-
Notifications
You must be signed in to change notification settings - Fork 2
/
plotting.py
executable file
·197 lines (160 loc) · 8.45 KB
/
plotting.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
import matplotlib.pyplot as plt
import rl.graph_includes as graph_inc
import os
import collections as c
import numpy as np
def print_scatter(results_root, budget):
r = graph_inc.get_files(results_root, budget)
sizes={"axis":30, "label":32, "title":36}
names = ['ba-degree-ineq', 'ba-degree-inverse-ineq', 'cl-degree-ineq', 'cl-degree-inverse-ineq', 'er-degree-ineq', 'er-degree-inverse-ineq', 'sbm-degree-ineq', 'sbm-degree-inverse-ineq']
field_names = {"original": "orig", "baseline": "baseline", "model":"model"}
names_plot = ["PA High Degree", "PA Low Degree", "CL High Degree", "CL Low Degree", "ER High Degree", "ER Low Degree", "SBM High Degree", "SBM Low Degree"]
colors = ["b", "b", "g", "g", "c", "c", "k", "k"]
ginis = {}
utilities = c.defaultdict(int)
counts = c.defaultdict(int)
for ff_print, ff in field_names.items():
fig, ax1 = plt.subplots(figsize=(20, 20))
ax1.tick_params(axis='both', which='major', labelsize=sizes["axis"])
for kk, kk_print in zip(names, names_plot):
for k,v in r[kk]["results"].items():
if k[0] == ff and k[1].startswith("flow"):
utilities[kk_print] += v[0]
counts[kk_print] += 1
if k[0] == ff and k[1] == "gini":
ginis[kk_print] = v[0]
for i,(x,y,l,kk, cc) in enumerate(zip(utilities.values(), ginis.values(), counts.values(), ginis.keys(), colors)):
if i % 2:
mm = "o"
else:
mm = "X"
plt.scatter(x/l,y, label=kk, s=2000,marker=mm, c=cc )
plt.xlim([0,1])
plt.ylim([0,1])
plt.title("Utility vs. Gini Index by graph: {}, budget: {}".format(ff_print, budget) , fontsize=sizes["title"])
plt.ylabel("Gini Index", fontsize=sizes["axis"])
plt.xlabel("Utility", fontsize=sizes["axis"])
plt.legend(fontsize=sizes["label"])
plt.savefig(os.path.join(results_root, "scatter_" + ff_print + ".png"))
def print_budget_vs(results_root, budgets = [25, 50, 75, 100], vs="gini", model="model"):
sizes = {"axis": 30, "label": 32, "title": 36}
names = ['ba-degree-ineq', 'ba-degree-inverse-ineq', 'cl-degree-ineq', 'cl-degree-inverse-ineq',
'er-degree-ineq', 'er-degree-inverse-ineq', 'sbm-degree-ineq', 'sbm-degree-inverse-ineq']
field_names = {"original": "orig", "baseline": "baseline", "model": "model"}
names_plot = ["PA High Degree", "PA Low Degree", "CL High Degree", "CL Low Degree", "ER High Degree",
"ER Low Degree", "SBM High Degree", "SBM Low Degree"]
colors = ["b", "b", "g", "g", "c", "c", "k", "k"]
fig, ax1 = plt.subplots(figsize=(20, 20))
ax1.tick_params(axis='both', which='major', labelsize=sizes["axis"])
ginis = c.defaultdict(list)
utilities = c.defaultdict(list)
plusses = c.defaultdict(list)
for i, b in enumerate(budgets):
r = graph_inc.get_files(results_root, b)
for kk, kk_print in zip(names, names_plot):
if model == "model":
plusses[kk_print].append(r[kk]["history"]['edit_num_edits_exceeded'][-1])
else:
plusses[kk_print].append(0)
for k,v in r[kk]["results"].items():
if k[0] == model and k[1].startswith("flow"):
utilities[(kk_print,b)].append(v[0])
if k[0] == model and k[1] == "gini":
ginis[kk_print].append(v[0])
utilities_new = c.defaultdict(list)
for (kk_print, b), v in utilities.items():
utilities_new[kk_print].append(np.mean(v))
if vs == "gini":
measure = ginis
vs_print = "Gini Index"
else:
measure = utilities_new
vs_print = "Utility"
for i,(y,kk, cc, pp) in enumerate(zip(measure.values(), measure.keys(), colors, plusses.values())):
if i % 2:
mm = "o"
else:
mm = "X"
plt.plot(np.array(budgets)+ np.array(pp), y, color=cc, marker=mm, linewidth=4, markersize=18, label=kk )
plt.ylim([0,1])
plt.title("Budget vs. {} by graph: {}".format(vs_print, model) , fontsize=sizes["title"])
plt.ylabel(vs_print, fontsize=sizes["axis"])
plt.xlabel("Budget", fontsize=sizes["axis"])
plt.legend(fontsize=sizes["label"])
plt.savefig(os.path.join(results_root, "budget_vs_{}_{}.png".format(vs, model)))
def print_training_trajectory(results_root, key="edit_num_edits_exceeded", max_size=750):
sizes = {"axis": 30, "label": 32, "title": 36}
names = ['ba-degree-ineq', 'ba-degree-inverse-ineq', 'cl-degree-ineq', 'cl-degree-inverse-ineq',
'er-degree-ineq', 'er-degree-inverse-ineq', 'sbm-degree-ineq', 'sbm-degree-inverse-ineq']
names_plot = ["PA High Degree", "PA Low Degree", "CL High Degree", "CL Low Degree", "ER High Degree",
"ER Low Degree", "SBM High Degree", "SBM Low Degree"]
colors = ["b", "b", "g", "g", "c", "c", "k", "k"]
fig, ax1 = plt.subplots(figsize=(20, 20))
ax1.tick_params(axis='both', which='major', labelsize=sizes["axis"])
print_vals = {"value_diff_bw_groups": "Mean Utility Difference Between Groups",
"edit_num_edits_exceeded": "Edits Over Budget",
"value_mean_value": "Mean Utility"}
r = graph_inc.get_files(results_root, budget=100)
for i, (kk, kk_print, cc) in enumerate(zip(names, names_plot, colors)):
if i % 2:
mm = "o"
else:
mm = "X"
plt.plot(r[kk]["history"][key][0:max_size], color=cc, marker=mm, linewidth=4, markersize=1, label=kk_print)
if key != "edit_num_edits_exceeded":
plt.ylim([0, 1])
plt.title("Training trajectory: {} by graph".format(print_vals[key]), fontsize=sizes["title"])
plt.ylabel(print_vals[key], fontsize=sizes["axis"])
plt.xlabel("Epoch", fontsize=sizes["axis"])
plt.legend(fontsize=sizes["label"])
plt.savefig(os.path.join(results_root, "training_trajectory_{}.png".format(key)))
def print_budget_vs_placement(results_root, budgets = [3, 6, 9, 12, 15, 18, 21], vs="gini", model="model"):
sizes = {"axis": 30, "label": 32, "title": 36}
names = ['sbm-degree-ineq']
field_names = {"original": "orig", "baseline": "baseline", "model": "model"}
names_plot = ["PA High Degree"]
colors = ["b", "b", "g", "g", "c", "c", "k", "k"]
ginis = c.defaultdict(list)
utilities = c.defaultdict(list)
plusses = c.defaultdict(list)
for i, b in enumerate(budgets):
r = graph_inc.get_files(results_root, b)
for kk, kk_print in zip(names, names_plot):
if model == "model":
plusses[kk_print].append(r[kk]["history"]['edit_num_edits_exceeded'][-1])
else:
plusses[kk_print].append(0)
for k,v in r[kk]["results"].items():
if k[0] == model and k[1].startswith("flow"):
utilities[(kk_print,b)].append(v[0])
if k[0] == model and k[1] == "gini":
ginis[kk_print].append(v[0])
utilities_new = c.defaultdict(list)
for (kk_print, b), v in utilities.items():
utilities_new[kk_print].append(np.mean(v))
if vs == "gini":
measure = ginis
vs_print = "Gini Index"
else:
measure = utilities_new
vs_print = "Utility"
fig, ax1 = plt.subplots(figsize=(20, 20))
plt.xlabel("Budget", fontsize=sizes["axis"])
ax1.tick_params(axis='both', which='major', labelsize=sizes["axis"])
colors = ['k', 'k']
for i,(y,kk, cc, pp) in enumerate(zip(measure.values(), measure.keys(), colors, plusses.values())):
mm = "o"
plt.plot(np.array(budgets) + np.array(pp), y, color=cc, marker=mm, linewidth=4, markersize=18, label=kk)
ax2 = ax1.twinx()
for i,(y,kk, cc, pp) in enumerate(zip(utilities_new.values(), measure.keys(), colors, plusses.values())):
mm = "o"
ax2.plot(np.array(budgets) + np.array(pp), y, color='b', marker=mm, linewidth=4, markersize=18, label=kk)
ax2.tick_params(axis='both', which='major', labelsize=sizes["axis"])
ax2.tick_params(axis='y', labelcolor='b')
ax1.set_ylabel("Gini Index", fontsize=sizes["label"])
ax2.set_ylabel("Average Utility", color='b', fontsize=sizes["label"])
plt.ylim([0,1])
plt.title("Facility Placement:\nBudget vs. {} by graph: {}".format("Utility and Gini Index", kk_print) , fontsize=sizes["title"])
#plt.ylabel(vs_print, fontsize=sizes["axis"])
#plt.legend(fontsize=sizes["label"])
plt.savefig(os.path.join(results_root, "facility_{}.png".format("base")))