import turtle,time,random
bob = turtle.Turtle()
sc = turtle.Screen()
WIDTH = 1000
HEIGHT = 500
sc.setup(WIDTH,HEIGHT)
ROWS = 10
COLUMNS = 20
START = (1,1)
END = (10,20)
centers = [[0 for i in range(COLUMNS)] for j in range(ROWS)]
if HEIGHT//ROWS < WIDTH//COLUMNS:
        side = HEIGHT//ROWS-5
else:
        side = WIDTH//COLUMNS-5
        
def square(x,y,side,fill):
    bob.penup()
    bob.setpos(x,y)
    bob.setheading(0)
    bob.pendown()
    if fill:
        bob.begin_fill()
    for i in range(4):
        bob.forward(side)
        bob.right(90)
    bob.penup()
    if fill:
        bob.end_fill()

def drawGrid(rows,columns):
    sc.tracer(1000)
    ycor = (-HEIGHT//2)+side+10
    for i in range(rows):
        xcor =  -WIDTH//2+10
        for j in range(columns):
            centers[i][j] = (xcor+side//2,ycor-side//2)
            square(xcor,ycor,side,0)
            xcor+=side
        ycor+=side

def goto(row,column):
    row = row-1
    column = column-1
    bob.setpos(centers[row][column][0],centers[row][column][1])
    
def gridsquare(row,column):
    
    row = row-1
    column = column-1
    bob.setpos(centers[row][column][0]-side//2,centers[row][column][1]+side//2)
    bob.begin_fill()
    for i in range(4):
        bob.forward(side)
        bob.right(90)
    bob.penup()
    bob.end_fill()
    
def setobstacles(number):
    obstacles = [0]*number
    n = 0
    while obstacles[-1] == 0:
        x = random.randint(1,ROWS)
        y = random.randint(1,COLUMNS)
        print(x,y)
        if (x,y) != START and (x,y) != END:
            obstacles[n] = (x,y)
            n += 1
    return obstacles

def drawobstacles(obstacles):
    
    for i in obstacles:
        print(i)
        gridsquare(i[0],i[1])
def drawstartend(START,END):
    bob.fillcolor("green")
    gridsquare(START[0],START[1])
    
    bob.fillcolor("red")
    gridsquare(END[0],END[1])

obstacles = setobstacles(30)
drawGrid(ROWS,COLUMNS)
drawobstacles(obstacles)
drawstartend(START,END)
sc.tracer(2)
bob.penup()
goto(START[0],START[1])
currentx = START[0]
currenty = START[1]
DONE = False
while not DONE:
    bob.pendown()
    candidates = [0]*4

    if currentx-1 > 0 and (currentx-1,currenty) not in obstacles:
        candidates[0]=(currentx-1,currenty)
    if currentx+1 <= ROWS and (currentx+1,currenty) not in obstacles:
        candidates[1]=(currentx+1,currenty)
    if currenty+1 <= COLUMNS and (currentx,currenty+1) not in obstacles:
        candidates[2]=(currentx,currenty+1)
    if currenty-1 > 0 and (currentx,currenty-1) not in obstacles:
        candidates[3]=(currentx,currenty-1)
    candidates = [x for x in candidates if x != 0]
    print(currentx,currenty)
    #print(candidates)
    future = random.choice(candidates)
    print(future)
    #time.sleep(0.1)
    goto(future[0],future[1])
    currentx = future[0]
    currenty = future[1]
    if currentx == END[0] and currenty == END[1]:
        DONE = True

Hello Hacker Gentlemen.

Below are the 3 pieces of code we have at the moment.  Your task is to edit the allrecipes.py code to make it more useful.  At current, it creates an enormous list of completely unsorted recipes, and is functionally equivalent to throwing a box with thousands of recipe cards into a pile on the floor.  You are now going to add some sorting to this to make it more useful.  In total, your changes will probably be no more than 2-3 lines of code in allrecipes.py.  Use regular expressions to, for example, create a separate text file for recipes that contain the word “pie” in the title.  That is just one possibility, the sorting criteria is entirely up to you!  If you would like to review/finish the lessons on the regex website, click here.

import re
#re stands for regular expressions
sentence = "The rain in Spain ComPLaiN, 8ain ain't 5ai9t ai"
# 01234567891111111
# 0123456

x = re.search("ai", sentence)

print(x.span())
# . is the metacharacter for any character (except for newline)
#[a-z] represents any lower case alpha character
#[A-Z] represents any upper case alpha character
#[0-9] represents any numeric character
#\w represents "word" characters, a-z, A-Z, 0-9, _ underscore
#\W represents anything that is NOT a "word" character
#\s string contains a whitespace character (space, tab, newline, return, feed)
#\S anything that is not a whitespace character
#* represents any number of the thing it follows
x = re.findall('[\w]*ai[\w]*',sentence)
print(x)

import requests, time, re
from recipe_scrapers import scrape_me
f = open("last_recipe_checked.txt",'rt')
content = f.readlines()  
start = int(content[-1])
f.close()
currentrecipe = start
def main():    
    validrecipes = []
    for i in range(start,9999999):
        f = open("last_recipe_checked.txt",'w')########
        f.write(str(i)+'\n')########
        f.close() ########
        url = 'https://cooking.nytimes.com/recipes/' + str(i)
        time.sleep(0.25)
        if requests.get(url).status_code == 200:
            currentrecipe = i
            recipe = scrape_me(url,wild_mode = True)
            print(recipe.title(),i)
            f = open("nytimesrecipes.txt","at")
            f.write(recipe.title() + "     " + url + '\n')
            f.close()
            validrecipes += [i]
if (__name__ == "__main__"):
    main()

 import requests, time
import requests, time
import shutil
from recipe_scrapers import scrape_me
url = 'https://cooking.nytimes.com/recipes/103'
recipe = scrape_me(url,wild_mode = True)
image_url = recipe.image()
file_name = recipe.title() + ' image.jpg'
res = requests.get(image_url, stream = True)
if res.status_code == 200:
    with open(file_name,'wb') as f:
        shutil.copyfileobj(res.raw, f)
    print('Image sucessfully Downloaded: ',file_name)
else:
    print('Image Couldn\'t be retrieved')
##################################################
from fpdf import FPDF
pdf = FPDF(orientation='P', unit = 'in', format = 'letter')
pdf.add_page()
pdf.set_xy(0.0,0.0)
pdf.set_font('Arial','B',16)
pdf.set_text_color(0,0,0)
pdf.cell(w=8.5,h=1.0, align = 'C', txt = recipe.title(), border = 0)
pdf.image(file_name, x = 2.75, y = 1, w = 3, h = 3, type = 'jpg', link = '')
pdf.set_font('Arial','',12)
pdf.set_text_color(0,0,0)
x = 4
for ingredient in recipe.ingredients():
    pdf.cell(w=3.5,h=.25, align = 'L', txt = ingredient, border = 0)
    x += 0.25
    pdf.set_xy(1.0,x)
x+=1
pdf.set_xy(1.0,x)
pdf.multi_cell(w=7,h=0.25, align = 'L', txt = recipe.instructions(), border = 0)
pdf.output(file_name+'.pdf','F')

import time
#runs scored per game
rsg = [4.26,4.75,4.19,4.52,4.28,4.36,4.41,4.39,4.54,3.24,4.49,3.82,3.92,5.19,3.99,4.67,4.56,4.66,
       5.34,3.44,4.63,3.57,4.37,4.06,4.62,4.61,4.13,4.55,4.86,3.85]
#runs allowed per game
rag = [4.6,3.98,4.25,4.74,4.94,4.58,5.33,4.26,5.31,4.4,3.37,5.06,4.28,3.28,4.27,4.23,4.29,3.74,3.29,
       4.5,4.12,5.04,4.07,3.85,4.36,4,3.93,4.47,4.29,5.45]

#is you winner or loser?
#I have included .500 teams as winner (you're welcome Chicago White Sox)
#1 means your team is losing (should be above the line), 0 means your team is winning (below the line)
w = [1,0,0,1,1,0,1,0,1,1,0,1,1,0,1,0,0,0,0,1,0,1,0,0,1,0,0,1,0,1]
teams = ['Arizona Diamondbacks','Atlanta Braves','Baltimore Orioles','Boston Red Sox','Chicago Cubs',
         'Chicago White Sox','Cincinnati Reds','Cleveland Guardians','Colorado Rockies','Detroit Tigers',
         'Houston Astros','Kansas City Royals','Los Angeles Angels','Los Angeles Dodgers','Miami Marlins',
         'Milwaukee Brewers','Minnesota Twins','New York Mets','New York Yankees','Oakland Athletics',
         'Philadelphia Phillies','Pittsburgh Pirates','San Diego Padres','Seattle Mariners',
         'San Francisco Giants','St. Louis Cardinals','Tampa Bay Rays','Texas Rangers','Toronto Blue Jays',
         'Washington Nationals']


#winner runs scored
winner_rsg = []
#loser runs scored
loser_rsg = []
#winner runs allowed
winner_rag = []
#loser runs allowed
loser_rag = []

#now separate all the data into 2 sets: winners and losers
for i in range(len(rsg)):
    if w[i] > 0:
        loser_rsg += [rsg[i]]
        loser_rag += [rag[i]]
    else:
        winner_rsg += [rsg[i]]
        winner_rag += [rag[i]]

import matplotlib.pyplot as plt
speed = 0.5 #how many seconds in between graphs
#draw a scatter plot of actual winners and losers as of 7/31/2022
def draw_scatter():
    for i, label in enumerate(teams):
        plt.annotate(label, (rsg[i]+0.02, rag[i]+0.02))
    plt.title("Separating Winning Teams By Runs Allowed and Scored")
    plt.xlabel("Runs Scored/Game")
    plt.ylabel("Runs Allowed/Game")
    plt.scatter(winner_rsg, winner_rag)
    plt.scatter(loser_rsg, loser_rag)
    plt.legend(['Winning Teams','Losing Teams'])

plt.ion()
draw_scatter()
x = list(range(30,55,1))
x = [data/10 for data in x]

#starting coefficients
#coefficients = [w0, w1, w2]

#you can basically start these coefficients with any random numbers because you haven't even looked a team yet
coeff = [0, 1, 0.5]
#now we will graph the first hypothesized line based on just a random guess
y = [(coeff[1]*data -1*coeff[0])/coeff[2] for data in x]
plt.plot(x,y) #use a line graph
plt.draw()    #draw the graph
plt.pause(speed)  #wait 1 second
plt.clf()     #clear the graph

n = 0.2 #this number determins how much the graph changes when you find a new point to classify

#now we are going to go through the real data points for all 30 of the teams
for i in range(len(rsg)):
    pointvalue = coeff[1]*rag[i] + coeff[2]*rsg[i] + coeff[0] * 1
    #these are the 2 options for properly classified points, so d = 1
    if (w[i] == 1 and pointvalue > 1) or (w[i] == 0 and pointvalue < 1):
        d = 1
    #these are the 2 options for misclassified points, so d = -1
    if (w[i] == 0 and pointvalue > 1) or (w[i] == 1 and pointvalue < 1):
        d = -1
    #if d=-1, we have found a misclassified point, so we will update the list of coefficients for the new point
    if d == -1:
        coeff[0] = coeff[0] + n * d * 1
        coeff[1] = coeff[1] + n * d * rag[i]
        coeff[2] = coeff[2] + n * d * rsg[i]
        #now replot the dividing line
        y = [(coeff[1]*data -1*coeff[0])/coeff[2] for data in x]
        draw_scatter()
        plt.plot(x,y)
        plt.draw()
        plt.pause(speed)
        #only clear if you aren't on the final data point
        if i != len(rsg)-1:
            plt.clf()
       
print(coeff)

#!/usr/bin/env python3
from math import inf as infinity
from random import choice
import platform
import time
from os import system

HUMAN = -1
COMP = +1
board = [
    [0, 0, 0],
    [0, 0, 0],
    [0, 0, 0],
]


def evaluate(state):
    """
    Function to heuristic evaluation of state.
    :param state: the state of the current board
    :return: +1 if the computer wins; -1 if the human wins; 0 draw
    """
    if wins(state, COMP):
        score = +1
    elif wins(state, HUMAN):
        score = -1
    else:
        score = 0

    return score


def wins(state, player):
    """
    This function tests if a specific player wins. Possibilities:
    * Three rows    [X X X] or [O O O]
    * Three cols    [X X X] or [O O O]
    * Two diagonals [X X X] or [O O O]
    :param state: the state of the current board
    :param player: a human or a computer
    :return: True if the player wins
    """
    win_state = [
        [state[0][0], state[0][1], state[0][2]],
        [state[1][0], state[1][1], state[1][2]],
        [state[2][0], state[2][1], state[2][2]],
        [state[0][0], state[1][0], state[2][0]],
        [state[0][1], state[1][1], state[2][1]],
        [state[0][2], state[1][2], state[2][2]],
        [state[0][0], state[1][1], state[2][2]],
        [state[2][0], state[1][1], state[0][2]],
    ]
    if [player, player, player] in win_state:
        return True
    else:
        return False


def game_over(state):
    """
    This function test if the human or computer wins
    :param state: the state of the current board
    :return: True if the human or computer wins
    """
    return wins(state, HUMAN) or wins(state, COMP)


def empty_cells(state):
    """
    Each empty cell will be added into cells' list
    :param state: the state of the current board
    :return: a list of empty cells
    """
    cells = []

    for x, row in enumerate(state):
        for y, cell in enumerate(row):
            if cell == 0:
                cells.append([x, y])

    return cells


def valid_move(x, y):
    """
    A move is valid if the chosen cell is empty
    :param x: X coordinate
    :param y: Y coordinate
    :return: True if the board[x][y] is empty
    """
    if [x, y] in empty_cells(board):
        return True
    else:
        return False


def set_move(x, y, player):
    """
    Set the move on board, if the coordinates are valid
    :param x: X coordinate
    :param y: Y coordinate
    :param player: the current player
    """
    if valid_move(x, y):
        board[x][y] = player
        return True
    else:
        return False


def minimax(state, depth, player):
    """
    AI function that choice the best move
    :param state: current state of the board
    :param depth: node index in the tree (0 <= depth <= 9),
    but never nine in this case (see iaturn() function)
    :param player: an human or a computer
    :return: a list with [the best row, best col, best score]
    """
    if player == COMP:
        best = [-1, -1, -infinity]
    else:
        best = [-1, -1, +infinity]

    if depth == 0 or game_over(state):
        score = evaluate(state)
        return [-1, -1, score]

    for cell in empty_cells(state):
        x, y = cell[0], cell[1]
        state[x][y] = player
        score = minimax(state, depth - 1, -player)
        state[x][y] = 0
        score[0], score[1] = x, y

        if player == COMP:
            if score[2] > best[2]:
                best = score  # max value
        else:
            if score[2] < best[2]:
                best = score  # min value

    return best


def clean():
    """
    Clears the console
    """
    os_name = platform.system().lower()
    if 'windows' in os_name:
        system('cls')
    else:
        system('clear')


def render(state, c_choice, h_choice):
    """
    Print the board on console
    :param state: current state of the board
    """

    chars = {
        -1: h_choice,
        +1: c_choice,
        0: ' '
    }
    str_line = '---------------'

    print('\n' + str_line)
    for row in state:
        for cell in row:
            symbol = chars[cell]
            print(f'| {symbol} |', end='')
        print('\n' + str_line)


def ai_turn(c_choice, h_choice):
    """
    It calls the minimax function if the depth < 9,
    else it choices a random coordinate.
    :param c_choice: computer's choice X or O
    :param h_choice: human's choice X or O
    :return:
    """
    depth = len(empty_cells(board))
    if depth == 0 or game_over(board):
        return

    clean()
    print(f'Computer turn [{c_choice}]')
    render(board, c_choice, h_choice)

    if depth == 9:
        x = choice([0, 1, 2])
        y = choice([0, 1, 2])
    else:
        move = minimax(board, depth, COMP)
        x, y = move[0], move[1]

    set_move(x, y, COMP)
    time.sleep(1)


def human_turn(c_choice, h_choice):
    """
    The Human plays choosing a valid move.
    :param c_choice: computer's choice X or O
    :param h_choice: human's choice X or O
    :return:
    """
    depth = len(empty_cells(board))
    if depth == 0 or game_over(board):
        return

    # Dictionary of valid moves
    move = -1
    moves = {
        1: [0, 0], 2: [0, 1], 3: [0, 2],
        4: [1, 0], 5: [1, 1], 6: [1, 2],
        7: [2, 0], 8: [2, 1], 9: [2, 2],
    }

    clean()
    print(f'Human turn [{h_choice}]')
    render(board, c_choice, h_choice)

    while move < 1 or move > 9:
        try:
            move = int(input('Use numpad (1..9): '))
            coord = moves[move]
            can_move = set_move(coord[0], coord[1], HUMAN)

            if not can_move:
                print('Bad move')
                move = -1
        except (EOFError, KeyboardInterrupt):
            print('Bye')
            exit()
        except (KeyError, ValueError):
            print('Bad choice')


def main():
    """
    Main function that calls all functions
    """
    clean()
    h_choice = ''  # X or O
    c_choice = ''  # X or O
    first = ''  # if human is the first

    # Human chooses X or O to play
    while h_choice != 'O' and h_choice != 'X':
        try:
            print('')
            h_choice = input('Choose X or O\nChosen: ').upper()
        except (EOFError, KeyboardInterrupt):
            print('Bye')
            exit()
        except (KeyError, ValueError):
            print('Bad choice')

    # Setting computer's choice
    if h_choice == 'X':
        c_choice = 'O'
    else:
        c_choice = 'X'

    # Human may starts first
    clean()
    while first != 'Y' and first != 'N':
        try:
            first = input('First to start?[y/n]: ').upper()
        except (EOFError, KeyboardInterrupt):
            print('Bye')
            exit()
        except (KeyError, ValueError):
            print('Bad choice')

    # Main loop of this game
    while len(empty_cells(board)) > 0 and not game_over(board):
        if first == 'N':
            ai_turn(c_choice, h_choice)
            first = ''

        human_turn(c_choice, h_choice)
        ai_turn(c_choice, h_choice)

    # Game over message
    if wins(board, HUMAN):
        clean()
        print(f'Human turn [{h_choice}]')
        render(board, c_choice, h_choice)
        print('YOU WIN!')
    elif wins(board, COMP):
        clean()
        print(f'Computer turn [{c_choice}]')
        render(board, c_choice, h_choice)
        print('YOU LOSE!')
    else:
        clean()
        render(board, c_choice, h_choice)
        print('DRAW!')

    exit()

main()

import itertools
import time

import numpy as np
import cv2

from moviepy.editor import VideoClip


WORLD_HEIGHT = 4
WORLD_WIDTH = 4
WALL_FRAC = .2
NUM_WINS = 5
NUM_LOSE = 10


class GridWorld:

    def __init__(self, world_height=3, world_width=4, discount_factor=.5, default_reward=-.5, wall_penalty=-.6,
                 win_reward=5., lose_reward=-10., viz=True, patch_side=120, grid_thickness=2, arrow_thickness=3,
                 wall_locs=[[1, 1], [1, 2]], win_locs=[[0, 3]], lose_locs=[[1, 3]], start_loc=[0, 0],
                 reset_prob=.2):
        self.world = np.ones([world_height, world_width]) * default_reward
        self.reset_prob = reset_prob
        self.world_height = world_height
        self.world_width = world_width
        self.wall_penalty = wall_penalty
        self.win_reward = win_reward
        self.lose_reward = lose_reward
        self.default_reward = default_reward
        self.discount_factor = discount_factor
        self.patch_side = patch_side
        self.grid_thickness = grid_thickness
        self.arrow_thickness = arrow_thickness
        self.wall_locs = np.array(wall_locs)
        self.win_locs = np.array(win_locs)
        self.lose_locs = np.array(lose_locs)
        self.at_terminal_state = False
        self.auto_reset = True
        self.random_respawn = True
        self.step = 0
        self.viz_canvas = None
        self.viz = viz
        self.path_color = (128, 128, 128)
        self.wall_color = (0, 255, 0)
        self.win_color = (0, 0, 255)
        self.lose_color = (255, 0, 0)
        self.world[self.wall_locs[:, 0], self.wall_locs[:, 1]] = self.wall_penalty
        self.world[self.lose_locs[:, 0], self.lose_locs[:, 1]] = self.lose_reward
        self.world[self.win_locs[:, 0], self.win_locs[:, 1]] = self.win_reward
        spawn_condn = lambda loc: self.world[loc[0], loc[1]] == self.default_reward
        self.spawn_locs = np.array([loc for loc in itertools.product(np.arange(self.world_height),
                                                                     np.arange(self.world_width))
                                    if spawn_condn(loc)])
        self.start_state = np.array(start_loc)
        self.bot_rc = None
        self.reset()
        self.actions = [self.up, self.left, self.right, self.down, self.noop]
        self.action_labels = ['UP', 'LEFT', 'RIGHT', 'DOWN', 'NOOP']
        self.q_values = np.ones([self.world.shape[0], self.world.shape[1], len(self.actions)]) * 1. / len(self.actions)
        if self.viz:
            self.init_grid_canvas()
            self.video_out_fpath = 'shm_dqn_gridsolver-' + str(time.time()) + '.mp4'
            self.clip = VideoClip(self.make_frame, duration=15)

    def make_frame(self, t):
        self.action()
        frame = self.highlight_loc(self.viz_canvas, self.bot_rc[0], self.bot_rc[1])
        return frame

    def check_terminal_state(self):
        if self.world[self.bot_rc[0], self.bot_rc[1]] == self.lose_reward \
                or self.world[self.bot_rc[0], self.bot_rc[1]] == self.win_reward:
            self.at_terminal_state = True
            # print('------++++---- TERMINAL STATE ------++++----')
            # if self.world[self.bot_rc[0], self.bot_rc[1]] == self.win_reward:
            #     print('GAME WON! :D')
            # elif self.world[self.bot_rc[0], self.bot_rc[1]] == self.lose_reward:
            #     print('GAME LOST! :(')
            if self.auto_reset:
                self.reset()

    def reset(self):
        # print('Resetting')
        if not self.random_respawn:
            self.bot_rc = self.start_state.copy()
        else:
            self.bot_rc = self.spawn_locs[np.random.choice(np.arange(len(self.spawn_locs)))].copy()
        self.at_terminal_state = False

    def up(self):
        action_idx = 0
        # print(self.action_labels[action_idx])
        new_r = self.bot_rc[0] - 1
        if new_r < 0 or self.world[new_r, self.bot_rc[1]] == self.wall_penalty:
            return self.wall_penalty, action_idx
        self.bot_rc[0] = new_r
        reward = self.world[self.bot_rc[0], self.bot_rc[1]]
        self.check_terminal_state()
        return reward, action_idx

    def left(self):
        action_idx = 1
        # print(self.action_labels[action_idx])
        new_c = self.bot_rc[1] - 1
        if new_c < 0 or self.world[self.bot_rc[0], new_c] == self.wall_penalty:
            return self.wall_penalty, action_idx
        self.bot_rc[1] = new_c
        reward = self.world[self.bot_rc[0], self.bot_rc[1]]
        self.check_terminal_state()
        return reward, action_idx

    def right(self):
        action_idx = 2
        # print(self.action_labels[action_idx])
        new_c = self.bot_rc[1] + 1
        if new_c >= self.world.shape[1] or self.world[self.bot_rc[0], new_c] == self.wall_penalty:
            return self.wall_penalty, action_idx
        self.bot_rc[1] = new_c
        reward = self.world[self.bot_rc[0], self.bot_rc[1]]
        self.check_terminal_state()
        return reward, action_idx

    def down(self):
        action_idx = 3
        # print(self.action_labels[action_idx])
        new_r = self.bot_rc[0] + 1
        if new_r >= self.world.shape[0] or self.world[new_r, self.bot_rc[1]] == self.wall_penalty:
            return self.wall_penalty, action_idx
        self.bot_rc[0] = new_r
        reward = self.world[self.bot_rc[0], self.bot_rc[1]]
        self.check_terminal_state()
        return reward, action_idx

    def noop(self):
        action_idx = 4
        # print(self.action_labels[action_idx])
        reward = self.world[self.bot_rc[0], self.bot_rc[1]]
        self.check_terminal_state()
        return reward, action_idx

    def qvals2probs(self, q_vals, epsilon=1e-4):
        action_probs = q_vals - q_vals.min() + epsilon
        action_probs = action_probs / action_probs.sum()
        return action_probs

    def action(self):
        # print('================ ACTION =================')
        if self.at_terminal_state:
            print('At terminal state, please call reset()')
            exit()
        # print('Start position:', self.bot_rc)
        start_bot_rc = self.bot_rc[0], self.bot_rc[1]
        q_vals = self.q_values[self.bot_rc[0], self.bot_rc[1]]
        action_probs = self.qvals2probs(q_vals)
        reward, action_idx = np.random.choice(self.actions, p=action_probs)()
        # print('End position:', self.bot_rc)
        # print('Reward:', reward)
        alpha = np.exp(-self.step / 10e9)
        self.step += 1
        qv = (1 - alpha) * q_vals[action_idx] + alpha * (reward + self.discount_factor
                                                         * self.q_values[self.bot_rc[0], self.bot_rc[1]].max())
        self.q_values[start_bot_rc[0], start_bot_rc[1], action_idx] = qv
        if self.viz:
            self.update_viz(start_bot_rc[0], start_bot_rc[1])
        if np.random.rand() < self.reset_prob:
            # print('-----> Randomly resetting to a random spawn point with probability', self.reset_prob)
            self.reset()

    def highlight_loc(self, viz_in, i, j):
        starty = i * (self.patch_side + self.grid_thickness)
        endy = starty + self.patch_side
        startx = j * (self.patch_side + self.grid_thickness)
        endx = startx + self.patch_side
        viz = viz_in.copy()
        cv2.rectangle(viz, (startx, starty), (endx, endy), (255, 255, 255), thickness=self.grid_thickness)
        return viz

    def update_viz(self, i, j):
        starty = i * (self.patch_side + self.grid_thickness)
        endy = starty + self.patch_side
        startx = j * (self.patch_side + self.grid_thickness)
        endx = startx + self.patch_side
        patch = np.zeros([self.patch_side, self.patch_side, 3]).astype(np.uint8)
        if self.world[i, j] == self.default_reward:
            patch[:, :, :] = self.path_color
        elif self.world[i, j] == self.wall_penalty:
            patch[:, :, :] = self.wall_color
        elif self.world[i, j] == self.win_reward:
            patch[:, :, :] = self.win_color
        elif self.world[i, j] == self.lose_reward:
            patch[:, :, :] = self.lose_color
        if self.world[i, j] == self.default_reward:
            action_probs = self.qvals2probs(self.q_values[i, j])
            x_component = action_probs[2] - action_probs[1]
            y_component = action_probs[0] - action_probs[3]
            magnitude = 1. - action_probs[-1]
            s = self.patch_side // 2
            x_patch = int(s * x_component)
            y_patch = int(s * y_component)
            arrow_canvas = np.zeros_like(patch)
            vx = s + x_patch
            vy = s - y_patch
            cv2.arrowedLine(arrow_canvas, (s, s), (vx, vy), (255, 255, 255), thickness=self.arrow_thickness,
                            tipLength=0.5)
            gridbox = (magnitude * arrow_canvas + (1 - magnitude) * patch).astype(np.uint8)
            self.viz_canvas[starty:endy, startx:endx] = gridbox
        else:
            self.viz_canvas[starty:endy, startx:endx] = patch

    def init_grid_canvas(self):
        org_h, org_w = self.world_height, self.world_width
        viz_w = (self.patch_side * org_w) + (self.grid_thickness * (org_w - 1))
        viz_h = (self.patch_side * org_h) + (self.grid_thickness * (org_h - 1))
        self.viz_canvas = np.zeros([viz_h, viz_w, 3]).astype(np.uint8)
        for i in range(org_h):
            for j in range(org_w):
                self.update_viz(i, j)

    def solve(self):
        if not self.viz:
            while True:
                self.action()
        else:
            self.clip.write_videofile(self.video_out_fpath, fps=460)


def gen_world_config(h, w, wall_frac=.5, num_wins=2, num_lose=3):
    n = h * w
    num_wall_blocks = int(wall_frac * n)
    wall_locs = (np.random.rand(num_wall_blocks, 2) * [h, w]).astype(np.int)
    win_locs = (np.random.rand(num_wins, 2) * [h, w]).astype(np.int)
    lose_locs = (np.random.rand(num_lose, 2) * [h, w]).astype(np.int)
    return wall_locs, win_locs, lose_locs


if __name__ == '__main__':
    wall_locs, win_locs, lose_locs = gen_world_config(WORLD_HEIGHT, WORLD_WIDTH, WALL_FRAC, NUM_WINS, NUM_LOSE)
    g = GridWorld(world_height=WORLD_HEIGHT, world_width=WORLD_WIDTH,
                  wall_locs=wall_locs, win_locs=win_locs, lose_locs=lose_locs, viz=True)
    g.solve()
    k = 0