matrix1to2.py

import numpy as np
import matplotlib.pyplot as plt
import sqlite3
from tqdm import tqdm
from datetime import datetime
from random import choice
from glob import glob

import gmplot
from utils import get_route_details, get_stops_details, get_matrix1_map, get_matrix2_map
import os 
from random import choice

stops_data = get_stops_details()
routes_data = get_route_details()

matrix1_map = get_matrix1_map(routes_data)


def m1to2(matrix1, route_id):
    assert route_id in routes_data
    stops = routes_data[route_id]
    num_stops = len(stops)
    matrix2 = np.zeros([int((num_stops*(num_stops-1))/2)+1, matrix1.shape[1]])
    print(matrix2.shape)

    for i in range(1, num_stops):
        matrix2[get_matrix2_map(i-1, i, num_stops)] = matrix1[matrix1_map['map'][stops[i-1]][stops[i]]]
        for j in range(i-2, -1, -1):
            prev_time = matrix2[get_matrix2_map(j, i-1, num_stops)]
            for k in range(matrix1.shape[1]):
                if prev_time[k] == 0:
                    continue
                
                ind = int(k + matrix1[i-1][k]//10)
                assert ind < matrix1.shape[1]

                matrix2[get_matrix2_map(j, i, num_stops), ind] = prev_time[k] + matrix1[matrix1_map['map'][stops[i-1]][stops[i]], ind]
                
    return matrix2


matrix1 = np.load("./assets/processed/matrix/dbs-d-bus_movements_2020_08_05.db.npz")['matrix']

matrix2 = m1to2(matrix1, 534)

np.savez_compressed("./assets/processed/matrix2/dbs-d-bus_movements_2020_08_05.db.npz", matrix = matrix2)