Pycord is a simple python wrapper on top of ffmpeg designed to load videos into numpy ndarrays quickly and easily.
You can use Pycord in place of more complex video loader as OpenCV or Decord in your Computer Vision or Machine Learning projects.
from pycord.video_reader import VideoReader
vr = VideoReader("video.mp4")
b = vr.get_batch([5, 8])There's no documentation at the moment, take a look to the tests to find some examples of use.
The main idea of this project is to provide a decord-like interface completely written in python (hance the name pycord, a python implementation of decord).
Any contribution is welcome as long as it is implemented in pure python.
If you think pycord is slow, and you don't mind to deal directly with some low-level code, take a look to my other project iterframes! It's a python module implemented in Rust with the aim to be as fast as possible.