kaushikbar
commited on
Commit
•
4198959
1
Parent(s):
d3269c3
cleaned up
Browse files
app.py
CHANGED
@@ -108,50 +108,8 @@ def prep_examples():
|
|
108 |
|
109 |
return examples
|
110 |
|
111 |
-
def detect_lang(sequence, labels):
|
112 |
-
DetectorFactory.seed = 0
|
113 |
-
seq_lang = 'en'
|
114 |
-
|
115 |
-
sequence = sequence.replace('\n', ' ')
|
116 |
-
|
117 |
-
try:
|
118 |
-
#seq_lang = detect(sequence)
|
119 |
-
#lbl_lang = detect(labels)
|
120 |
-
seq_lang = fasttext_model.predict(sequence, k=1)[0][0].split("__label__")[1]
|
121 |
-
lbl_lang = fasttext_model.predict(labels, k=1)[0][0].split("__label__")[1]
|
122 |
-
except:
|
123 |
-
print("Language detection failed!",
|
124 |
-
"Date:{}, Sequence:{}, Labels:{}".format(
|
125 |
-
str(datetime.datetime.now()),
|
126 |
-
labels))
|
127 |
-
|
128 |
-
if seq_lang != lbl_lang:
|
129 |
-
print("Different languages detected for sequence and labels!",
|
130 |
-
"Date:{}, Sequence:{}, Labels:{}, Sequence Language:{}, Label Language:{}".format(
|
131 |
-
str(datetime.datetime.now()),
|
132 |
-
sequence,
|
133 |
-
labels,
|
134 |
-
seq_lang,
|
135 |
-
lbl_lang))
|
136 |
-
|
137 |
-
if seq_lang in models:
|
138 |
-
print("Sequence Language detected.",
|
139 |
-
"Date:{}, Sequence:{}, Sequence Language:{}".format(
|
140 |
-
str(datetime.datetime.now()),
|
141 |
-
sequence,
|
142 |
-
seq_lang))
|
143 |
-
else:
|
144 |
-
print("Language not supported. Defaulting to English!",
|
145 |
-
"Date:{}, Sequence:{}, Sequence Language:{}".format(
|
146 |
-
str(datetime.datetime.now()),
|
147 |
-
sequence,
|
148 |
-
seq_lang))
|
149 |
-
seq_lang = 'en'
|
150 |
-
|
151 |
-
return seq_lang
|
152 |
-
|
153 |
def sequence_to_classify(sequence, labels, multi_label):
|
154 |
-
lang = 'en'
|
155 |
classifier = classifiers[lang]
|
156 |
|
157 |
label_clean = str(labels).split(";;")
|
|
|
108 |
|
109 |
return examples
|
110 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
111 |
def sequence_to_classify(sequence, labels, multi_label):
|
112 |
+
lang = 'en'
|
113 |
classifier = classifiers[lang]
|
114 |
|
115 |
label_clean = str(labels).split(";;")
|