-
Notifications
You must be signed in to change notification settings - Fork 0
/
preprocess.py
108 lines (90 loc) · 2.99 KB
/
preprocess.py
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
"""Sample analyses."""
import json
import matplotlib.pyplot as plt
from matplotlib import cm
import numpy as np
import dallinger
# b2c86911-a372-2bc2-eb6f-1179ced58e9d
# d8f46512-efcb-6f3f-969c-84a6351621b0
def data2states(data):
states = []
timestamps = []
for info in data.infos.list:
if info[10] == "state":
state = json.loads(info[13])
states.append(state)
timestamps.append(info[1])
times = np.array([np.datetime64(t).astype("int64")/1e6 for t in timestamps])
times = times - times.min()
sorted_idx = np.argsort(times)
times = np.sort(times)
states = [states[sorted_idx[i]] for i in range(len(sorted_idx))]
return (states, times)
def unique_player_ids(data):
(states, timestamps) = data2states(data)
ids = set()
for i, state in enumerate(states):
ids = ids.union(set(int(p["id"]) for p in state["players"]))
return list(ids)
def timecourse_num_players(data):
(states, times) = data2states(data)
num_players = np.zeros(len(states))
for i, state in enumerate(states):
num_players[i] = len(state["players"])
return (times, num_players)
def timecourse_player_scores(data):
(states, times) = data2states(data)
players = unique_player_ids(data)
scores = np.zeros((len(states), len(players)))
for i, state in enumerate(states):
for player in state["players"]:
scores[i, players.index(int(player["id"]))] = float(player["score"])
return (times, scores)
def timecourse_player_positions(data):
(states, times) = data2states(data)
players = unique_player_ids(data)
positions_x = np.zeros((len(states), len(players)))
positions_y = np.zeros((len(states), len(players)))
for i, state in enumerate(states):
for player in state["players"]:
player_id = players.index(int(player["id"]))
positions_x[i, player_id] = player["position"][0]
positions_y[i, player_id] = player["position"][1]
return (times, positions_x, positions_y)
# Subplots of player positions over time:
# (times, positions_x, positions_y) = timecourse_player_positions(data)
# print(times)
# print(positions_x)
# print(positions_y)
# plt.xlim([0, 48])
# plt.ylim([0, 48])
# plt.axis("off")
#
# for i in range(positions_x.shape[1]):
# plt.subplot(3, 5, i+1)
# plt.plot(positions_x[:, i], positions_y[:, i], alpha=0.50)
# plt.show()
# # Plot score per player over time.
# (t, s) = timecourse_player_scores(data)
# plt.ylim([0, np.max(s)])
# plt.xlabel("Time")
# plt.ylabel("Score")
# for column in s.T:
# plt.plot(t, column)
# plt.show()
#
# # Plot total collected resource over time.
# (t, s) = timecourse_player_scores(data)
# plt.ylim([0, 1000])
# plt.xlabel("Time")
# plt.ylabel("Total score")
# plt.plot(t, np.sum(s, axis=1))
# plt.show()
#
# # Plot the number of players over time.
# (t, n) = timecourse_num_players(data)
# plt.ylim([0, 20])
# plt.xlabel("Time")
# plt.ylabel("Number of players in game")
# plt.plot(t, n)
# plt.show()