decoder_stochastic.py

#################### COPY OF DECODER.PY WITH MINOR CHANGES FOR PROBABILTY ##########################################


import sys
import math
from random import randint

possible_directions = ['N','S','E','W']

"""
Helper function to read a file
It takes a file name a input and populates a list with the all_actions values.
It returns a list of number of states and all actions
"""

def fun_file_read(file_name):

	curr_line = 0

	all_actions = []

	with open(file_name) as f:

		for line in f:
			#iterating for all lines
			curr_line = curr_line + 1
			#splitting a line to extract values
			all_words = line.split()

			if all_words[0]=='iterations':
				number_of_states = curr_line-1

				return [number_of_states,all_actions]

				continue
			else:

				all_actions.append(int(all_words[1]))

"""
Helper function to read a grid from a file.
It takes the file name as the only argument.
It returns a grid as output.
"""

def extract_grid(file_name):

	with open(file_name) as f:
		line_no = 0
		for line in f:

			all_words = line.split()

			if line_no == 0:
				#define a matrix of size all_words * all_words
				grid = [[0 for i in range(len(all_words))] for j in range(len(all_words))]
			for col in range(len(all_words)):
				#populating the values
				grid[line_no][col] = int(all_words[col])

			line_no = line_no + 1
		return grid

"""
Function to print the path of the agent
It takes start position, end position, number of states and all actions as input
and print the path.
====> added probability
It does not explicitly return anything
"""

def extract_map(start_position, end_position, number_of_states, all_actions, probab):
		curr_state = start_position
		end_state = end_position
		
		W = int(math.sqrt(number_of_states))

		while curr_state != end_state:

			curr_action = all_actions[curr_state]

			print(possible_directions[curr_action], end = " ")

			i = int(curr_state / W)
			j = int(curr_state % W)

			probab_val_true = [True] * int(100*probab)
			probab_val_false  = [False] * (100 - int(100*probab))
			all_probab = probab_val_true + probab_val_false


			#checking for the action and assigning the nearby state with the value
			if curr_action == 0 and all_probab[randint(0, 99)] == True:

				nearby_state =  (i-1, j)

			elif curr_action == 1 and all_probab[randint(0, 99)] == True:

				nearby_state =  (i+1,j) 

			elif curr_action == 2 and all_probab[randint(0, 99)] == True:

				nearby_state =  (i,j+1)

			elif curr_action == 3 and all_probab[randint(0, 99)] == True:

				nearby_state =  (i,j-1)

			curr_state = W * nearby_state[0] + nearby_state[1]

		print(" ", end = "\n")





if __name__ == "__main__":

	input_grid = sys.argv[1]

	file_name = sys.argv[2]

	probab = sys.argv[3]

	grid = extract_grid(input_grid)

	W = len(grid[0])

	start_position = [(i,j) for i in range(W) for j in range(W) if grid[i][j]==2]

	end = [(i,j) for i in range(W) for j in range(W) if grid[i][j]==3]

	start_position = start_position[0]

	end = end[0] 
	start_position = W * start_position[0] + start_position[1]

	end = W*end[0] + end[1]

	[number_of_states,all_actions] = fun_file_read(file_name)

	extract_map(start_position,end,number_of_states,all_actions, probab)