Join videos quickly row-wise or column-wise using Python

join_videos.py

import cv2
import argparse
import glob
from torchvision.io import write_video
import numpy as np
import os

"""
A quick piece of code I put together to join some video samples I had for creating visualizations. 

Limitations: 
Cannot handle different aspect ratio videos (fix should be easy -- you can just rescale to fixed aspect ratio)

Args:

dir: path to directory with videos
row_vids: maximum no. of videos in a single row

Output: 

A merged video made of n x row_vids  grid of passed video snippets.

"""


parser = argparse.ArgumentParser()
parser.add_argument('--dir', type=str, default='')
parser.add_argument('--row_vids', type=int, default=10)

def merge(vids_frames, type="horizontal"):
	svf = []
	for fi in range(len(vids_frames[0])):
		sf = []

		for vi in range(len(vids_frames)):
			if fi < len(vids_frames[vi]):
				sf.append(vids_frames[vi][fi])
			else:
				# handle when videos are different lengths
				sf.append(vids_frames[vi][-1])

		if type == "horizontal":
			# stack horizontally 
			sf = np.concatenate(sf, axis=1)
		else:
			# stack vertically 
			sf = np.concatenate(sf, axis=0)
		svf.append(sf)
	return svf

args = parser.parse_args()

vids = glob.glob(args.dir + "/**", recursive=True)
vids = [v for v in vids if v.endswith(".mp4")]

vids_frames = []
for v in vids:
	vf = []
	vc = cv2.VideoCapture(v)
	while vc.isOpened():
		success, frame = vc.read()
		if not success:
			break
		else:
			frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
			vf.append(frame)
	vids_frames.append(vf)

# concatenate horizontally
n_rows = int(np.ceil( len(vids_frames) / args.row_vids))
vids_rows = [merge(vids_frames[i*args.row_vids:(i+1)*args.row_vids]) for i in range(n_rows)]

# stack vertically
svf = merge(vids_rows, type="vertical")

write_video(os.path.join(args.dir, 'combined.mp4'), svf, fps=10)
print(f'Written combined video to: {args.dir}')