Feature/sg 747 support predict video full pipeline master

[Louis-Dupont]

Adding support for image folder and video as part of the base predict function.
Also adding support for predict_webcam.
This also introduced some refactoring on the Pipeline to support lazy batch processing.

[BloodAxe]

Regarding processing video/image folder, I wonder whether we can extract common logic via generators:

def process_series(self, images: Iterable[np.ndarray]) -> Iterable[Result]:
   for image in images:
      yield self.process(image)

So processing logic for a collection of images is in one place, regardless what if it is an image folder or video.

So now even if we want to process and save entire LoTR movie we can do this without storing all frames in memory:

class VideoFileIterator:
  def __init__(self, input_path):
     self.cap = cv2.VideoCapture(input_path)

  def __next__(self):
     success, frame = self.cap.grab()
     if not success:
        raise StopIteration()
     return frame

def process_video(self, input_path, output_path):
    input = VideoFileIterator(input_path)
    output = VideoWriter(output_path)
    for result in process_series(input):
        output.write_frame(result)

[shaydeci]

1 small comment then LGTM on my end.

[shaydeci]

LGTM