DQN-2048/plot.py at master · timohe/DQN-2048 · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
import json
import matplotlib.pyplot as plt
import os
#plotting info from https://matplotlib.org/gallery/ticks_and_spines/multiple_yaxis_with_spines.html


#####SET PARAMETERS HERE######
stepsize = 100
plot_max_instead_of_avg = False
inputname = "test"
debug = False
#####SET PARAMETERS HERE######

path = os.getcwd()

def plot():
    with open(path + '/data/'+inputname+'output.txt', "r") as infile:
      inputlist = json.load(infile)
      if (debug): print("path is: " + str(path))
      if (debug): print("list is: " + str(inputlist))
      #extract first item from list which is the configuration info

      plottitle = inputlist[0]
      inputlist.pop(0)
      calculate_data(inputlist, plottitle)


def calculate_data(inputlist, plottitle):
  number_of_steps = int(len(inputlist) / stepsize)
  if(debug):print("nr of steps is: " + str(number_of_steps))

  # EPISODES
  episodes_list = []
  for i in range(1, number_of_steps + 1):
    episodes_list.append(i * stepsize);
  if (debug):print("episode list: " + str(episodes_list))

  # Average Score
  avg_score_list = []
  for i in range(0, number_of_steps):
    sum = 0
    for j in range(0, stepsize):
      sum = sum + inputlist[stepsize * i + j][2]
    result = sum / (stepsize)
    avg_score_list.append(result)
  if (debug):print("avg_score_list: " + str(avg_score_list))

  # Average MaxValue
  avg_max_value_list = []
  for i in range(0, number_of_steps):
    sum = 0
    for j in range(0, stepsize):
      sum = sum + inputlist[stepsize * i + j][1]
    result = sum / float(stepsize)
    avg_max_value_list.append(result)
  if (debug):print("avg_max_value_list: " + str(avg_max_value_list))

  # Epsilon
  epsilon_list = []
  for i in range(0, number_of_steps):
    result = inputlist[stepsize * i][3]
    epsilon_list.append(result)
  if (debug):print("epsilon_list: " + str(epsilon_list))

  # MaxScore
  max_score_list = []
  # make now list consisting only of scores
  templist = [i[2] for i in inputlist]
  for i in range(0, number_of_steps):
    result = max(templist[i * stepsize:((i + 1) * stepsize)])
    max_score_list.append(result)
  if (debug):print("max_score: " + str(max_score_list))

  # MaxValue
  max_value_list = []
  # make now list consisting only of scores
  templist_two = [i[1] for i in inputlist]
  for i in range(0, number_of_steps):
    result = max(templist_two[i * stepsize:((i + 1) * stepsize)])
    max_value_list.append(result)
  if (debug):print("max_value_list: " + str(max_value_list))


  if(plot_max_instead_of_avg):
      plot_data(plottitle, max_x=len(inputlist)-(len(inputlist)%stepsize), para_episodes_list=episodes_list, para_value_list=max_value_list, para_score_list=max_score_list, para_epsilon_list=epsilon_list)
  else:
      plot_data(plottitle, max_x=len(inputlist)-(len(inputlist)%stepsize), para_episodes_list=episodes_list, para_value_list=avg_max_value_list, para_score_list=avg_score_list, para_epsilon_list=epsilon_list)


def plot_data(plottitle, max_x, para_episodes_list, para_value_list, para_score_list, para_epsilon_list):
    def make_patch_spines_invisible(ax):
        ax.set_frame_on(True)
        ax.patch.set_visible(False)
        for sp in ax.spines.values():
            sp.set_visible(False)


    fig, host = plt.subplots()
    fig.subplots_adjust(right=0.75)
    par1 = host.twinx()
    par2 = host.twinx()
    par2.spines["right"].set_position(("axes", 1.2))
    make_patch_spines_invisible(par2)
    par2.spines["right"].set_visible(True)

    #Score
    p1, = host.plot(para_episodes_list, para_score_list, "b-", label="Score")
    #MaxValue
    p2, = par1.plot(para_episodes_list, para_value_list, "r-", label="MaxValue")
    #Epsilon
    p3, = par2.plot(para_episodes_list, para_epsilon_list, "g-", label="Epsilon")

    #Episodes
    host.set_xlim(stepsize, max_x)
    #Score
    #host.set_ylim(0, 9000)
    host.set_ylim(0, 2500)
    if (plot_max_instead_of_avg): host.set_ylim(0, 9000)
    #MaxValue
    par1.set_ylim(6, 8)
    if (plot_max_instead_of_avg): par1.set_ylim(6, 11)
    #Epsilon
    par2.set_ylim(0, 1)


    host.set_xlabel("Episodes")
    host.set_ylabel("Score")
    par1.set_ylabel("Max_Value")
    par2.set_ylabel("Epsilon")
    host.yaxis.label.set_color(p1.get_color())
    par1.yaxis.label.set_color(p2.get_color())
    par2.yaxis.label.set_color(p3.get_color())
    tkw = dict(size=4, width=1.5)
    host.tick_params(axis='y', colors=p1.get_color(), **tkw)
    par1.tick_params(axis='y', colors=p2.get_color(), **tkw)
    par2.tick_params(axis='y', colors=p3.get_color(), **tkw)
    host.tick_params(axis='x', **tkw)
    lines = [p1, p2, p3]
    host.legend(lines, [l.get_label() for l in lines])
    plt.title(plottitle, fontsize=10)
    additionalname = "_avg_"
    if (plot_max_instead_of_avg): additionalname = "_max_"

    figurepath = path + "/graphs/"
    if not os.path.exists(figurepath):
      os.makedirs(figurepath)
    fig.savefig(figurepath + str(inputname)+additionalname+"graph.pdf", bbox_inches='tight')
    plt.show()

plot()