Utilities.py

import sys
import numpy as np
from tqdm import tqdm

from collections import namedtuple


Endpoint = namedtuple('Endpoint', ['id', 'lat', 'con'])
Request = namedtuple('Request', ['vid', 'eid', 'n'])
Scoring = namedtuple('Scoring', ['vid', 'cid', 'score'])


def csv_print(mat, fmt='%.5f'):
    np.savetxt(sys.stdout.buffer, mat, fmt=fmt, newline="\n")


def sort_array_with_id(arr):
    """
    Sort array ascending and keep track of ids.
    :param arr: array with values
    :return: array with tuples (id, val)
    """
    tuple_arr = [(id, arr[id]) for id in range(len(arr))]
    return sorted(tuple_arr, key=lambda t: t[1])


def compute_solution_score(videos_on_cache, requests, endpoints):
    score = 0.0
    total_req = 0

    for req in tqdm(requests, desc="Computing score"):
        # find caches that are connected to eid and have vid
        _lats = [endpoints[req.eid].lat]

        # check if video vid is on cache
        for _ci, _c in enumerate(videos_on_cache):
            if req.vid in _c:
                # check if endpoint eid is connected to cache
                for _cc in endpoints[req.eid].con:
                    if _ci == _cc[0]:
                        _lats.append(_cc[1])

        min_lat = np.min(_lats)
        diff_lat = endpoints[req.eid].lat - min_lat
        saved_lad = req.n * diff_lat
        score += saved_lad
        total_req += req.n

    return np.floor(score * 1000.0 / total_req)