Add respective frames to YOLO dump

[rishacha]

Description
I don't know if this feature already exists, but to train a darknet YOLOv3 model, I need to also get the frames and obj.data generated. The model gives an obj.names but there is no obj.data present and I would like to generate this for training a detector.

Can I try making a change for the same?

Proposed change

## Add the image file somewhere here?
with ZipFile(file_object, "w") as output_zip:
        for frame_annotation in annotations.group_by_frame():
            image_name = frame_annotation.name
            annotation_name = "{}.txt".format(os.path.splitext(os.path.basename(image_name))[0])
            width = frame_annotation.width
            height = frame_annotation.height

            # TODO : Add this line
'''
            output_zip.write(image_name,{}.jpg".format(os.path.splitext(os.path.basename(image_name))[0]), 
compression=Optional)
'''

            yolo_annotation = ""
            for shape in frame_annotation.labeled_shapes:
                if shape.type != "rectangle":
                    continue

                label = shape.label
                yolo_bb = convert_to_yolo((width, height), shape.points)
                yolo_bb = " ".join("{:.6f}".format(p) for p in yolo_bb)
                yolo_annotation += "{} {}\n".format(labels_ids[label], yolo_bb)
     
            output_zip.writestr(annotation_name, yolo_annotation)
        output_zip.writestr("obj.names", "\n".join(l[0] for l in sorted(labels_ids.items(), key=lambda x:x[1])))

[nmanovic]

@rishacha , please work with @zhiltsov-max who is own the feature. He is developing a new framework to export datasets in different formats to facilitate the training procedure.

[rishacha]

@zhiltsov-max , Please let me know how I can help. Thank you 👍