Object Tracking

######################### DO NOT CHANGE THIS CELL WHEN SUBMITTING ##########################

# The functions below make it easier to work with the provided data
# In the following cells you will see examples of their use
from utils.utils import get_level_info, get_video_data, display_video, download_and_replace_data

FINAL_EVALUATION_MODE = False
# While checking your solution, we will change this value to True
# The value of this flag M U S T be set to False in the solution you send us!

images, coordinates, target, path_to_images = get_video_data(level=1,video_id=0,dataset="example")
display_video(images,rescale=0.7,FINAL_EVALUATION_MODE=FINAL_EVALUATION_MODE)

# The libraries below are sufficient to complete all the tasks
# If, however, you want to use others, check whether they are available on the server (requirements.txt)
import numpy as np
import os
import matplotlib.pyplot as plt
import torch
import IPython.display
import json
import PIL
import sklearn as sk

# Helper function for loading the data
images, _, _, _ = get_video_data(level=1,video_id=0,dataset="example")

with open(os.path.join(os.getcwd(),'example_tracks','tracks_1_0.json'), 'r') as f:
    tracks = json.load(f)

for key in tracks.keys():
    tracks[key] = [tuple(el) for el in tracks[key]]

# Helper function for displaying the data
display_video(images,
                tracks=tracks,
                rescale=0.7,
                FINAL_EVALUATION_MODE=FINAL_EVALUATION_MODE)

# Downloading the data for subtasks 1, 2 and 3 (about ~646Mb); the script will run for a few minutes
# It is enough to download the data only once. On the grading server the data will already be downloaded
# The file structure will be identical to the one here
if not FINAL_EVALUATION_MODE:
    download_and_replace_data()

######################### DO NOT CHANGE THIS CELL ##########################

# Helper function for testing the algorithm
def submission_script(algorithm,level,verbose=False,dataset="valid"):
    num_videos, _ = get_level_info(level=level,dataset=dataset)
    correct = []
    exception_messages = set()
    for video_number in range(num_videos):
        _, coordinates, target, _ = get_video_data(level=level,video_id=video_number,dataset=dataset)
        try:
            prediction = algorithm(coordinates)
            if tuple(target) == tuple(prediction):
                correct.append(1)
            else:
                correct.append(0)
            if verbose:
                print(f"Video: animation_{str(video_number).zfill(4)}")
                print(f"Prediction: {prediction}")
                print(f"Target:     {target}")
                print(f"Score: {tuple(target) == tuple(prediction)}", end='\n\n')
        except Exception as e:
            correct.append(0)
            exception_messages.add(str(e))
    if verbose:
        print(f"Accuracy: {np.mean(correct)}")
        print(f"Correctness: {correct}")
    return np.sum(correct) / num_videos, correct, exception_messages

def your_algorithm_task_1(coordinates): # do not change the function name
    #TODO - implement your algorithm 

    permutation = [0,1,2]
    return permutation

# Check how your algorithm works
accuracy, correctness, _ = submission_script(
    algorithm=your_algorithm_task_1,
    level=1,
    verbose=True,
    dataset="train")

# Save your report in the variable below, so that we can later read it automatically with the checker
raport_1 = \
"""
Task report:
. 
.
.
"""