-
Notifications
You must be signed in to change notification settings - Fork 12
/
plot_tools.py
301 lines (258 loc) · 10.4 KB
/
plot_tools.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
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
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
'''
UFPA - LASSE - Telecommunications, Automation and Electronics Research and Development Center - www.lasse.ufpa.br
CAVIAR - Communication Networks and Artificial Intelligence Immersed in Virtual or Augmented Reality
Ailton Oliveira, Felipe Bastos, João Borges, Emerson Oliveira, Daniel Suzuki, Lucas Matni, Rebecca Aben-Athar, Aldebaro Klautau (UFPA): [email protected]
CAVIAR: https://github.com/lasseufpa/ITU-Challenge-ML5G-PHY-RL.git
Auxiliary methods for plotting
V1.0
'''
import argparse
from typing import List
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import matplotlib as mpl
class PlotTools():
def __init__(self, model_name, dummy=True, oracle=True):
# Defines eps as a list
self.model_name = model_name
self.dummy = dummy
self.oracle = oracle
self.metrics = {
'chosen_ue':'User',
'ue_index':'ue_index',
'beam_index': 'Beam Index',
'pkts_dropped': 'Dropped packets',
'pkts_transmitted': 'Transmitted packets',
'pkts_buffered': 'Buffered packets',
'bit_rate_gbps': 'Bit Rate (Gbps)',
'channel_mag': 'Channel Magnitude',
'reward': 'Reward'
}
def read_data(self, eps):
"""
Read baseline data within an episode or an episode interval
"""
if len(eps) >= 2:
self.eps = range(eps[0], eps[1] + 1)
else:
self.eps = eps
# Read the data
dfs = []
titles = []
if self.dummy:
df_dummy = []
ep = []
for ep in self.eps:
df = pd.read_csv('./data/dummy/'+'output_dummy_ep'+ str(ep)+'.csv')
self._rename(df)
df['ep'] = ep
df_dummy.append(df)
df_dummy = pd.concat(df_dummy, axis=0)
df_dummy['title'] = 'B-Dummy'
dfs.append(df_dummy)
titles.append('B-Dummy')
if self.oracle:
df_oracle = []
for ep in self.eps:
df = pd.read_csv('./data/beamoracle/'+'output_b_beam_oracle_ep'+ str(ep)+ '.csv')
self._rename(df)
df['ep'] = ep
df_oracle.append(df)
df_oracle = pd.concat(df_oracle, axis=0)
df_oracle['title'] = 'B-BeamOracle'
dfs.append(df_oracle)
titles.append('B-BeamOracle')
df_rl = []
for ep in self.eps:
df = pd.read_csv('./data/'+'output_' + self.model_name +'_ep'+str(ep)+'.csv')
self._rename(df)
df['ep'] = ep
df_rl.append(df)
df_rl = pd.concat(df_rl, axis=0)
df_rl['title'] = 'B-A2C'
dfs.append(df_rl)
titles.append('B-A2C')
self.titles = titles
self.df_list = dfs
def plot_density(self, feature, subplot=False, definition=.05, size=4):
"""
Plot the density (KDE) graphic of given feature
"""
if subplot:
self._density(df=self.df_list, feature=feature,definition=definition, size=size)
else:
for df in self.df_list:
self._density(df=[df], feature=feature,definition=definition, size=size)
plt.show()
def plot_hist(self, feature, size=3, definition=.05):
"""
Plot a histogram of a feature
"""
for title, df in zip(self.titles, self.df_list):
binwidth = max(df[feature])*definition
sns.histplot(df, x=feature, hue='Legend', binwidth=binwidth, element='step', legend=True)
plt.title(title)
plt.show()
def plot_time(self, feature, ep=False):
"""
Plot a feature over time, if given an episode 'ep', then the plot is over the timesteps of episode 'ep'
"""
# If an episode is given, the dataframe list is restricted to that given episode
if ep:
df_list = []
for df in self.df_list:
df_list.append(self._episode(df, ep))
else:
df_list = self.df_list
rows = len(self.titles)
# Gerenate a subplot with 3 rows (or less) for each baseline/rl agent
self._plot(rows = rows, df_list = df_list, feature = feature, ylabels = feature, xlabel = 'Time Steps', titles = self.titles)
plt.show()
def plot_time_ep(self, feature):
"""
Plot a the total sum of a feature over the episodes
"""
for df, title in zip(self.df_list, self.titles):
value_per_ep = []
for ep in self.eps:
data = df.loc[df.ep == ep]
value_per_ep.append(np.sum(data[feature]))
plt.plot(self.eps, value_per_ep)
plt.title(title + '-' + feature + ' per Episode')
plt.ylabel(feature)
plt.xlabel('Episodes')
plt.show()
def plot_hist_ep(self, feature, definition=.1):
"""
Plot histogram of the occurences of total sum of a feature in all episodes
"""
mpl.style.use('seaborn-dark')
rows = len(self.titles)
fig = plt.figure(figsize=[6,6])
lim_x = []
# Gather the total sum of 'feature' for each episode
for df, title, row in zip(self.df_list, self.titles, range(rows)):
value_per_ep = []
for ep in self.eps:
# Select the related to the episode ep from the dataframe df
data = self._episode(df, ep)
value_per_ep.append(np.sum(data[feature]))
lim_x.append(value_per_ep)
# Define the bins according to the definition
bins = int(1/definition)
plt.hist(value_per_ep, bins=bins, alpha=0.5, label=title)
plt.title('Total '+ feature + ' per Episode')
plt.legend()
plt.xlabel(feature)
plt.xlim(np.min(lim_x)-30, np.max(lim_x))
plt.ylabel('Occurrence in Episodes')
plt.grid()
plt.show()
def plot_metrics(self):
"""
Plot general metrics over time steps (Reward, Transmitted packets, Dropped packets, Buffered packets)
"""
metrics = ['Reward', 'Transmitted packets', 'Dropped packets', 'Buffered packets']
for df, baseline in zip(self.df_list, self.titles):
df = self._episode(df, self.eps[0])
fig, ax = plt.subplots(nrows=4, ncols=1, figsize=(8, 8))
ax[0].set_title(baseline + ' Episode ' + str(self.eps[0]))
for i, feature in zip(range(4),metrics):
ax[i].plot(df[feature])
# Set label y for every plot
for a, feature in zip(ax.flat, metrics):
a.set(ylabel=feature)
# Hide x labels and tick labels for top plots and y ticks for right plots.
for a in ax.flat:
a.label_outer()
plt.xlabel('Time Steps')
plt.show()
def plot_cumulative(self, feature, size=4):
"""
Plot the cumulative sum of an feature over the time steps
"""
for df in self.df_list:
feature_sum = 0
cache = []
for data in df[feature]:
feature_sum += data
cache.append(feature_sum)
plt.plot(range(len(cache)), cache)
plt.legend(self.titles)
plt.ylabel('Cumulative '+feature)
plt.xlabel('Time steps')
plt.show()
def plot_all(self):
"""
Plot general metrics at once (Recommended)
"""
self.plot_cumulative( feature='Reward')
self.plot_metrics()
self.plot_density(feature='Reward')
self.plot_density(feature='Transmitted packets')
self.plot_density(feature='Dropped packets')
self.plot_density(feature='Buffered packets')
self.plot_hist_ep(feature='Reward')
def _rename(self, df):
# Change users names
legend={
'uav1': 'UAV',
'simulation_car1': "Car",
'simulation_pedestrian1': 'Pedestrian'
}
df['Legend'] = df.chosen_ue.map(legend)
return df.rename(columns=self.metrics, inplace=True)
def _plot(self, rows, df_list, feature, ylabels, xlabel, titles = [''], figsize = [6,8]):
fig, ax = plt.subplots(nrows=rows, ncols=1, figsize=figsize)
for title, df, row in zip(titles, df_list, range(rows)):
ax[row].plot(df[feature])
ax[row].set_title(title)
# Set label for y for every plot
for a in ax.flat:
a.set(ylabel=ylabels)
# Hide x labels and tick labels for top plots and y ticks for right plots.
for a in ax.flat:
a.label_outer()
plt.xlabel(xlabel)
def _episode(self, dataframe, ep):
return dataframe.loc[dataframe.ep == ep]
def _density(self, df, feature, definition=0.05, size=3):
dfs = pd.concat(df, axis=0)
g = sns.displot(dfs, x=feature, hue='Legend', row='title', kind="kde", bw_adjust=definition, fill=True, legend=True, aspect=2, height=size)
g.set_titles("{row_name}")
'''
The file provides a class with several plot utilities.
However, the user can also call from the terminal to perform
a quick plot.
Usage:
$ python3 plot_tools.py -m <model_name> -ep <test_ep_id#first> <test_ep_id#last> --dummy --oracle
Example:
$ python3 plot_tools.py -m baseline -ep 0 1 --dummy --oracle
'''
parser = argparse.ArgumentParser()
parser.add_argument("--model", "-m",
help="Pass RL model name",
action="store",
dest="model",
type=str)
parser.add_argument("--episode", "-ep",
nargs='+',
help="IDs of the first and " +
"last episodes to test",
action="store",
dest="episode",
type=int)
parser.add_argument("--dummy", "-d",
help="If activated, plots dummy output as well",
action="store_true",
dest="dummy")
parser.add_argument("--oracle", "-o",
help="If activated, plots oracle output as well",
action="store_true",
dest="oracle")
args = parser.parse_args()
plotter = PlotTools(model_name=args.model, dummy=args.dummy, oracle=args.oracle)
plotter.read_data(eps=args.episode)
plotter.plot_all()