-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathtest_cases.py
More file actions
209 lines (193 loc) · 8.79 KB
/
test_cases.py
File metadata and controls
209 lines (193 loc) · 8.79 KB
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
import nltk
import re
import math
import numpy
from nltk.tokenize.texttiling import TextTilingTokenizer
from parameter import Parameter
from texttiling import *
from evaluation import *
from output import *
def baseline_test():
def read_target(fname):
targets = []
text = ""
with open(fname) as f:
content = f.readlines()
for line in content:
text+=line
if line[0:2] == '[[':
targets.append((line.split('[[')[1]).split(']')[0].lower())
f.close()
lowercase_text = text.lower()
tt = nltk.tokenize.texttiling.TextTilingTokenizer(w=40, k=28, demo_mode=True)
paragraph_breaks = tt._mark_paragraph_breaks(text)
text_length = len(lowercase_text)
# Remove punctuation
nopunct_text = ''.join(c for c in lowercase_text
if re.match("[a-z\-\' \n\t]", c))
nopunct_par_breaks = tt._mark_paragraph_breaks(nopunct_text)
tokseqs = tt._divide_to_tokensequences(nopunct_text)
target_boundry = find_target_boundry(tt, nopunct_text, targets)
return target_boundry
lecs = [
"data/Microeconomics/Microecon_1.train", "data/Microeconomics/Microecon_2.train",
"data/Microeconomics/Microecon_3.train", "data/Microeconomics/Microecon_4.train",
"data/Microeconomics/Microecon_5.train",
"data/Microeconomics/Microecon_18.train", "data/Microeconomics/Microecon_19.train",
"data/Microeconomics/Microecon_20.train", "data/Microeconomics/Microecon_21.train",
"data/Microeconomics/Microecon_22.train", "data/Microeconomics/Microecon_23.train",
"data/Microeconomics/Microecon_24.train", "data/Microeconomics/Microecon_25.train",
"data/Microeconomics/Microecon_26.train"]
lecs2 = [
"data/Psychology/Psych_1.train", "data/Psychology/Psych_2.train",
"data/Psychology/Psych_3.train", "data/Psychology/Psych_4.train",
"data/Psychology/Psych_5.train", "data/Psychology/Psych_6.train",
"data/Psychology/Psych_7.train", "data/Psychology/Psych_8.train",
"data/Psychology/Psych_9.train", "data/Psychology/Psych_10.train",
"data/Psychology/Psych_11.train", "data/Psychology/Psych_12.train",
"data/Psychology/Psych_13.train", "data/Psychology/Psych_14.train",
"data/Psychology/Psych_15.train", "data/Psychology/Psych_16.train",
"data/Psychology/Psych_17.train",
"data/Psychology/Psych_18.train", "data/Psychology/Psych_19.train",
"data/Psychology/Psych_20.train", "data/Psychology/Psych_21.train",
"data/Psychology/Psych_22.train", "data/Psychology/Psych_23.train",
"data/Psychology/Psych_24.train"]
lecs3 = [
"data/Engineering_Dynamics/EngDyn_1.train", "data/Engineering_Dynamics/EngDyn_2.train",
"data/Engineering_Dynamics/EngDyn_3", "data/Engineering_Dynamics/EngDyn_4",
"data/Engineering_Dynamics/EngDyn_5", "data/Engineering_Dynamics/EngDyn_6",
"data/Engineering_Dynamics/EngDyn_7", "data/Engineering_Dynamics/EngDyn_8",
"data/Engineering_Dynamics/EngDyn_9", "data/Engineering_Dynamics/EngDyn_10",
"data/Engineering_Dynamics/EngDyn_11", "data/Engineering_Dynamics/EngDyn_12",
"data/Engineering_Dynamics/EngDyn_13", "data/Engineering_Dynamics/EngDyn_14",
"data/Engineering_Dynamics/EngDyn_15", "data/Engineering_Dynamics/EngDyn_16",
"data/Engineering_Dynamics/EngDyn_17",
"data/Engineering_Dynamics/EngDyn_18"]
p=[]
r=[]
f1=[]
for lec in lecs2:
target_boundry = read_target(lec)
print (lec, len(target_boundry))
a,b,c = baseline(target_boundry)
p.append(a)
r.append(b)
f1.append(c)
print (sum(p)/len(p), sum(r)/len(r), sum(f1)/len(f1))
def test_best_setup(text, targets):
tt = nltk.tokenize.texttiling.TextTilingTokenizer(w=40, k=28, demo_mode=True)
new_text, s, ss, d, b,t = tokenize(tt, text, targets, 80, 9, True, 0, 0, 0)
plot(s, ss, d, b, t)
precision, recall, f1 = evaluate(b,t)
print(precision, recall, f1)
baseline(t)
return new_text
def test_cue_word(text, targets):
for i in range(50, 100, 10):
for j in range(7,12):
for k in [True, False]:
tt = nltk.tokenize.texttiling.TextTilingTokenizer(w=38, k=23, demo_mode=True)
new_text, s, ss, d, b,t = tokenize(tt, text, targets, i, j, k)
# plot(s, ss, d, b, t)
precision, recall, f1 = evaluate(b,t)
print('percent: ', i, 'distance: ', j, 'cue', k, precision, recall, f1)
def test_w_k(text, targets):
# test for w and k
for ww in range(30, 50, 2):
for kk in range(20,30):
tt = nltk.tokenize.texttiling.TextTilingTokenizer(w = ww, k=kk, demo_mode=True)
new_text, s, ss, d, b,t = tokenize(tt, text, targets, 70, 9)
# plot(s, ss, d, b, t)
precision, recall, f1 = evaluate(b,t)
print('w: ', ww, 'k: ', kk, precision, recall, f1)
def test_np(text, targets):
for i in range(70, 90, 10):
for j in range(7,11):
for k in [0,0.05,0.1,0.15,0.2,0.25,0.3]:
tt = nltk.tokenize.texttiling.TextTilingTokenizer(w=38, k=23, demo_mode=True)
new_text, s, ss, d, b,t = tokenize(tt, text, targets, i, j, False, k)
# plot(s, ss, d, b, t)
precision, recall, f1 = evaluate(b,t)
print('percent: ', i, 'distance: ', j, 'np_percent', k, precision, recall, f1)
def test_all(text, targets):
parameters = []
# for w in range(36, 48, 2):
# for k in range(20,30, 4):
# for percentile in range(70, 90, 10):
# for boundary_diff in range(7,10):
# for np_percent in [0]:
# for cue_percent in [0]:
for w in [40]:
for k in [28]:
for percentile in [80]:
for boundary_diff in [9]:
for np_percent in [ 0.8, 0.9, 1.0]:
for cue_percent in [0]:
# for np_percent in [0,0.05]:
# for cue_percent in [0,0.05]:
tt = nltk.tokenize.texttiling.TextTilingTokenizer(w=w, k=k, demo_mode=True)
new_text, s, ss, d, b,t = tokenize(tt, text, targets, percentile, boundary_diff, False, np_percent, cue_percent)
precision, recall, f1 = evaluate(b,t)
parameters.append(Parameter(k, w, percentile, boundary_diff, np_percent, cue_percent, precision, recall, f1))
print (w, k, percentile, boundary_diff, np_percent, cue_percent, precision, recall, f1)
sorted_para = sorted(parameters, key=lambda parameter: parameter.f1, reverse = True)
print (sorted_para)
f = open('test_all.out','w')
for para in sorted_para:
output = str(para) + '\n'
f.write(output)
f.close()
return new_text
def test_cuewords_weight(lecs):
b0 = []
t0 = []
b10 = []
t10 = []
b20 = []
t20 = []
b30 = []
t30 = []
b40 = []
t40 = []
bf = []
tf = []
for lec in lecs:
text = ""
targets = []
with open(lec) as f:
content = f.readlines()
for line in content:
text+=line
if line[0:2] == '[[':
# print (line)
targets.append((line.split('[[')[1]).split(']')[0].lower())
f.close()
tt = nltk.tokenize.texttiling.TextTilingTokenizer(w=38, k=23, demo_mode=True)
new_text, s, ss, d, b,t = tokenize(tt, text, targets, 75, 8, False, 0, 0)
b0.extend(b)
t0.extend(t)
#print t
new_text, s, ss, d, b,t = tokenize(tt, text, targets, 75, 8, False, 0, 0.1)
b10.extend(b)
t10.extend(t)
#print t
new_text, s, ss, d, b,t = tokenize(tt, text, targets, 75, 8, False, 0, 0.2)
b20.extend(b)
t20.extend(t)
new_text, s, ss, d, b,t = tokenize(tt, text, targets, 75, 8, False, 0, 0.3)
b30.extend(b)
t30.extend(t)
new_text, s, ss, d, b,t = tokenize(tt, text, targets, 75, 8, False, 0, 0.4)
b40.extend(b)
t40.extend(t)
new_text, s, ss, d, b,t = tokenize(tt, text, targets, 75, 8, True, 0, 0)
bf.extend(b)
tf.extend(t)
#print t
print 0, evaluate(b0, t0)
print 10, evaluate(b10, t10)
print 20, evaluate(b20, t20)
print 30, evaluate(b30, t30)
print 40, evaluate(b40, t40)
print "filter", evaluate(bf, tf)
# baseline_test()