= target_deviation or iteration >= max_iters"
break
# Step 4: Identify and process top outliers
image_stack, top_outliers_count = self.process_top_outliers(image_stack, region_diff_norm, top_threshold,
rife_vfi)
# Step 5: Identify and drop bottom outliers
image_stack, bottom_outliers_count = self.drop_bottom_outliers(image_stack, region_diff_norm,
bottom_threshold)
# Log information
log += self.log_info(iteration, top_outliers_count, bottom_outliers_count, mean_diff, std_diff, max_diff,
max_dev, stats_str)
# Increment the iteration counter
iteration += 1
# Step 6: Check if the image stack is empty
if image_stack.shape[0] == 0:
return (None, "Image stack is empty after processing. No frames left.")
return (image_stack.to(torch.float32), log)
class MovingCircle(metaclass=ABCWidgetMetaclass):
# VideoWidget class definition
def calculate_start(self, x1, y1, width, height):
# Convert to 0-1 range
start_x = x1 * width
start_y = y1 * height
return (start_x, start_y)
def calculate_end(self, x2, y2, width, height):
# Convert to 0-1 range
end_x = x2 * width
end_y = y2 * height
return (end_x, end_y)
def generate_frames(self, num_frames, width, height):
# Match tensor shape from example
frames = torch.zeros(num_frames, width, height, 3)
return frames
def draw_circle_or_oval(self, kp, width, height, radius, aspect_ratio, color=(255, 255, 255)):
import torch.nn.functional as F
# Initialize black background
img = torch.zeros((3, height, width), dtype=torch.uint8)
# Draw circle as keypoints
keypoints = torch.tensor([[kp[0], kp[1]]]).unsqueeze(0)
img = draw_keypoints(img, keypoints, colors=color, radius=radius)
# convert to (w,h,c)
# img = img.permute(1, 2, 0).unsqueeze(0)
img = img.unsqueeze(0)
# Calculate new size
if aspect_ratio < 1:
new_size = (height, round(width / aspect_ratio))
else:
new_size = (round(height * aspect_ratio), width)
# Stretch the image
img_stretched = F.interpolate(img, size=new_size, mode='bilinear', align_corners=False)
# Shrink it back
img_shrunk = F.interpolate(img_stretched, size=(height, width), mode='bilinear', align_corners=False)
# Convert back to float values
img_shrunk = img_shrunk.float() / 255
img_shrunk = img_shrunk.permute(0, 2, 3, 1)
return img_shrunk
def draw_circle(self, kp, width, height, radius, color=(255, 255, 255)):
# Initialize black background
img = torch.zeros((3, height, width), dtype=torch.uint8)
# Draw circle as keypoints
keypoints = torch.tensor([[kp[0], kp[1]]]).unsqueeze(0)
img = draw_keypoints(img, keypoints, colors=color, radius=radius)
# convert to (w,h,c)
img = img.permute(1, 2, 0).unsqueeze(0)
# convert back to float values
img = img.float() / 255
return img
# Metadata for documentation
RETURN_TYPES = ("IMAGE",)
CATEGORY = "ETK/video"
FUNCTION = "handler"
# Match parent class method
@classmethod
def INPUT_TYPES(cls):
# Validate types, ranges
return {"required":
{
"x1": ("FLOAT", {"min": 0, "max": 1, "step": 0.1}),
"y1": ("FLOAT", {"min": 0, "max": 1, "step": 0.1}),
"x2": ("FLOAT", {"min": 0, "max": 1, "step": 0.1}),
"y2": ("FLOAT", {"min": 0, "max": 1, "step": 0.1}),
"num_frames": ("INT", {"min": 1}),
"width": ("INT", {"min": 1}),
"height": ("INT", {"min": 1}),
"radius": ("INT", {"min": 1}),
"color": ("STRING", {"default": "255 255 255"}),
"aspect_ratio": ("FLOAT", {"min": 0.1, "max": 10})
}}
def handler(self, x1, y1, x2, y2, num_frames, width, height, radius, color="255 255 255", aspect_ratio=1):
import torch
# Avoid errors on bad values
assert 0 <= x1 <= 1
assert 0 <= y1 <= 1
# Split calculations
start = torch.tensor(self.calculate_start(x1, y1, width, height))
# Split calculations
end = torch.tensor(self.calculate_end(x2, y2, width, height))
# color is supposed to be a tuple
color = tuple(map(lambda x: int(x), color.split()))
# Initialize for compositing
frames = []
for i in range(num_frames):
# Leverage PyTorch
cur_pos = start + (end - start) * i / num_frames
circle = self.draw_circle_or_oval(cur_pos, width, height, radius, aspect_ratio, color)
# Use broadcasting to overlay
frames.append(circle)
# cat
frames = torch.cat(frames, dim=0)
return (frames,)
if __name__ == "__main__":
import doctest
doctest.testmod()
class FindFrameFolders(metaclass=ABCWidgetMetaclass):
"""Finds all folders directly off a given root path containing sequential image files"""
@classmethod
def INPUT_TYPES(self):
return {
"required": {
"root_path": ("STRING", {"default": "path to root of project"})
}
}
RETURN_TYPES = ("LIST",)
RETURN_NAMES = ("image_sequence_folders",)
CATEGORY = "video"
OUTPUT_NODE = True
FUNCTION = "handler"
def handler(self, **kwargs):
root_path = kwargs["root_path"]
image_sequence_folders = []
for folder_name in os.listdir(root_path):
folder_path = os.path.join(root_path, folder_name)
if os.path.isdir(folder_path):
image_files = [file for file in os.listdir(folder_path) if file.endswith(('.png', '.jpg', '.jpeg'))]
if len(image_files) >= 2:
image_sequence_folders.append(folder_path)
return (image_sequence_folders,)
class CombineAudioAndVideoFiles(metaclass=ABCWidgetMetaclass):
"""Combines lists of audio files and video files using imageio-ffmpeg"""
@classmethod
def INPUT_TYPES(cls):
return {
"required": {
"audio_files": ("LIST", {"default": []}),
"video_files": ("LIST", {"default": []}),
"resample": (["None", "video", "audio", ],),
"fps": ("FLOAT", {"default": 30, "min": 1, "max": 120}),
},
"optional": {
"move_to_folder": ("STRING", {"default": ""}),
}
}
RETURN_TYPES = ("LIST",)
RETURN_NAMES = ("combined_video_paths",)
CATEGORY = "video"
OUTPUT_NODE = True
FUNCTION = "handler"
@classmethod
def has_audio_stream(cls, video_file):
"""Check if the video file has an audio stream."""
command = ['ffprobe', '-v', 'error', '-select_streams', 'a', '-show_entries', 'stream=codec_name', '-of',
'default=nw=1:nk=1', video_file]
try:
output = subprocess.check_output(command, stderr=subprocess.STDOUT)
return bool(output.strip())
except subprocess.CalledProcessError:
return False
import subprocess
import os
from os.path import splitext, join
import shutil
def get_duration(self, file):
"""Get the duration of a file in seconds."""
command = ['ffprobe', '-v', 'error', '-show_entries', 'format=duration', '-of',
'default=nw=1:nk=1', file]
try:
output = subprocess.check_output(command, stderr=subprocess.STDOUT)
return float(output.strip())
except subprocess.CalledProcessError:
return None
def handler(self, **kwargs):
audio_files = kwargs["audio_files"]
video_files = kwargs["video_files"]
move_to_folder = kwargs.get("move_to_folder", "")
resample = kwargs.get("resample", "None")
fps = kwargs.get("fps", 30.0)
combined_video_paths = []
video_audio_pairs = list(zip(video_files, audio_files))
ffmpeg_exe = ffmpeg.get_ffmpeg_exe()
if ffmpeg_exe is None:
raise RuntimeError("FFmpeg could not be found.")
for video_file, audio_file in video_audio_pairs:
video_file = f'{video_file}'
audio_file = f'{audio_file}'
video_out_file_fp = f'{splitext(video_file)[0]}_v_and_a.mp4'
input_args = ['-i', video_file, '-i', audio_file]
if self.has_audio_stream(video_file):
# The video has an audio stream, proceed with mixing
filter_complex_arg = '[0:a][1:a]amix=inputs=2:duration=longest[a]'
output_args = ['-c:v', 'copy', '-map', '0:v:0', '-map', '[a]', '-c:a', 'aac', '-strict', 'experimental',
'-y', video_out_file_fp]
else:
filter_complex_arg = ''
output_args = ['-c:v', 'copy', '-map', '0:v:0', '-map', '1:a:0', '-c:a', 'aac', '-strict',
'experimental', '-y', video_out_file_fp]
if resample == "video":
video_duration = self.get_duration(video_file)
audio_duration = self.get_duration(audio_file)
if video_duration and audio_duration:
speed = audio_duration / video_duration
filter_complex_arg = f'[0:v]setpts={speed}*PTS[v];[v][1:a]concat=n=1:v=1:a=1[v][a]'
output_args = ['-map', '[v]', '-map', '[a]', '-c:v', 'libx264', '-c:a', 'aac', '-strict',
'experimental', '-y', video_out_file_fp]
else:
raise RuntimeError("Could not determine duration of video or audio file.")
elif resample == "None":
output_args.extend(['-r', str(fps)])
elif resample == "audio":
# Not implemented: Finding silence in audio and removing it
raise NotImplementedError(
"Resampling audio is not implemented. Consider finding silence in the audio and removing it.")
try:
# try to run it with the list of parts, using the list
if filter_complex_arg:
commands_to_run = [ffmpeg_exe, *input_args, '-filter_complex', filter_complex_arg, *output_args]
else:
commands_to_run = [ffmpeg_exe, *input_args, *output_args]
commands_to_run = [str(c).replace("\"", "") for c in commands_to_run]
subprocess.run(commands_to_run, check=True)
except subprocess.CalledProcessError as e:
print(f"An error occurred: {e}")
if move_to_folder:
destination_path = join(move_to_folder, os.path.basename(video_out_file_fp))
if video_out_file_fp != destination_path:
shutil.move(video_out_file_fp, destination_path)
combined_video_paths.append(destination_path)
else:
combined_video_paths.append(video_out_file_fp)
os.chdir("..")
os.chdir("..")
os.chdir("..")
return (combined_video_paths,)
class CombineFoldersWithFramesToVideoFiles(metaclass=ABCWidgetMetaclass):
"""Combines lists of folders containing frames of individual video files using imageio and ffmpeg to create video files"""
@classmethod
def INPUT_TYPES(cls):
return {
"required": {
"folders_with_frames": ("LIST", {"default": []}),
"fps/length": ("FLOAT", {"default": 30, "min": 1, "max": 10000000000}),
"bitrate": ("STRING", {"default": "20000k"}),
"codec": (
["libx264", "libx265", "mpeg4", "vp9", "huffyuv", "flv1", "h264_nvenc", "hevc_nvenc", "rawvideo"],
{"default": "libx264"}
),
"file_extension": (
[".mp4", ".mov", ".avi", ".flv", ".mkv", ".webm"],
{"default": ".mp4"}
),
"fps or length": (["fps", "length"], {"default": "fps"}),
},
"optional": {
"[duration or fps]": ("LIST", {"default": []}),
"q": ("INT", {"default": 20, "min": 0, "max": 31}),
"pix_fmt": ("STRING", {"default": "yuv420p"}),
},
}
RETURN_TYPES = ("LIST", "STRING",)
RETURN_NAMES = ("combined_video_paths",)
CATEGORY = "video"
OUTPUT_NODE = True
FUNCTION = "handler"
def handler(self, **kwargs):
folders_with_frames = kwargs.get("folders_with_frames", [])
fps_length_value = kwargs.get("fps/length", 30)
file_extension = kwargs.get("file_extension", ".mkv")
fps_or_length = kwargs.get("fps or length", "fps")
duration_or_fps_list = kwargs.get("[duration or fps]", [])
codec = kwargs.get("codec", "libx264")
q = kwargs.get("q", 20)
pix_fmt = kwargs.get("pix_fmt", "yuv420p")
combined_video_paths = []
messages = []
ffmpeg_exe = 'ffmpeg' # Ensure ffmpeg is in your PATH or provide the full path to the ffmpeg executable
if not shutil.which(ffmpeg_exe):
raise RuntimeError("FFmpeg could not be found.")
for i, folder_path in enumerate(folders_with_frames):
try:
output_file = os.path.join(os.path.dirname(folder_path),
f"{os.path.basename(folder_path)}tmp_{i}{file_extension}")
if fps_or_length == 'fps':
fps = fps_length_value
if duration_or_fps_list:
fps = duration_or_fps_list[i]
else:
frame_files = sorted(
[f for f in os.listdir(folder_path) if f.lower().endswith(('.jpg', '.jpeg', '.png'))]
)
total_frames = len(frame_files)
fps = total_frames / fps_length_value if fps_length_value != 0 else 30
if duration_or_fps_list:
fps = total_frames / duration_or_fps_list[i]
create_video_from_images(folder_path, output_file, fps, codec, q, pix_fmt)
combined_video_paths.append(output_file)
except Exception as e:
import traceback
# Capture full traceback information
full_traceback = traceback.format_exc()
messages.append(f"Error processing folder '{folder_path}': {str(e)}\nFull Traceback:\n{full_traceback}")
return (combined_video_paths, "\n".join(messages),)
def create_file_list(folder_path):
frame_files = sorted(
[f for f in os.listdir(folder_path) if f.lower().endswith(('.jpg', '.jpeg', '.png'))]
)
list_file_path = os.path.join(folder_path, 'filelist.txt')
with open(list_file_path, 'w') as f:
for frame_file in frame_files:
f.write(f"file '{os.path.join(folder_path, frame_file)}'\n")
return list_file_path
import subprocess
def create_video_from_images(folder_path, output_file, fps, codec, q, pix_fmt):
import os
import subprocess
import shutil
import traceback
# Assuming create_file_list is part of the larger framework, not redefining it.
list_file_path = create_file_list(folder_path)
# Ensure ffmpeg is available in PATH
ffmpeg_exe = shutil.which('ffmpeg')
if ffmpeg_exe is None:
raise RuntimeError("FFmpeg is not found. Please ensure it is installed and available in the system's PATH.")
# Normalize paths for cross-platform compatibility
list_file_path = os.path.normpath(list_file_path)
output_file = os.path.normpath(output_file)
# FFmpeg command to create the video
command = [
ffmpeg_exe, '-y', '-f', 'concat', '-safe', '0', '-i', list_file_path,
'-vf', f"fps={fps}", '-c:v', codec, '-pix_fmt', pix_fmt, '-qp', str(q), output_file
]
try:
# Capture both stdout and stderr to diagnose errors
subprocess.run(command, check=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
except subprocess.CalledProcessError as e:
# Capture the full traceback and the error output from the ffmpeg process
error_message = e.stderr.decode() if e.stderr else "No additional error information"
full_traceback = traceback.format_exc()
raise RuntimeError(f"ffmpeg failed with error: {error_message}\nFull Traceback:\n{full_traceback}")
class ImageStackToVideoFile(metaclass=ABCWidgetMetaclass):
"""Really just saves the image stack to a video file"""
@classmethod
def INPUT_TYPES(self):
return {"required":
{
"image_stack": ("IMAGE",),
"video_out": ("STRING", {"multiline": False, "default": "video01.mp4"}),
"fps or length": (["fps", "length"], {"default": "fps"}),
"fps/length": ("FLOAT", {"default": 30, "min": 1, "max": 10000000000}),
"bitrate": ("STRING", {"default": "20000k"}),
"codec": (
["libx264", "libx265", "mpeg4", "libvpx-vp9", "huffyuv", "flv1"],
{"default": "libx264"}
),
},
"optional":
{
"audio_file": ("STRING", {"default": "NONE"}),
},
}
RETURN_TYPES = ("FUNC", "STRING",)
RETURN_NAMES = ("FUNC", "video_out_path",)
CATEGORY = "video"
OUTPUT_NODE = True
FUNCTION = "handler"
def handler(self, **kwargs):
import os
import numpy as np
import uuid
import shutil
import imageio
import audioread
import cv2
import torchvision
use_torchvision = False
# video_dir = kwargs["video_dir"]
audio_file = kwargs.get("audio_file")
if audio_file == "NONE":
audio_file = None
image_stack = kwargs["image_stack"]
fps_or_length = kwargs["fps or length"]
fps_length_value = kwargs["fps/length"]
bitrate = kwargs["bitrate"]
codec = kwargs["codec"]
video_out = kwargs["video_out"]
if fps_or_length == "fps":
fps_length_value = np.array([fps_length_value], dtype=np.float16)[0]
video_out_path = video_out
if os.path.exists(video_out_path):
random_folder = os.path.join(os.path.dirname(video_out_path), str(uuid.uuid4()))
os.makedirs(random_folder)
shutil.move(video_out_path, random_folder)
if audio_file is not None:
audio_file = audio_file
if not os.path.exists(audio_file):
# raise Exception(f"Audio file {audio_file} does not exist")
audio_length = 0
else:
with audioread.audio_open(audio_file) as audio_info:
audio_length = audio_info.duration
frames_length = image_stack.shape[0]
if fps_or_length == "fps":
video_length = frames_length / fps_length_value
else:
video_length = fps_length_value
if fps_or_length == "fps":
video_fps = fps_length_value
else:
video_fps = frames_length / fps_length_value
if use_torchvision == False:
writer = imageio.get_writer(video_out_path, fps=float(video_fps))
# convert the image stack to uint8
image_stack = (image_stack * 255).to(torch.uint8)
for frame_data in image_stack:
frame = np.array(frame_data) # Convert torch tensor to numpy array if needed
frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
writer.append_data(frame)
if audio_file is not None:
difference = video_length - audio_length
needed_extra_frames = int(difference * video_fps)
if needed_extra_frames > 0:
for i in range(needed_extra_frames):
writer.append_data(frame)
writer.close()
args = [
"-i", video_out_path,
"-i", audio_file,
"-c:v", codec,
"-b:v", str(bitrate),
"-c:a", "aac",
"-strict", "experimental",
"-y",
video_out_path.replace(".mp4", "_with_audio.mp4")
]
run_ffmpeg_command(args)
else:
writer.close()
if use_torchvision:
audio_tensor = None
if audio_file is not None:
audio_tensor, aud_sr = torchaudio.load(audio_file)
fps_frac = fractions.Fraction(float(video_fps))
torchvision.io.write_video(filename=video_out_path,
video_array=image_stack * 255,
fps=fps_frac,
audio_array=audio_tensor,
video_codec=codec,
audio_codec="aac",
audio_fps=aud_sr
)
return (self, video_out_path,)
class CreateConsistentVideo(metaclass=ABCWidgetMetaclass):
"""Creates a single video from a folder of videos with a consistent rate of change"""
@classmethod
def INPUT_TYPES(cls):
return {
"required": {
"video_folder": ("STRING", {"default": ""}),
"rate_of_change_target": ("FLOAT", {"default": 0.5}),
}
}
RETURN_TYPES = ("STRING", "LIST", "STRING",)
RETURN_NAMES = ("joined_video_path", "individual_video_paths", "error_messages")
CATEGORY = "video"
OUTPUT_NODE = True
FUNCTION = "handler"
def handler(self, **kwargs):
import asyncio
import os
video_folder = kwargs["video_folder"]
rate_of_change_target = kwargs["rate_of_change_target"]
individual_video_paths = []
error_messages = []
async def process_video(file_path):
try:
# Calculate visual differences
difference = calculate_visual_difference(file_path)
# Adjust frame rate based on the rate of change target
adjusted_video_path = adjust_frame_rate(
video_path=file_path,
target_rate_of_change=rate_of_change_target,
visual_difference=difference
)
return adjusted_video_path
except Exception as e:
error_messages.append(str(e))
return None
# Gather video files
video_files = [f for f in os.listdir(video_folder) if
f.lower().endswith(('.mp4', '.mov', '.avi', '.flv', '.mkv', '.webm'))]
# now get their full path
video_files = [os.path.join(video_folder, f) for f in video_files]
# Use Asyncio to process videos in parallel
# There is no current loop so we need to create a NEW one
loop = asyncio.new_event_loop()
asyncio.set_event_loop(loop)
processed_videos = loop.run_until_complete(asyncio.gather(
*[process_video(file_path) for file_path in video_files]
))
# Combine processed videos into the final video
joined_video_path = combine_videos(processed_videos, video_folder, video_folder)
return (joined_video_path, individual_video_paths, "\n".join(error_messages),)
def calculate_visual_difference(video_path):
# Open the video file
cap = cv2.VideoCapture(video_path)
if not cap.isOpened():
raise Exception("Could not open the video file")
# Initialize variables to hold previous frame
prev_frame_tensor = None
total_difference = 0
# Iterate through frames
while True:
ret, frame = cap.read()
if not ret:
break
# Convert frame to a PyTorch tensor and normalize
frame_tensor = torch.from_numpy(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)).permute(2, 0, 1).float() / 255
# If previous frame exists, calculate the difference using SSIM
if prev_frame_tensor is not None:
difference = get_img_diff_torch(prev_frame_tensor, frame_tensor)
total_difference += difference
# Update previous frame
prev_frame_tensor = frame_tensor
# Close the video file
cap.release()
return total_difference
def get_img_diff_torch(image1: torch.Tensor, image2: torch.Tensor, K1=0.01, K2=0.03, L=255):
# Convert images to grayscale using weighted average of color channels
gray1 = 0.2989 * image1[0] + 0.5870 * image1[1] + 0.1140 * image1[2]
gray2 = 0.2989 * image2[0] + 0.5870 * image2[1] + 0.1140 * image2[2]
# Calculate mean and variance of input images
mu1 = torch.mean(gray1)
mu2 = torch.mean(gray2)
var1 = torch.var(gray1)
var2 = torch.var(gray2)
covar = np.cov(gray1.cpu().numpy().ravel(), gray2.cpu().numpy().ravel())[0][1]
# Set constants
C1 = (K1 * L) ** 2
C2 = (K2 * L) ** 2
# Calculate SSIM
numerator = (2 * mu1 * mu2 + C1) * (2 * covar + C2)
denominator = (mu1 ** 2 + mu2 ** 2 + C1) * (var1 + var2 + C2)
ssim = numerator / denominator
# Return difference score
return 1 - ssim
def adjust_frame_rate(video_path, target_rate_of_change, visual_difference):
filename, _ = os.path.splitext(os.path.basename(video_path))
base_dir = os.path.dirname(video_path)
output_filename = f"{filename}_adjusted.mp4"
output_filename = os.path.join(base_dir, output_filename)
# use the full path
cap = cv2.VideoCapture(video_path)
if not cap.isOpened():
raise Exception("Could not open the video file")
original_fps = int(cap.get(cv2.CAP_PROP_FPS))
frame_size = (int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)), int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT)))
if visual_difference == 0:
desired_fps = float(original_fps)
else:
desired_fps = max(float(original_fps * target_rate_of_change / visual_difference), 1.0)
desired_fps = int(desired_fps * 8) / 8
fourcc = cv2.VideoWriter_fourcc(*'mp4v')
out = cv2.VideoWriter(output_filename, fourcc, desired_fps, frame_size)
while True:
ret, frame = cap.read()
if not ret:
break
out.write(frame)
cap.release()
out.release()
return output_filename
def combine_videos(video_paths, base_input_directory, output_path):
import subprocess
import tempfile
# Create a temporary text file
with tempfile.NamedTemporaryFile(mode="w+t", delete=False) as temp_file:
for video_path in video_paths:
temp_file.write(f"file '{video_path}'\n")
# TODO: Check cross-platform compatibility (linux)
cmd = f"ffmpeg -f concat -safe 0 -i \"{temp_file.name}\" \"{output_path}combined{os.path.splitext(video_paths[0])[1]}\""
try:
subprocess.run(cmd, shell=True, check=True)
return output_path
except subprocess.CalledProcessError as e:
return str(e)
finally:
# Clean up the temporary file
temp_file.close()
os.remove(temp_file.name)
def transcribe_to_segs(input_filename):
import whisper as wh
import folder_paths as fp
if not os.sep in input_filename:
input_dir = fp.input_directory
full_path = os.path.join(input_dir, input_filename)
else:
full_path = input_filename
# check that the file exists
if not os.path.exists(full_path):
raise ValueError(f"File {full_path} does not exist.")
# check that it is a file
if not os.path.isfile(full_path):
raise ValueError(f"{full_path} is not a file.")
model = wh.load_model("medium.en")
res = wh.transcribe(model=model, audio=full_path, word_timestamps=True)
segs = res["segments"]
return segs
class WhisperTranscription(metaclass=ABCWidgetMetaclass):
@classmethod
def INPUT_TYPES(cls):
return {"required": {"input_filename": ("STRING", {"default": "fullpath"})}}
RETURN_TYPES = ("STRING", "STRING",)
RETURN_NAMES = ("lines_out", "words_out",)
FUNCTION = "transcribe_movie"
CATEGORY = "transcription"
def transcribe_movie(self, input_filename):
full_path = os.path.abspath(input_filename)
try:
segs = transcribe_to_segs(full_path)
lines_out = []
for seg in segs:
lines_out.append(f"{seg['start']:3.3f} - {seg['end']:3.3f}: {seg['text']}")
lines_out = '\n'.join(lines_out)
words_out = []
for seg in segs:
words_out.append([])
for word in seg['words']:
line = f"{word['start']:3.3f} - {word['end']:3.3f}: {word['word']}"
words_out[-1].append(line)
words_out = '\n'.join(['\n'.join(x) for x in words_out])
return (lines_out, words_out,)
except Exception as e:
print(e)
return ("Error", str(e))
class WhisperTranscribeToFrameIDXandWords(metaclass=ABCWidgetMetaclass):
"""Just like WhisperTranscription, but returns a list of frame numbers and words instead of a string."""
@classmethod
def INPUT_TYPES(cls):
# needs full path to video file
# needs fps for conversion
ret = {"required": {"input_filename": ("STRING", {"default": "fullpath"}),
"fps": ("INT", {"default": 30})}}
return ret
RETURN_TYPES = ("LIST", "LIST", "LIST",)
RETURN_NAMES = ("frame_idx", "words_out", "combined",)
FUNCTION = "transcribe_movie"
CATEGORY = "transcription"
def transcribe_movie(self, input_filename, fps):
import folder_paths as fp
input_dir = fp.input_directory
full_path = os.path.join(input_dir, input_filename)
try:
segs = transcribe_to_segs(full_path)
frame_idx = []
words_out = []
combined = []
for seg in segs:
for word in seg['words']:
start_frame = int(word['start'] * fps)
end_frame = int(word['end'] * fps)
word_text = word['word']
frame_idx.append((start_frame, end_frame))
words_out.append(word_text)
combined.append((start_frame, end_frame, word_text))
# now we fill in the gaps so the entire video is covered by some word
# we do this by finding if there is a gap then changing the frame stamps of the last word and the next word
# to cover the gap
for i in range(len(frame_idx) - 1):
if frame_idx[i][1] < frame_idx[i + 1][0]:
# there is a gap
gap_len = frame_idx[i + 1][0] - frame_idx[i][1]
to_previous_word = gap_len // 2
to_next_word = gap_len - to_previous_word
frame_idx[i] = (frame_idx[i][0], frame_idx[i][1] + to_previous_word)
frame_idx[i + 1] = (frame_idx[i + 1][0] - to_next_word, frame_idx[i + 1][1])
combined[i] = (combined[i][0], combined[i][1] + to_previous_word, combined[i][2])
combined[i + 1] = (combined[i + 1][0] - to_next_word, combined[i + 1][1], combined[i + 1][2])
# leave words_out alone
return (frame_idx, words_out, combined,)
except Exception as e:
print(e)
return ("Error",)
# @profile
def get_text_frames(word_idx, words_data, x: callable, y: callable, size: (), text_kwargs: dict, style="default"):
from .image import TextRender as text_render_class
text_render = text_render_class().render_text
start = words_data[word_idx][0]
end = words_data[word_idx][1]
word_text = words_data[word_idx][2]
frame_length = end - start
# if x and y are not callables
if not callable(x):
x = lambda i: x
if not callable(y):
y = lambda i: y
out_frames = []
for i in range(start, end):
it_dict = dict()
it_dict["text"] = word_text
it_dict["x"] = x(i / float(frame_length))
it_dict["y"] = y(i / float(frame_length))
it_dict["width"] = size[0]
it_dict["height"] = size[1]
it_dict.update(text_kwargs)
# it_dict["size"] = 128
# it_dict["color"] = "#ffffff"
# it_dict["func_only"] = False
# it_dict["stroke fill"] = "#050505"
# it_dict["stroke width"] = 5
text_img = text_render(**it_dict)
out_frames.append(text_img[0])
return out_frames
class AddTranscriptionToVideo(metaclass=ABCWidgetMetaclass):
"""
Add a transcription to a video. Requires a video and transcription data.
transcription data must be [(start_frame, end_frame, word), ...]
"""
@classmethod
def INPUT_TYPES(cls):
# needs full path to video file
# needs fps for conversion
ret = {"required": {"video_file": ("STRING", {"default": "fullpath"}),
"words": ("LIST", {"default": None}),
}
}
return ret
RETURN_TYPES = ("IMAGE", "IMAGE",)
RETURN_NAMES = ("1 fps preview", "Full Video")
FUNCTION = "add_transcription_to_video"
CATEGORY = "transcription"
# @profile
def add_transcription_to_video(self, **kwargs):
import torch
import gc
video_file = kwargs["video_file"]
words = kwargs["words"]
def create_video_frames_folder(video_file):
fnc = VideoToFramesFolderFFMPEG().handler
call_dict = dict()
call_dict["video_in"] = video_file
call_dict["folder_out"] = os.path.join(video_file, os.sep, "frames_tmp")
call_dict["format"] = "PNG"
tmp_video_folder = fnc(**call_dict)[0]
del fnc
del call_dict
gc.collect()
return tmp_video_folder
def get_video_frames(start, end):
fnc = VideoFramesFolderToImageStack().handler
call_dict = dict()
call_dict["use_subfolders"] = False
call_dict["by_date_or_name"] = "NAME"
call_dict["folder_in"] = os.path.join(video_file, os.sep, "frames_tmp")
call_dict["idx_slice_start"] = 0
call_dict["idx_slice_stop"] = end - start
call_dict["slice_idx_step"] = 1
call_dict["idx"] = start
call_dict["use_float16"] = False
ret = fnc(**call_dict)[0]
return ret
def overlay_text_on_video(video_frames, text_frames):
import torch
# Create a mask for the text frames (where text is black)
text_mask = (text_frames == 0) * 1 # Assuming black is represented as 0 in text_frames
# Use the ~ operator to invert the boolean mask
# inverted_mask = text_mask
# Overlay text_frames on video_frames using the inverted_mask
overlaid_frames = video_frames * text_mask + text_frames
return overlaid_frames
frames_out = []
num_frames = 0
frames_folder = create_video_frames_folder(video_file)
for w_idx in range(len(words)):
word_start = words[w_idx][0]
word_end = words[w_idx][1]
word = words[w_idx][2].strip()
print(f"word_start: {word_start}, word_end: {word_end}, word: {word}")
words_video_frames = get_video_frames(word_start, word_end)
i_width = words_video_frames.shape[2]
i_height = words_video_frames.shape[1]
text_frames = get_text_frames(word_idx=w_idx,
words_data=words,
x=lambda i: 0,
y=lambda i: 0,
size=(i_width, i_height),
text_kwargs={"size": 128, "color": "#ffffff", "func_only": False,
"stroke fill": "#050505", "stroke width": 5,
"font_name": "c:/Windows/Fonts/arial.ttf"})
text_frames = torch.cat(text_frames)
# Overlay text on video frames
video_frames_with_text = overlay_text_on_video(words_video_frames, text_frames)
frames_out.append(video_frames_with_text)
num_frames += video_frames_with_text.shape[0]
gc.collect()
# free memory
del words_video_frames
del text_frames
del video_frames_with_text
num_stks = len(frames_out)
# write to disk
for i, stk in enumerate(frames_out):
torch.save(stk, f"tmp_str_{i}.pt")
del frames_out
gc.collect()
torch.cuda.empty_cache()
# pre allocate tensor
y = torch.zeros((num_frames, i_height, i_width, 3), dtype=torch.float32, device="cpu")
# load each save and put into tensor
new_start = 0
for i in range(num_stks):
stk = torch.load(f"tmp_str_{i}.pt")
y[new_start:new_start + stk.shape[0]] = stk
new_start = new_start + stk.shape[0]
del stk
gc.collect()
torch.cuda.empty_cache()
preview = y[::30].clone()
return (preview, y,)
class TransformImageStack(metaclass=ABCWidgetMetaclass):
"""
Transform a stack of images
given x change (in pixels)
given y change (in pixels)
"""
# TODO: support sheer,rotate,scale
# TODO: support transparancy
# TODO: support fill color specification
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"ImageStack": ("IMAGE", {"default": None}),
"pan x": ("INT", {"default": 0, "min": -2048}),
"pan y": ("INT", {"default": 0, "min": -2048}),
},
}
return ret
RETURN_TYPES = ("IMAGE",)
FUNCTION = "transform_image_stack"
def transform_image_stack(self, **kwargs):
import os
import torch
from PIL import Image
import torchvision.transforms as transforms
import torchvision.transforms.functional as TF
x_image = kwargs["ImageStack"]
x_image = x_image.detach().clone()
x = kwargs["pan x"]
y = kwargs["pan y"]
time_index, H, W, C = x_image.shape # Assuming x_image has shape (time_index, H, W, C)
out_list = []
stop_x = x
stop_y = y
dx_step = stop_x / time_index - 1
dy_step = stop_y / time_index - 1
for t in range(time_index):
tx = dx_step * t
ty = dy_step * t
current_frame = x_image[t].detach().clone()
# Change from (H, W, C) to (C, H, W)
current_frame = current_frame.permute(2, 0, 1)
transformed_frame = TF.affine(current_frame, angle=0, translate=(tx, ty), scale=1,
shear=[0, 0]).detach().clone()
# Convert back to (H, W, C)
transformed_frame = transformed_frame.permute(1, 2, 0)
# Store in the list
x_image[t] = transformed_frame
return (x_image.detach().clone(),)
class PanByStep(metaclass=ABCWidgetMetaclass):
"""
Transform a stack of images
given x change (in pixels)
given y change (in pixels)
"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"ImageStack": ("IMAGE", {"default": None}),
"pan x": ("INT", {"default": 0, "min": -2048}),
"pan y": ("INT", {"default": 0, "min": -2048}),
"wrap": ("BOOLEAN", {"default": False}),
},
}
return ret
RETURN_TYPES = ("IMAGE",)
FUNCTION = "pan_by_step"
def pan_by_step(self, **kwargs):
import os
import torch
from PIL import Image
import torchvision.transforms as transforms
import torchvision.transforms.functional as TF
x_image = kwargs["ImageStack"]
x_image = x_image.detach().clone()
x = kwargs["pan x"]
y = kwargs["pan y"]
wrap = kwargs["wrap"]
time_index, H, W, C = x_image.shape # Assuming x_image has shape (time_index, H, W, C)
out_list = []
stop_x = x
stop_y = y
dx_step = 0 if x == 0 else stop_x / time_index - 1
dy_step = stop_y / time_index - 1
for t in range(time_index):
tx = dx_step * t
ty = dy_step * t
current_frame = x_image[t].detach().clone()
# Change from (H, W, C) to (C, H, W)
current_frame = current_frame.permute(2, 0, 1)
# transformed_frame = TF.affine(current_frame, angle=0, translate=(tx, ty), scale=1,
# shear=[0, 0]).detach().clone()
# If wrap is True, apply the wrapping operation
if wrap:
transformed_frame = current_frame.roll(
(int(tx) % W, int(ty) % H), dims=(2, 1)
)
else:
transformed_frame = TF.affine(current_frame, angle=0, translate=(tx, ty), scale=1,
shear=[0, 0]).detach().clone()
# Convert back to (H, W, C)
transformed_frame = transformed_frame.permute(1, 2, 0)
# Store in the list
x_image[t] = transformed_frame
return (x_image.detach().clone(),)
class VideoFolderProvider(metaclass=ABCWidgetMetaclass):
"""
given a folder path, creates the folder or empties it entirely returns the folder path as
a string but as the node type "VIDEOFOLDER"
"""
@classmethod
def INPUT_TYPES(cls):
ret = {
"required": {
"folder_path": ("STRING", {"default": ""}),
"clear_folder": ("BOOLEAN", {"default": True}),
},
"optional": {
"activator:": ("*", {"default": None})
}
}
return ret
RETURN_TYPES = ("VIDEOFOLDER",)
FUNCTION = "video_folder_provider"
def video_folder_provider(self, **kwargs):
import shutil
import os
# check if the folder_path is none
if kwargs["folder_path"] is None:
raise ValueError("folder_path must be specified")
# set folder exists flag by using the os.path.exists function
folder_exists = os.path.exists(kwargs["folder_path"])
if not folder_exists:
os.makedirs(kwargs["folder_path"])
folder_exists = True
# set files in folder flag by using the os.listdir function
files_in_folder = os.listdir(kwargs["folder_path"])
# set accessible flag by using the os.access function
accessible = os.access(kwargs["folder_path"], os.W_OK)
if not accessible:
raise ValueError("Folder for video files is not accessible")
if kwargs["clear_folder"]:
# if the folder exists, delete all files in the folder
if folder_exists:
for file in files_in_folder:
# use shutil to remove the file
fl = os.path.join(kwargs["folder_path"], file)
if os.path.isfile(fl):
os.remove(fl)
# use shutil to remove the entire path
check_this_path = file
if len(check_this_path) > 3:
shutil.rmtree(check_this_path, ignore_errors=True)
else:
raise ValueError("Folder path is too short")
# if the folder does not exist, create the folder
return (kwargs["folder_path"],)
class AudioFolderProvider(metaclass=ABCWidgetMetaclass):
"""
given a folder path, creates the folder or empties it entirely returns the folder path as
a string but as the node type "AUDIOFOLDER"
"""
@classmethod
def INPUT_TYPES(cls):
ret = {
"required": {
"folder_path": ("STRING", {"default": ""}),
"clear_folder": ("BOOLEAN", {"default": True}),
},
"optional": {
"activator:": ("*", {"default": None})
}
}
return ret
RETURN_TYPES = ("AUDIOFOLDER",)
FUNCTION = "audio_folder_provider"
def audio_folder_provider(self, **kwargs):
import shutil
import os
# check if the folder_path is none
if kwargs["folder_path"] is None:
raise ValueError("folder_path must be specified")
# set folder exists flag by using the os.path.exists function
folder_exists = os.path.exists(kwargs["folder_path"])
if not folder_exists:
os.makedirs(kwargs["folder_path"])
folder_exists = True
# set files in folder flag by using the os.listdir function
files_in_folder = os.listdir(kwargs["folder_path"])
# set accessible flag by using the os.access function
accessible = os.access(kwargs["folder_path"], os.W_OK)
if not accessible:
raise ValueError("Folder for audio files is not accessible")
if kwargs["clear_folder"]:
# if the folder exists, delete all files in the folder
if folder_exists:
for file in files_in_folder:
# use shutil to remove the file
fl = os.path.join(kwargs["folder_path"], file)
if os.path.isfile(fl):
os.remove(fl)
return (kwargs["folder_path"],)
class VideoDefinitionProvider(metaclass=ABCWidgetMetaclass):
"""
given a VIDEOFOLDER node, and inputs for height, width, and fps, creates a VIDEOFOLDERDEF variable
"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"video_folder": ("VIDEOFOLDER", {"default": None}),
"height": ("INT", {"default": 1080}),
"width": ("INT", {"default": 1920}),
"fps": ("FLOAT", {"default": 30}),
"frame format": (["JPG", "PNG"], {"default": "JPG"}),
},
}
return ret
RETURN_TYPES = ("VIDEO_FOLDER_DEF",)
FUNCTION = "video_definition_provider"
def video_definition_provider(self, **kwargs):
from addict import Dict as adict
ret = adict()
ret["video_folder"] = kwargs["video_folder"]
ret["height"] = kwargs["height"]
ret["width"] = kwargs["width"]
ret["fps"] = kwargs["fps"]
ret["frame format"] = kwargs["frame format"]
# a function that gets the frame count from the disk
ret.get_frame_count = lambda: len([f for f in os.listdir(ret["video_folder"]) if "frame_" in f])
# a function that gets the next file name and full path dont forget to add 10 zeros to the frame count
ret.get_next_file_name = lambda x=0: os.path.join(ret['video_folder'],
f"frame_{str(ret.get_frame_count() + x).zfill(10)}.{ret['frame format'].lower()}")
ret.get_last_file_name = lambda offset=0: os.path.join(ret['video_folder'],
f"frame_{str(max(ret.get_frame_count() - offset - 1, 0)).zfill(10)}.{ret['frame format'].lower()}")
ret.get_current_file_name = lambda: os.path.join(ret['video_folder'],
f"frame_{str(ret.get_frame_count() - 1).zfill(10)}.{ret['frame format'].lower()}")
def reserve_next_frame():
fn = ret.get_next_file_name()
open(fn, "w").close()
return fn
ret.reserve_next_frame = reserve_next_frame
return (ret,)
class AudioDefinitionProvider(metaclass=ABCWidgetMetaclass):
"""
given a AUDIOFOLDER node, and inputs for sample rate, and channels, creates a AUDIOFOLDERDEF variable
"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"audio_folder": ("AUDIOFOLDER", {"default": None}),
"sample rate": ("INT", {"default": 44100}),
"channels": ("INT", {"default": 2}),
"audio format": (["WAV", "MP3"], {"default": "WAV"}),
},
}
return ret
RETURN_TYPES = ("AUDIO_FOLDER_DEF",)
FUNCTION = "audio_definition_provider"
def get_num_audio_files_that_match(self, audio_folder):
audio_files = self.get_all_audio_files(audio_folder)
return len(audio_files)
def get_all_audio_files(self, audio_folder):
import os
audio_files = os.listdir(audio_folder)
audio_files = [os.path.join(audio_folder, x) for x in audio_files if "audio_" in x]
return audio_files
def audio_definition_provider(self, **kwargs):
from addict import Dict as adict
ret = adict()
ret["audio_folder"] = kwargs["audio_folder"]
ret["sample rate"] = kwargs["sample rate"]
ret["channels"] = kwargs["channels"]
ret["audio format"] = kwargs["audio format"]
# a function that gets the frame count from the disk
ret.get_file_count = lambda: self.get_num_audio_files_that_match(ret["audio_folder"])
# a function that gets the next file name and full path dont forget to zfill
ret.get_next_file_name = lambda: f"{ret['audio_folder']}{os.sep}audio_{str(ret.get_file_count()).zfill(10)}.{ret['audio format'].lower()}"
ret.get_last_file_name = lambda: f"{ret['audio_folder']}{os.sep}audio_{str(ret.get_file_count() - 1).zfill(10)}.{ret['audio format'].lower()}"
ret.get_all_audio_files = lambda: self.get_all_audio_files(ret["audio_folder"])
return (ret,)
class AddAudioToAudioFolder(metaclass=ABCWidgetMetaclass):
"""
given a AUDIO_FOLDER_DEF node, and a list of audio files, adds the audio files to the audio folder
"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"audio_folder_def": ("AUDIO_FOLDER_DEF", {"default": None}),
"AUDIO": ("LIST", {"default": None}),
"delete after": ("BOOLEAN", {"default": False}),
},
}
return ret
RETURN_TYPES = ("AUDIO_FOLDER_DEF",)
FUNCTION = "add_audio_to_audio_folder"
def add_audio_to_audio_folder(self, **kwargs):
# check each file in the list to make sure we can convert it to the correct format
import shutil
import os
from subprocess import run
ffmpeg_exe = ffmpeg.get_ffmpeg_exe()
audio_folder_def = kwargs["audio_folder_def"]
audio_files = kwargs["AUDIO"]
delete = kwargs["delete after"]
for audio_file in audio_files:
# use ffmpeg_exe to convert the audio file to the correct format which is audio_folder_def["audio format"]
# use the audio_folder_def["audio format"] to determine the output file name
# use the audio_folder_def["audio folder"] to determine the output folder
# use the audio_folder_def["sample rate"] to determine the sample rate
# use the audio_folder_def["channels"] to determine the number of channels
# get the output file name
output_file_name = audio_folder_def.get_next_file_name()
input_file_name = audio_file
CMD = f"{ffmpeg_exe} -y -i {input_file_name} -ar {audio_folder_def['sample rate']} -ac {audio_folder_def['channels']} '{output_file_name}'"
CMD_AS_LIST = [ffmpeg_exe, "-y", "-i", input_file_name, "-ar", str(audio_folder_def['sample rate']), "-ac",
str(audio_folder_def['channels']), output_file_name]
run(CMD_AS_LIST)
if delete:
os.remove(input_file_name)
return (audio_folder_def,)
class AddSilenceToAudioFolder(metaclass=ABCWidgetMetaclass):
"""create and add a new audio file that is silence for the specified duration"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"audio_folder_def": ("AUDIO_FOLDER_DEF", {"default": None}),
"duration": ("FLOAT", {"default": 1.0}),
},
}
return ret
RETURN_TYPES = ("AUDIO_FOLDER_DEF",)
FUNCTION = "add_silence_to_audio_folder"
def add_silence_to_audio_folder(self, **kwargs):
# this is to be multi-platform
import shutil
import os
audio_folder_def = kwargs["audio_folder_def"]
duration = kwargs["duration"]
# get the output file name
output_file_name = audio_folder_def.get_next_file_name()
sample_rate = audio_folder_def["sample rate"]
channels = audio_folder_def["channels"]
fmt = audio_folder_def["audio format"]
# create the silence audio file
if fmt == "WAV":
fmt_part = "pcm_s16le"
elif fmt == "MP3":
fmt_part = "mp3"
else:
fmt_part = "pcm_s16le"
ffmpeg_exe = ffmpeg.get_ffmpeg_exe()
# "-i anullsrc" is the source of the silence this will work on all platforms
# "-c:a" is the codec for the audio file
# "-f" is the format of the audio file, "lavfi" is equivalent to "ffmpeg filter"
# "-i anullsrc=r=44100:cl=2:d=1" is the source of the silence
# CMD = f"{ffmpeg_exe} -f lavfi -i aevalsrc=0:d={duration}:cl={channels}:s={sample_rate} -c:a {fmt_part} {output_file_name}"
# CMD = f"{ffmpeg_exe} -f lavfi -i anullsrc=channel_layout={channels}:sample_rate={sample_rate}:duration={duration} -c:a {fmt_part} {output_file_name}"
CMD = f"{ffmpeg_exe} -f lavfi -i aevalsrc=0:sample_rate={sample_rate}:channel_layout={channels},atrim=0:{duration} -c:a {fmt_part} {output_file_name}"
# now we call it in a way to that we will know if there is an error
import subprocess
# the plan is to call the command and wait for it to finish
process = subprocess.Popen(CMD, shell=True, stderr=subprocess.PIPE)
process.wait()
# check if the process was successful
if process.returncode != 0:
# if the process was not successful raise an exception
raise Exception(f"ffmpeg failed with error code {process.returncode}")
return (audio_folder_def,)
class GetAudioFilesLengthsFromAudioFolder(metaclass=ABCWidgetMetaclass):
"""Get the lengths of all audio files in a given audio folder"""
@classmethod
def INPUT_TYPES(cls):
return {"required": {"audio_folder_def": ("AUDIO_FOLDER_DEF", {"default": None})}}
RETURN_TYPES = ("AUDIO_FOLDER_DEF", "LIST",)
RETURN_NAMES = ("audio_folder_def", "[lens(seconds)]",)
FUNCTION = "get_audio_files_lengths"
def get_audio_files_lengths(self, audio_folder_def):
import os
import torchaudio
audio_folder = audio_folder_def["audio_folder"]
audio_files = os.listdir(audio_folder)
audio_files_lengths = []
for audio_file in audio_files:
if audio_file[-3:] not in ["wav", "mp3"]:
continue
audio_path = os.path.join(audio_folder, audio_file)
waveform, sample_rate = torchaudio.load(audio_path)
duration = waveform.shape[1] / sample_rate
audio_files_lengths.append(duration)
return (audio_folder_def, audio_files_lengths,)
class AudioFolderRenderAudio(metaclass=ABCWidgetMetaclass):
""" Joins all the audio files """
@classmethod
def INPUT_TYPES(cls):
return {
"required": {
"audio_folder_def": ("AUDIO_FOLDER_DEF", {"default": None}),
"output_path": ("STRING", {"multiline": False, "default": ""}),
}
}
RETURN_TYPES = ("STRING",)
RETURN_NAMES = ("output_path",)
FUNCTION = 'join_audio_files'
CATEGORY = 'ETK/audio'
def join_audio_files(self, **kwargs):
import librosa
import copy
import soundfile as sf
audio_folder_def = kwargs['audio_folder_def']
kwargs = copy.deepcopy(kwargs)
paths = audio_folder_def.get_all_audio_files()
output_path = kwargs['output_path']
# Read the first file to get metadata
audio_data_list = []
for path in paths:
audio_data, sr = librosa.load(path, sr=None)
audio_data_list.append(audio_data)
# Concatenate all audio data
joined_audio_data = np.concatenate(audio_data_list)
# TODO: need to implement writing in the correct format
# Save the concatenated audio
sf.write(output_path, joined_audio_data, sr)
return (output_path,)
import subprocess
import platform
import subprocess
import platform
import os
def duplicate_file(source_file, target_files):
"""
Duplicates a given source file into multiple target files using a single CLI command
in a non-blocking way. The base directory is automatically determined from the
source file's location.
Args:
- source_file (str): Path to the source file to be duplicated.
- target_files (list of str): A list of target file paths where the source file should be duplicated.
Returns:
- None
"""
# Determine the base directory from the source file's directory
base_dir = os.path.dirname(os.path.abspath(source_file))
# Change the current working directory to the base directory
os.chdir(base_dir)
# Get the source file name (relative to the base directory)
source_file_name = os.path.basename(source_file)
# get the relative paths of the target files
target_files = [os.path.relpath(target_file, base_dir) for target_file in target_files]
# Determine the shell command based on the operating system
copy_command = 'copy' if os.name == 'nt' else 'cp'
# Construct the command string
command_parts = [f'{copy_command} "{source_file_name}" "{target}"' for target in target_files]
command = ' && '.join(command_parts)
command = command.replace("\"", "")
command = command.replace("\'", "")
# print(len(command))
try:
# Execute the command in a non-blocking way
process = subprocess.Popen(command, shell=True, stderr=subprocess.DEVNULL)
process.wait()
except subprocess.CalledProcessError as e:
pass
except ValueError as e:
pass
except Exception as e:
pass
if "process" not in locals():
pass
if "process" not in locals() or process.returncode != 0:
# If the command is too long, split it into multiple commands
# split the command into commands that do 64 at a time
command_parts = [command_parts[i:i + 64] for i in range(0, len(command_parts), 64)]
processes = []
for command_part in command_parts:
command = ' && '.join(command_part)
process = subprocess.Popen(command, shell=True, stderr=subprocess.DEVNULL)
processes.append(process)
for process in processes:
process.wait()
return process
class AddFramesToVideoFolder(metaclass=ABCWidgetMetaclass):
"""
given a VIDEO_FOLDER_DEF node, and a list of frames, adds the frames to the video folder
"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"video_folder_def": ("VIDEO_FOLDER_DEF", {"default": None}),
"IMAGE": ("IMAGE", {"default": None}),
"Rescale method": (
["NEAREST", "BOX", "BILINEAR", "HAMMING", "BICUBIC", "LANCZOS"],
{"default": "LANCZOS"}),
"Release vram?": ("BOOLEAN", {"default": True}),
},
}
return ret
RETURN_TYPES = ("VIDEO_FOLDER_DEF",)
FUNCTION = "add_frames_to_video_folder"
@staticmethod
def save_or_copy_frame(current_frame, previous_frame, frame_fp, previous_frame_fp):
from torchvision.utils import save_image
import shutil
from PIL import Image
if previous_frame is None or not torch.equal(current_frame, previous_frame):
# If the current frame is different from the previous frame, save it to disk
current_frame = current_frame
if current_frame.dtype != torch.uint8: # handle float type frames
ndarr = current_frame.permute(2, 0, 1).mul(255).add_(0.5).clamp_(0, 255).permute(1, 2, 0).to("cpu",
torch.uint8).numpy()
else: # handle uint8 frames
ndarr = current_frame.permute(2, 0, 1).permute(1, 2, 0).numpy()
im = Image.fromarray(ndarr)
im.save(frame_fp)
previous_frame = current_frame
else:
# If the current frame is the same as the previous frame, copy the previously saved frame file
shutil.copy2(previous_frame_fp, frame_fp)
return previous_frame
@staticmethod
def process_frames(frames, rescale_method, video_folder_def):
import gc
num_frames = frames.shape[0]
previous_frame = None
frame_fp = None
frames_area_all_the_same = torch.all(frames[0] == frames)
if frames_area_all_the_same and frames.shape[0] > 1:
# get a list of file names using reservenexframe
frame_f_names = [video_folder_def.reserve_next_frame() for _ in range(num_frames)]
# save the first frame
AddFramesToVideoFolder.save_or_copy_frame(frames[0], previous_frame, frame_f_names[0], None)
# copy the first frame to the rest of the frames using os by calling the cli
duplicate_file(frame_f_names[0], frame_f_names[1:])
else:
for i in range(num_frames):
next_fn = video_folder_def.get_next_file_name()
previous_frame_fp = frame_fp
frame_fp = next_fn
current_frame = frames[i]
previous_frame = AddFramesToVideoFolder.save_or_copy_frame(current_frame, previous_frame, frame_fp,
previous_frame_fp)
del frames
gc.collect()
return None
def add_frames_to_video_folder(self, **kwargs):
from glob import glob
from PIL import Image
import os
import torch
import gc
from torchvision.transforms.functional import resize
import psutil
from custom_nodes.EternalKernelLiteGraphNodes.modules.image_utils import lanczos_resize
initial_memory = psutil.virtual_memory().available
print(f"Initial free memory: {initial_memory / (1024 * 1024):.2f} MB")
import shutil
# get the video folder path
video_folder = kwargs["video_folder_def"]["video_folder"]
# get the frame format
frame_format = kwargs["video_folder_def"]["frame format"]
# get the height
height = kwargs["video_folder_def"]["height"]
# get the width
width = kwargs["video_folder_def"]["width"]
# get the fps
fps = kwargs["video_folder_def"]["fps"]
# get the frames
frames = kwargs["IMAGE"]
# get the rescale method
rescale_method = kwargs.get("Rescale method", None)
if not rescale_method:
rescale_method = kwargs.get("Rescale_method", None)
free_vram = kwargs.get("Release vram?", None)
if not free_vram:
free_vram = kwargs.get("Release_vram_", True)
orig_frames = frames
# num_files = len(glob(os.path.join(video_folder, f"frame_*.{frame_format.lower()}")))
# rescale if needed to the correct size
if height != frames.shape[1] or width != frames.shape[2]:
# first since transforms.Resize requires a tensor of shape B,C, H, W
# we need to permute the tensor to B, C, H, W
frames = frames.permute(0, 3, 1, 2)
frames = frames.to("cuda")
# get the right value for interpolation
interpolation_methods = {
"NEAREST": Image.Resampling.NEAREST,
"BOX": Image.Resampling.BOX,
"BILINEAR": Image.Resampling.BILINEAR,
"HAMMING": Image.Resampling.HAMMING,
"BICUBIC": Image.Resampling.BICUBIC,
"LANCZOS": Image.Resampling.LANCZOS
}
rescale_type = interpolation_methods[rescale_method]
# rescale_method = transforms.Resize((height, width), interpolation=rescale_type) # this doesnt work so well lets use a different function
if rescale_type != Image.Resampling.LANCZOS:
rescale_method = lambda x: resize(x, (height, width), interpolation=rescale_type)
frames = rescale_method(frames)
else:
rescale_method = lambda x: lanczos_resize(x, (width, height))
frames = rescale_method(frames)
# now we need to permute back to B, H, W, C
frames = frames.permute(0, 2, 3, 1)
self.process_frames(frames, rescale_method, kwargs["video_folder_def"])
else:
self.process_frames(frames, rescale_method, kwargs["video_folder_def"])
gc.collect()
if free_vram:
import gc
if isinstance(frames, torch.Tensor):
frames.to("cpu").detach().numpy()
del frames
if isinstance(orig_frames, torch.Tensor):
orig_frames = orig_frames.to("cpu").detach().numpy()
del orig_frames
torch.cuda.empty_cache()
gc.collect()
# torch.cuda.empty_cache()
# gc.collect()
initial_memory = psutil.virtual_memory().available
print(f"after adding frames free memory: {initial_memory / (1024 * 1024):.2f} MB")
# return the VIDEO_FOLDER_DEF
return (kwargs["video_folder_def"],)
class RepeatFrames(metaclass=ABCWidgetMetaclass):
"""
does nothing more than repeat the frames a number of times
"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"IMAGE": ("IMAGE", {"default": None}),
"repeat": ("INT", {"default": 1}),
},
}
return ret
RETURN_TYPES = ("IMAGE",)
RETURN_NAMES = ("repeated images",)
FUNCTION = "repeat_frames"
def repeat_frames(self, **kwargs):
import torch
frames = kwargs["IMAGE"]
repeat = kwargs["repeat"]
return (frames.repeat(repeat, 1, 1, 1),)
class HoldFrameForSecsInVideoFolder(metaclass=ABCWidgetMetaclass):
"""
given a VIDEO_FOLDER_DEF node, and a list of frames, first uses RepeatFrames to repeat the frames
for the correct amount and then adds the frames to the video folder using AddFramesToVideoFolder
"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"video_folder_def": ("VIDEO_FOLDER_DEF", {"default": None}),
"IMAGE": ("IMAGE", {"default": None}),
"seconds": ("FLOAT", {"default": 1.0}),
"Rescale method": (
["NEAREST", "BOX", "BILINEAR", "HAMMING", "BICUBIC", "LANCZOS"],
{"default": "LANCZOS"}),
"Release vram?": ("BOOLEAN", {"default": True}),
},
}
return ret
RETURN_TYPES = ("VIDEO_FOLDER_DEF",)
FUNCTION = "hold_frame_for_secs_in_video_folder"
def saver(self, x, kwargs):
# this just saves in an async way
rescale_method = kwargs.get("Rescale method", None)
if rescale_method is None:
rescale_method = kwargs.get("Rescale_method", None)
rel_vram = kwargs.get("Release vram?", None)
if rel_vram is None:
rel_vram = kwargs.get("Release_vram_", True)
AddFramesToVideoFolder().add_frames_to_video_folder(
video_folder_def=kwargs["video_folder_def"],
IMAGE=x,
Rescale_method=rescale_method,
Release_vram_=rel_vram
)
def hold_frame_for_secs_in_video_folder(self, **kwargs):
import torch
import asyncio
vide_folder_ref = kwargs["video_folder_def"]
video_folder = kwargs["video_folder_def"]["video_folder"]
# get the frame format
frame_format = kwargs["video_folder_def"]["frame format"]
# get the height
height = kwargs["video_folder_def"]["height"]
# get the width
width = kwargs["video_folder_def"]["width"]
# get the fps
fps = kwargs["video_folder_def"]["fps"]
# get the frames
frames = kwargs["IMAGE"]
# get the number of frames
num_frames = frames.shape[0]
# get the rescale method
rescale_method = kwargs.get("Rescale method", None)
if rescale_method is None:
rescale_method = kwargs.get("Rescale_method", None)
# get the number of seconds
seconds = kwargs["seconds"]
# get the number of frames to hold
num_frames_to_hold = int(fps * seconds)
num_frames_in_this_stack = frames.shape[0]
# get the number of repeats needed
num_repeats_needed = num_frames_to_hold // num_frames_in_this_stack
# get the number of frames left over
num_frames_left_over = num_frames_to_hold % num_frames_in_this_stack
# write the first set of frames to disk
self.saver(frames, kwargs)
# get the name of the last frame
source_last_frame_name = vide_folder_ref.get_last_file_name()
# from the last frame find the rest of the frames names/paths
source_last_frame_num = int(source_last_frame_name.split("_")[-1].split(".")[0])
source_first_frame_num = source_last_frame_num - num_frames_in_this_stack + 1
template = source_last_frame_name.replace(str(source_last_frame_num).zfill(10), "{}")
range_of_source_frame_numbers = list(range(source_first_frame_num, source_last_frame_num + 1))
source_frame_names = [template.format(str(x).zfill(10)) for x in range_of_source_frame_numbers]
target_done_frame_num = source_last_frame_num + num_frames_to_hold - num_frames_in_this_stack
num_frames_remaining = num_frames_to_hold - num_frames_in_this_stack
target_start_frame_num = source_last_frame_num + 1
target_frame_range = list(range(target_start_frame_num, target_done_frame_num + 1))
target_frame_names = [template.format(str(x).zfill(10)) for x in target_frame_range]
set_todo = {fn: [] for fn in source_frame_names}
# we can now repeat the frames
for i, target_frame_name in enumerate(target_frame_names):
# keep track of what stack frame we are on
stack_frame_idx = i % num_frames_in_this_stack
stack_frame_name_to_use = source_frame_names[stack_frame_idx]
set_todo[stack_frame_name_to_use].append(target_frame_name)
# now we can repeat the frames
for frame_name in set_todo:
src_frame = frame_name
target_frames = set_todo[src_frame]
duplicate_file(src_frame, target_frames)
return (kwargs["video_folder_def"],)
class Listed_HoldFramesForSecsInVideoFolder(metaclass=ABCWidgetMetaclass):
"""This just calls HoldFrameForSecsInVideoFolder multiple times for a list of frames and a list of durations"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"video_folder_def": ("VIDEO_FOLDER_DEF", {"default": None}),
"IMAGE": ("LIST", {"default": None}),
"seconds": ("LIST", {"default": None}),
"Rescale method": (
["NEAREST", "BOX", "BILINEAR", "HAMMING", "BICUBIC", "LANCZOS"],
{"default": "LANCZOS"}),
"Release vram?": ("BOOLEAN", {"default": True}),
},
}
return ret
RETURN_TYPES = ("VIDEO_FOLDER_DEF",)
FUNCTION = "listed_hold_frames_for_secs_in_video_folder"
def listed_hold_frames_for_secs_in_video_folder(self, **kwargs):
# the plan is to call HoldFrameForSecsInVideoFolder for each frame and duration
# and then return the video_folder_def
if "IMAGE" not in kwargs or "seconds" not in kwargs:
return (kwargs["video_folder_def"],)
for frame, duration in zip(kwargs["IMAGE"], kwargs["seconds"]):
HoldFrameForSecsInVideoFolder().hold_frame_for_secs_in_video_folder(
video_folder_def=kwargs["video_folder_def"],
IMAGE=frame,
seconds=duration,
Rescale_method=kwargs["Rescale method"],
Release_vram_=kwargs["Release vram?"])
return (kwargs["video_folder_def"],)
class VideoFolderDefInfo(metaclass=ABCWidgetMetaclass):
"""
Returns everything in the VIDEO_FOLDER_DEF, and also returns information about the status of the folder
"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"video_folder_def": ("VIDEO_FOLDER_DEF", {"default": None}),
},
}
return ret
RETURN_TYPES = ("VIDEO_FOLDER_DEF", "STRING", "INT", "INT", "FLOAT", "STRING", "INT", "LIST",)
RETURN_NAMES = ("VFD", "video_folder", "height", "width", "fps", "frame format", "num_files", "list of files")
OUTPUT_NODE = True
FUNCTION = "video_folder_def_info"
def video_folder_def_info(self, **kwargs):
import os
video_folder = kwargs["video_folder_def"]["video_folder"]
height = kwargs["video_folder_def"]["height"]
width = kwargs["video_folder_def"]["width"]
fps = kwargs["video_folder_def"]["fps"]
frame_format = kwargs["video_folder_def"]["frame format"]
files = os.listdir(video_folder)
num_files = len(files)
return (kwargs["video_folder_def"], video_folder, height, width, fps, frame_format, num_files, files,)
class FadeToInVideoFolder(metaclass=ABCWidgetMetaclass):
"""
given a VIDEO_FOLDER_DEF node, and a stack of frames, fades the frames in and adds them to the video folder
"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"video_folder_def": ("VIDEO_FOLDER_DEF", {"default": None}),
"IMAGE": ("IMAGE", {"default": None}),
"seconds": ("FLOAT", {"default": 1.0}),
"transition ms": ("FLOAT", {"default": 0.1}),
"Rescale method": (
["NEAREST", "BOX", "BILINEAR", "HAMMING", "BICUBIC", "LANCZOS"],
{"default": "LANCZOS"}),
"Release vram?": ("BOOLEAN", {"default": True}),
},
}
return ret
RETURN_TYPES = ("VIDEO_FOLDER_DEF",)
FUNCTION = "fade_to_in_video_folder"
def fade_to_in_video_folder(self, **kwargs):
from PIL import Image
import torch
# usee HoldFrameForSecsInVideoFolder to hold the frame after creating and AddFramesToVideoFolder the transition
transition_length_secs = int(kwargs["transition ms"] * kwargs["video_folder_def"]["fps"]) / 1000
hold_frame_length_secs = int(kwargs["seconds"] * kwargs["video_folder_def"]["fps"]) - transition_length_secs
transition_frame_count = int(kwargs["video_folder_def"]["fps"] * transition_length_secs)
base_frame_fn = kwargs["video_folder_def"].get_last_file_name()
# load it into a pytorch tensor (B, H, W, C)
base_frame = Image.open(base_frame_fn)
base_frame = torch.tensor(base_frame).permute(2, 0, 1).unsqueeze(0).to("cuda")
transition_stack = torch.zeros((transition_frame_count,
base_frame.shape[1],
base_frame.shape[2],
base_frame.shape[3]),
dtype=torch.float16, device="cuda")
# fade from base_frame to the first frame in the stack over the transition length
new_frame = kwargs["IMAGE"][0].to("cuda")
for i in range(transition_frame_count):
alpha = i / transition_frame_count
new_frame = (1 - alpha) * base_frame + alpha * new_frame
new_frame = new_frame.to("cpu")
transition_stack[i] = new_frame
# add to the video folder
AddFramesToVideoFolder().add_frames_to_video_folder(
video_folder_def=kwargs["video_folder_def"],
IMAGE=transition_stack,
Rescale_method=kwargs["Rescale method"],
Release_vram_=kwargs["Release vram?"])
# hold the frame
HoldFrameForSecsInVideoFolder().hold_frame_for_secs_in_video_folder(
video_folder_def=kwargs["video_folder_def"],
IMAGE=kwargs["IMAGE"],
seconds=hold_frame_length_secs,
Rescale_method=kwargs["Rescale method"],
Release_vram_=kwargs["Release vram?"])
return (kwargs["video_folder_def"],)
class ListedSVDConditioning(metaclass=ABCWidgetMetaclass):
"""
repeats SVD_img2vid_Conditioning for each set of params
"""
@classmethod
def INPUT_TYPES(cls):
from comfy_extras.nodes_video_model import SVD_img2vid_Conditioning
base = SVD_img2vid_Conditioning.INPUT_TYPES()
base["required"]["init_image"] = ("LIST", {"default": None})
# base["required"]["clip_vision"] = ("LIST", {"default": None})
base["required"]["video_frames"] = ("LIST", {"default": None})
return base
RETURN_TYPES = ("LIST", "LIST", "LIST")
RETURN_NAMES = ("positive", "negative", "latent")
FUNCTION = "listed_SVD_conditioning"
def listed_SVD_conditioning(self, **kwargs):
# just like Listed_HoldFramesForSecsInVideoFolder, we will call SVD_img2vid_Conditioning for each set of params
positives = []
negatives = []
latents = []
init_images = kwargs["init_image"]
clip_vision = kwargs["clip_vision"]
# vae
vae = kwargs["vae"]
# width and height
width = kwargs["width"]
height = kwargs["height"]
video_frames = kwargs["video_frames"]
motion_bucket_id = kwargs["motion_bucket_id"]
fps = kwargs["fps"]
augmentation_level = kwargs["augmentation_level"]
lens = [len(init_images), len(video_frames)]
# check if one length is divisible by the other
if lens[0] % lens[1] != 0 and lens[1] % lens[0] != 0:
raise ValueError("The length of init_image and video_frames must be divisible by each other")
# repeat the shorter
if lens[0] < lens[1]:
init_images = init_images * (lens[1] // lens[0])
else:
video_frames = video_frames * (lens[0] // lens[1])
# at this point the lengths should be the same
for init_image, video_frame in zip(init_images, video_frames):
pos, neg, latent = SVD_img2vid_Conditioning().encode(
clip_vision=clip_vision,
init_image=init_image,
video_frames=video_frame,
vae=vae,
width=width,
height=height,
motion_bucket_id=motion_bucket_id,
fps=fps,
augmentation_level=augmentation_level
)
positives.append(pos)
negatives.append(neg)
latents.append(latent)
return (positives, negatives, latents,)
class VideoFolderFFMPEGVideoInterpolate(metaclass=ABCWidgetMetaclass):
"""
use run_ffmpeg to interpolate the frames in the video folder
"""
@classmethod
def INPUT_TYPES(cls):
ret = {"required": {"video_folder_def": ("VIDEO_FOLDER_DEF", {"default": None}),
"last_n_frames": ("INT", {"default": 2}),
"target_n_frames": ("INT", {"default": 3}),
},
}
return ret
RETURN_TYPES = ("VIDEO_FOLDER_DEF",)
RETURN_NAMES = ("video_folder_def",)
FUNCTION = "video_folder_ffmpeg_video_interpolate"
def video_folder_ffmpeg_video_interpolate(self, **kwargs):
import os
import shutil
video_folder_def = kwargs["video_folder_def"]
last_n_frames = kwargs["last_n_frames"]
target_n_frames = kwargs["target_n_frames"]
# Get the current frame count
current_frame_count = video_folder_def.get_frame_count()
# Calculate the target frame rate for interpolation
target_fps = video_folder_def["fps"] * (target_n_frames / last_n_frames)
# Get the file names for the last n frames
last_frame_names = [video_folder_def.get_last_file_name(offset=i) for i in range(last_n_frames)]
last_frame_names = list(set(last_frame_names))
last_frame_names.sort()
# Create a temporary folder off of the video frame folder's base path
temp_folder = os.path.join(video_folder_def["video_folder"], "temp")
os.makedirs(temp_folder, exist_ok=True)
# put the input frames into the new folder
for frame_name in last_frame_names:
# do this in a way that overwrites the files and MOVES them
shutil.move(frame_name, os.path.join(temp_folder, os.path.basename(frame_name)))
# new input frame names list
last_frame_names = [os.path.join(temp_folder, os.path.basename(frame_name)) for frame_name in last_frame_names]
last_frame_names.sort()
# Create a file with the list of input frame paths in the temporary folder
input_frame_list_file = os.path.join(temp_folder, "input_frames.txt")
input_frame_list_file = os.path.abspath(input_frame_list_file)
with open(input_frame_list_file, 'w') as file:
# need to do this in a manner that puts the first frame at the top of the file.
file.write('\n'.join([f"file '{frame_name}'" for frame_name in last_frame_names]))
# Get the output frame names for the interpolated frames
# output_frame_names = [video_folder_def.reserve_next_frame() for _ in range(target_n_frames)]
output_frame_names = [video_folder_def.get_next_file_name(i) for i in range(target_n_frames)]
# delete the frames on the list
# for frame_name in output_frame_names:
# os.remove(frame_name)
temp_output_frame_pattern = os.path.join(video_folder_def["video_folder"], "temp",
"tmp_frame_%010d." + video_folder_def["frame format"].lower())
temp_output_frame_pattern = os.path.abspath(temp_output_frame_pattern)
# Create the FFmpeg command for motion interpolation using the frame list file
ffmpeg_interpolate_cmd = [
"-f", "concat", "-safe", "0",
"-i", input_frame_list_file,
"-vsync", "0", "-filter_complex",
# f"[0:v]minterpolate=fps={target_fps}:mi_mode=mci:mc_mode=aobmc:me_mode=bidir:me=ds:mb_size=16:search_param=8:vsbmc=1[v]",
# f"[0:v]framerate=fps={target_fps}:interp_start=0:interp_end=255:scene=0[v]",
# f"[0:v]minterpolate=fps={target_fps}:vsbmc=1[v]",
f"[0:v]minterpolate=fps={target_fps}:mi_mode=blend[v]",
"-map", "[v]", "-q:v", "2", "-r", str(target_fps), "-frames", str(target_n_frames),
"-threads", "0",
"-y", temp_output_frame_pattern
]
run_ffmpeg_command(ffmpeg_interpolate_cmd)
# Move the interpolated frames to the video folder with the correct names
for i, frame_name in enumerate(output_frame_names):
temp_frame_name = os.path.join(temp_folder,
f"tmp_frame_{i + 1:010d}.{video_folder_def['frame format'].lower()}")
try:
shutil.move(temp_frame_name, frame_name)
except FileNotFoundError:
pass
# Remove the temporary folder and frame list file
os.remove(input_frame_list_file)
# remove this without generating an error even if it is not empty
try:
os.rmdir(temp_folder)
except OSError:
shutil.rmtree(temp_folder)
return (video_folder_def,)
class AddMetadataNode(metaclass=ABCWidgetMetaclass):
@classmethod
def INPUT_TYPES(cls):
return {
"required": {
"input_file": ("STRING", {"default": "input.mp4"}),
"output_file": ("STRING", {"default": "output.mp4"}),
"metadata": ("STRING", {"default": "title:My Video"}),
}
}
RETURN_TYPES = ("STRING",) # This node returns a string message indicating success or failure
RETURN_NAMES = ("status_message",) # Providing a name for the return type
FUNCTION = "add_metadata"
CATEGORY = "video processing"
def add_metadata(self, input_file, output_file, metadata):
from ffmpeg import FFmpeg
import shutil
import torch # Importing torch as it seems to be a convention in your framework
# Parse the metadata from string to dictionary
# strip input and output file
input_file = input_file.strip()
output_file = output_file.strip()
# so actually if the input_file and output_file are the same
# it will by default have an error
# co if this is the case copy then input file to a tempfile then use that as the input file
import tempfile
if input_file == output_file:
# copy the input file to a temp file
with tempfile.NamedTemporaryFile(delete=False) as temp_file:
temp_file_name = temp_file.name
shutil.copy(input_file, temp_file_name)
input_file = temp_file_name
metadata_dict = {}
try:
for line in metadata.splitlines():
key, value = line.split(':', 1) # Split on the first colon
mdkey = "metadata"
# metadata_dict.append(f'metadata:{key.strip()}={value.strip()}')
metadata_dict[mdkey] = f'{key.strip()}={value.strip()}'
# Run ffmpeg to add metadata
ffmpeg = FFmpeg().option("y")
try:
# make sure to tell it to overwrite outputfile
ffmpeg.input(input_file).output(output_file, **metadata_dict).execute()
return (f"status_message : Metadata added successfully to {output_file}.",)
except Exception as e:
return (f"status_message : Error adding metadata to {output_file}.{str(e)}",)
return (f"status_message : Metadata added successfully to {output_file}.",)
except Exception as e:
return (f"status_message : Error adding metadata to {output_file}.",)
@staticmethod
def describe():
return "This node adds metadata to video files using ffmpeg. Metadata should be provided as 'key: value' pairs separated by newlines."
Reference in New Issue View Git Blame Copy Permalink