Spaces:
Sleeping
Sleeping
hasanriaz121
commited on
Commit
•
98eb826
1
Parent(s):
14bb62b
ambiguity detection added
Browse files- .ipynb_checkpoints/app-checkpoint.py +32 -0
- .ipynb_checkpoints/detector-checkpoint.py +99 -0
- __pycache__/detector.cpython-39.pyc +0 -0
- app.py +42 -1
- coordination_encoded.pickel +0 -0
- detector.py +130 -0
- lexical_encoded.pickel +0 -0
- referential_encoded.pickel +0 -0
- scope_encoded.pickel +0 -0
- vague_encoded.pickel +0 -0
.ipynb_checkpoints/app-checkpoint.py
ADDED
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
|
3 |
+
|
4 |
+
def update_array(array_a, array_b):
|
5 |
+
# Create a dictionary to store the mappings from the first value to the second value
|
6 |
+
mapping = {}
|
7 |
+
|
8 |
+
# Populate the dictionary using the tuples from array A
|
9 |
+
for a, b in array_a:
|
10 |
+
mapping[a] = b
|
11 |
+
|
12 |
+
# Iterate through the tuples in array B
|
13 |
+
for i, (a, b) in enumerate(array_b):
|
14 |
+
if b is None and a in mapping:
|
15 |
+
# Replace the tuple in array B with the value from array A
|
16 |
+
array_b[i] = (a, mapping[a])
|
17 |
+
|
18 |
+
def amb_texts(text):
|
19 |
+
tokens = re.split(r'(\s+)', text)
|
20 |
+
# tokens = [token for token in tokens if token.strip() != '']
|
21 |
+
ambs=a.sentence_ambiguity(text)
|
22 |
+
res=list()
|
23 |
+
for i in tokens:
|
24 |
+
res.append((i,None))
|
25 |
+
update_array(ambs,res)
|
26 |
+
# print(tokens,text)
|
27 |
+
return res
|
28 |
+
def greet(name):
|
29 |
+
return "Hello " + name + "!!"
|
30 |
+
|
31 |
+
iface = gr.Interface(fn=greet, inputs="text", outputs="text")
|
32 |
+
iface.launch()
|
.ipynb_checkpoints/detector-checkpoint.py
ADDED
@@ -0,0 +1,99 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import nltk
|
2 |
+
from nltk.tokenize import word_tokenize
|
3 |
+
from sentence_transformers import SentenceTransformer, util
|
4 |
+
import numpy
|
5 |
+
# from nltk.stem import WordNetLemmatizer
|
6 |
+
import pickle
|
7 |
+
import re
|
8 |
+
|
9 |
+
nltk.download('punkt')
|
10 |
+
|
11 |
+
|
12 |
+
class AmbguityDetector:
|
13 |
+
|
14 |
+
def __init__(self):
|
15 |
+
self.model = SentenceTransformer(
|
16 |
+
'sentence-transformers/all-MiniLM-L6-v2')
|
17 |
+
|
18 |
+
def sentence_ambiguity(self, sentence):
|
19 |
+
|
20 |
+
model = self.model
|
21 |
+
tokens = word_tokenize(sentence)
|
22 |
+
filtered_tokens = list()
|
23 |
+
for token in tokens:
|
24 |
+
if token not in stopwords_custom:
|
25 |
+
filtered_tokens.append(token)
|
26 |
+
|
27 |
+
for i in filtered_tokens:
|
28 |
+
filtered_tokens[filtered_tokens.index(i)] = i.lower()
|
29 |
+
if i in punctuation:
|
30 |
+
filtered_tokens.remove(i)
|
31 |
+
|
32 |
+
lexical = dict()
|
33 |
+
scope = dict()
|
34 |
+
referential = dict()
|
35 |
+
vague = dict()
|
36 |
+
coordination = dict()
|
37 |
+
ambiguity = dict()
|
38 |
+
ambiguous_words = list()
|
39 |
+
words_set=list()
|
40 |
+
|
41 |
+
for i in filtered_tokens:
|
42 |
+
temp = model.encode(i, convert_to_tensor=True)
|
43 |
+
for j in lexical_AMB:
|
44 |
+
temp2 = lexical_encoded[j]
|
45 |
+
cos_sim = util.pytorch_cos_sim(
|
46 |
+
temp, temp2).numpy().reshape([1, ])
|
47 |
+
if(cos_sim[0] >= 0.6):
|
48 |
+
ambiguous_words.append(i)
|
49 |
+
words_set.append((i,"lexical"))
|
50 |
+
lexical[i+"+"+j] = cos_sim[0]
|
51 |
+
|
52 |
+
for j in scope_AMB:
|
53 |
+
temp2 = scope_encoded[j]
|
54 |
+
cos_sim = util.pytorch_cos_sim(
|
55 |
+
temp, temp2).numpy().reshape([1, ])
|
56 |
+
if(cos_sim[0] >= 0.6):
|
57 |
+
ambiguous_words.append(i)
|
58 |
+
words_set.append((i,"scope"))
|
59 |
+
scope[i+"+"+j] = cos_sim[0]
|
60 |
+
|
61 |
+
for j in referential_AMB:
|
62 |
+
temp2 = referential_encoded[j]
|
63 |
+
cos_sim = util.pytorch_cos_sim(
|
64 |
+
temp, temp2).numpy().reshape([1, ])
|
65 |
+
if(cos_sim[0] >= 0.6):
|
66 |
+
ambiguous_words.append(i)
|
67 |
+
words_set.append((i,"referential"))
|
68 |
+
referential[i+"+"+j] = cos_sim[0]
|
69 |
+
|
70 |
+
for j in vague_AMB:
|
71 |
+
temp2 = vague_encoded[j]
|
72 |
+
cos_sim = util.pytorch_cos_sim(
|
73 |
+
temp, temp2).numpy().reshape([1, ])
|
74 |
+
if(cos_sim[0] >= 0.6):
|
75 |
+
ambiguous_words.append(i)
|
76 |
+
words_set.append((i,"vague"))
|
77 |
+
vague[i+"+"+j] = cos_sim[0]
|
78 |
+
|
79 |
+
for j in coordination_AMB:
|
80 |
+
temp2 = coordination_encoded[j]
|
81 |
+
cos_sim = util.pytorch_cos_sim(
|
82 |
+
temp, temp2).numpy().reshape([1, ])
|
83 |
+
if(cos_sim[0] >= 0.6):
|
84 |
+
ambiguous_words.append(i)
|
85 |
+
words_set.append((i,"coordination"))
|
86 |
+
coordination[i+"+"+j] = cos_sim[0]
|
87 |
+
|
88 |
+
ambiguous_words = list(dict.fromkeys(ambiguous_words))
|
89 |
+
ambiguity["lexical"] = lexical
|
90 |
+
ambiguity["referential"] = referential
|
91 |
+
ambiguity["scope"] = scope
|
92 |
+
ambiguity["vague"] = vague
|
93 |
+
ambiguity["coordination"] = coordination
|
94 |
+
ambiguity["words"] = ambiguous_words
|
95 |
+
ambiguity["lexical_st"]=words_set
|
96 |
+
|
97 |
+
# print(filtered_tokens)
|
98 |
+
# print(ambiguity)
|
99 |
+
return ambiguity["lexical_st"]
|
__pycache__/detector.cpython-39.pyc
ADDED
Binary file (4.38 kB). View file
|
|
app.py
CHANGED
@@ -1,7 +1,48 @@
|
|
1 |
import gradio as gr
|
|
|
|
|
2 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
def greet(name):
|
4 |
return "Hello " + name + "!!"
|
5 |
|
6 |
-
iface = gr.Interface(fn=
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
iface.launch()
|
|
|
1 |
import gradio as gr
|
2 |
+
import re
|
3 |
+
from detector import AmbguityDetector
|
4 |
|
5 |
+
a=AmbguityDetector()
|
6 |
+
|
7 |
+
def update_array(array_a, array_b):
|
8 |
+
# Create a dictionary to store the mappings from the first value to the second value
|
9 |
+
mapping = {}
|
10 |
+
|
11 |
+
# Populate the dictionary using the tuples from array A
|
12 |
+
for a, b in array_a:
|
13 |
+
mapping[a] = b
|
14 |
+
|
15 |
+
# Iterate through the tuples in array B
|
16 |
+
for i, (a, b) in enumerate(array_b):
|
17 |
+
if b is None and a in mapping:
|
18 |
+
# Replace the tuple in array B with the value from array A
|
19 |
+
array_b[i] = (a, mapping[a])
|
20 |
+
|
21 |
+
def amb_texts(text):
|
22 |
+
tokens = re.split(r'(\s+)', text)
|
23 |
+
# tokens = [token for token in tokens if token.strip() != '']
|
24 |
+
ambs=a.sentence_ambiguity(text)
|
25 |
+
res=list()
|
26 |
+
for i in tokens:
|
27 |
+
res.append((i,None))
|
28 |
+
update_array(ambs,res)
|
29 |
+
# print(tokens,text)
|
30 |
+
return res
|
31 |
def greet(name):
|
32 |
return "Hello " + name + "!!"
|
33 |
|
34 |
+
iface = gr.Interface(fn=amb_texts, inputs= [
|
35 |
+
gr.Textbox(
|
36 |
+
label="Input",
|
37 |
+
info="Find ambiguities in the following",
|
38 |
+
lines=3,
|
39 |
+
value="The test can only continue if it receives all inputs from previous page.",
|
40 |
+
),
|
41 |
+
], outputs= gr.HighlightedText(
|
42 |
+
label="Ambiguity Detection",
|
43 |
+
combine_adjacent=True,
|
44 |
+
show_legend=True,
|
45 |
+
color_map={"lexical": "blue","scope":"yellow","referential":"orange","coordination":"pink","vague":"red"}),
|
46 |
+
theme=gr.themes.Base())
|
47 |
+
|
48 |
iface.launch()
|
coordination_encoded.pickel
ADDED
Binary file (7.45 kB). View file
|
|
detector.py
ADDED
@@ -0,0 +1,130 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import nltk
|
2 |
+
from nltk.tokenize import word_tokenize
|
3 |
+
from sentence_transformers import SentenceTransformer, util
|
4 |
+
import numpy
|
5 |
+
# from nltk.stem import WordNetLemmatizer
|
6 |
+
import pickle
|
7 |
+
import re
|
8 |
+
|
9 |
+
nltk.download('punkt')
|
10 |
+
|
11 |
+
lexical_AMB = ['bound', 'break', 'content', 'call', 'continue', 'contract', 'count', 'direct', 'even', 'express', 'form', 'forward', 'function', 'job',
|
12 |
+
'level', 'name', 'notice', 'number', 'out', 'position', 'record', 'reference', 'subject', 'string', 'switch', 'throw', 'translate', 'try', 'under']
|
13 |
+
referential_AMB = ['everyone', 'everything', 'someone',
|
14 |
+
'something', 'anything', 'anyone', 'itself', 'yourself']
|
15 |
+
coordination_AMB = ['also', 'if then', 'unless', 'if and only if']
|
16 |
+
scope_AMB = ['all', 'any', 'few', 'little', 'many', 'much', 'several', 'some']
|
17 |
+
vague_AMB = ['good', 'better', 'worse', 'available', 'common', 'capability', 'easy', 'full', 'maximum',
|
18 |
+
'minimum', 'quickly', 'random', 'recently', 'sufficient', 'sufficiently', 'simple', 'useful', 'various']
|
19 |
+
stopwords_custom = ['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', "you're", "you've", "you'll", "you'd", 'your', 'yours', 'yourselves',
|
20 |
+
'he', 'him', 'his', 'himself', 'she', "she's", 'her', 'hers', 'herself', 'it', "it's", 'its', 'they',
|
21 |
+
'them', 'their', 'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', "that'll", 'these',
|
22 |
+
'those', 'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does',
|
23 |
+
'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', 'at', 'by',
|
24 |
+
'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below',
|
25 |
+
'to', 'from', 'up', 'down', 'in', 'on', 'off', 'over', 'again', 'further', 'then', 'once', 'here', 'there', 'when',
|
26 |
+
'where', 'why', 'how', 'both', 'each', 'more', 'most', 'other', 'such', 'no', 'nor', 'not', 'only', 'own', 'same',
|
27 |
+
'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don', "don't", 'should', "should've", 'now', 'd', 'll',
|
28 |
+
'm', 'o', 're', 've', 'y', 'ain', 'aren', "aren't", 'couldn', "couldn't", 'didn', "didn't", 'doesn', "doesn't", 'hadn',
|
29 |
+
"hadn't", 'hasn', "hasn't", 'haven', "haven't", 'isn', "isn't", 'ma', 'mightn', "mightn't", 'mustn', "mustn't", 'needn',
|
30 |
+
"needn't", 'shan', "shan't", 'shouldn', "shouldn't", 'wasn', "wasn't", 'weren', "weren't", 'won', "won't", 'wouldn', "wouldn't"]
|
31 |
+
|
32 |
+
punctuation = ['.', ',', ';', '?']
|
33 |
+
|
34 |
+
# model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2')
|
35 |
+
# tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
|
36 |
+
|
37 |
+
lexical_encoded = pickle.load(open("lexical_encoded.pickel", "rb"))
|
38 |
+
vague_encoded = pickle.load(open("vague_encoded.pickel", "rb"))
|
39 |
+
referential_encoded = pickle.load(open("referential_encoded.pickel", "rb"))
|
40 |
+
coordination_encoded = pickle.load(open("coordination_encoded.pickel", "rb"))
|
41 |
+
scope_encoded = pickle.load(open("scope_encoded.pickel", "rb"))
|
42 |
+
|
43 |
+
class AmbguityDetector:
|
44 |
+
|
45 |
+
def __init__(self):
|
46 |
+
self.model = SentenceTransformer(
|
47 |
+
'sentence-transformers/all-MiniLM-L6-v2')
|
48 |
+
|
49 |
+
def sentence_ambiguity(self, sentence):
|
50 |
+
|
51 |
+
model = self.model
|
52 |
+
tokens = word_tokenize(sentence)
|
53 |
+
filtered_tokens = list()
|
54 |
+
for token in tokens:
|
55 |
+
if token not in stopwords_custom:
|
56 |
+
filtered_tokens.append(token)
|
57 |
+
|
58 |
+
for i in filtered_tokens:
|
59 |
+
filtered_tokens[filtered_tokens.index(i)] = i.lower()
|
60 |
+
if i in punctuation:
|
61 |
+
filtered_tokens.remove(i)
|
62 |
+
|
63 |
+
lexical = dict()
|
64 |
+
scope = dict()
|
65 |
+
referential = dict()
|
66 |
+
vague = dict()
|
67 |
+
coordination = dict()
|
68 |
+
ambiguity = dict()
|
69 |
+
ambiguous_words = list()
|
70 |
+
words_set=list()
|
71 |
+
|
72 |
+
for i in filtered_tokens:
|
73 |
+
temp = model.encode(i, convert_to_tensor=True)
|
74 |
+
for j in lexical_AMB:
|
75 |
+
temp2 = lexical_encoded[j]
|
76 |
+
cos_sim = util.pytorch_cos_sim(
|
77 |
+
temp, temp2).numpy().reshape([1, ])
|
78 |
+
if(cos_sim[0] >= 0.6):
|
79 |
+
ambiguous_words.append(i)
|
80 |
+
words_set.append((i,"lexical"))
|
81 |
+
lexical[i+"+"+j] = cos_sim[0]
|
82 |
+
|
83 |
+
for j in scope_AMB:
|
84 |
+
temp2 = scope_encoded[j]
|
85 |
+
cos_sim = util.pytorch_cos_sim(
|
86 |
+
temp, temp2).numpy().reshape([1, ])
|
87 |
+
if(cos_sim[0] >= 0.6):
|
88 |
+
ambiguous_words.append(i)
|
89 |
+
words_set.append((i,"scope"))
|
90 |
+
scope[i+"+"+j] = cos_sim[0]
|
91 |
+
|
92 |
+
for j in referential_AMB:
|
93 |
+
temp2 = referential_encoded[j]
|
94 |
+
cos_sim = util.pytorch_cos_sim(
|
95 |
+
temp, temp2).numpy().reshape([1, ])
|
96 |
+
if(cos_sim[0] >= 0.6):
|
97 |
+
ambiguous_words.append(i)
|
98 |
+
words_set.append((i,"referential"))
|
99 |
+
referential[i+"+"+j] = cos_sim[0]
|
100 |
+
|
101 |
+
for j in vague_AMB:
|
102 |
+
temp2 = vague_encoded[j]
|
103 |
+
cos_sim = util.pytorch_cos_sim(
|
104 |
+
temp, temp2).numpy().reshape([1, ])
|
105 |
+
if(cos_sim[0] >= 0.6):
|
106 |
+
ambiguous_words.append(i)
|
107 |
+
words_set.append((i,"vague"))
|
108 |
+
vague[i+"+"+j] = cos_sim[0]
|
109 |
+
|
110 |
+
for j in coordination_AMB:
|
111 |
+
temp2 = coordination_encoded[j]
|
112 |
+
cos_sim = util.pytorch_cos_sim(
|
113 |
+
temp, temp2).numpy().reshape([1, ])
|
114 |
+
if(cos_sim[0] >= 0.6):
|
115 |
+
ambiguous_words.append(i)
|
116 |
+
words_set.append((i,"coordination"))
|
117 |
+
coordination[i+"+"+j] = cos_sim[0]
|
118 |
+
|
119 |
+
ambiguous_words = list(dict.fromkeys(ambiguous_words))
|
120 |
+
ambiguity["lexical"] = lexical
|
121 |
+
ambiguity["referential"] = referential
|
122 |
+
ambiguity["scope"] = scope
|
123 |
+
ambiguity["vague"] = vague
|
124 |
+
ambiguity["coordination"] = coordination
|
125 |
+
ambiguity["words"] = ambiguous_words
|
126 |
+
ambiguity["lexical_st"]=words_set
|
127 |
+
|
128 |
+
# print(filtered_tokens)
|
129 |
+
# print(ambiguity)
|
130 |
+
return ambiguity["lexical_st"]
|
lexical_encoded.pickel
ADDED
Binary file (53.2 kB). View file
|
|
referential_encoded.pickel
ADDED
Binary file (14.8 kB). View file
|
|
scope_encoded.pickel
ADDED
Binary file (14.7 kB). View file
|
|
vague_encoded.pickel
ADDED
Binary file (33.1 kB). View file
|
|