Update functions.py
Browse files- functions.py +1 -18
functions.py
CHANGED
@@ -76,29 +76,12 @@ def inference(link, upload, _asr_model):
|
|
76 |
results = _asr_model.trasncribe(upload)
|
77 |
|
78 |
return results['text'], "Transcribed Earnings Audio"
|
79 |
-
|
80 |
-
@st.experimental_memo(suppress_st_warning=True)
|
81 |
-
def chunk_long_text(text,threshold):
|
82 |
-
'''Chunk long text'''
|
83 |
-
sentences = sent_tokenize(text)
|
84 |
-
out = []
|
85 |
-
|
86 |
-
for chunk in sentences:
|
87 |
-
if len(chunk.split()) < threshold:
|
88 |
-
out.append(chunk)
|
89 |
-
else:
|
90 |
-
words = chunk.split()
|
91 |
-
num = int(len(words)/threshold)
|
92 |
-
for i in range(0,num*threshold+1,threshold):
|
93 |
-
out.append(' '.join(words[i:threshold+i]))
|
94 |
-
|
95 |
-
return out
|
96 |
|
97 |
@st.experimental_memo(suppress_st_warning=True)
|
98 |
def sentiment_pipe(earnings_text):
|
99 |
'''Determine the sentiment of the text'''
|
100 |
|
101 |
-
earnings_sentences = chunk_long_text(earnings_text,
|
102 |
earnings_sentiment = sent_pipe(earnings_sentences)
|
103 |
|
104 |
return earnings_sentiment, earnings_sentences
|
|
|
76 |
results = _asr_model.trasncribe(upload)
|
77 |
|
78 |
return results['text'], "Transcribed Earnings Audio"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
79 |
|
80 |
@st.experimental_memo(suppress_st_warning=True)
|
81 |
def sentiment_pipe(earnings_text):
|
82 |
'''Determine the sentiment of the text'''
|
83 |
|
84 |
+
earnings_sentences = chunk_long_text(earnings_text,200,1)
|
85 |
earnings_sentiment = sent_pipe(earnings_sentences)
|
86 |
|
87 |
return earnings_sentiment, earnings_sentences
|