def load(self, path):
""" Load the model.
Parameters
----------
path : str
a full path to a file from which a model will be loaded
"""
self.estimator = pickle.load(path)
def modelpredict(text:str):
try:
model = joblib.load('pipeline_lr.pkl')
if not text.endswith('txt'):
string = [text]
preds = model.predict(string)[0]
pred_prb = round(model.predict_proba(string)[:,1][0], 2)
return f'Verbatism: {text}\n Sentiment Value: {preds}\n Sentiment score: {pred_prb}%'
else:
df = pd.read_csv(text, names =['sentiment'], delimiter='\t', encoding='utf-8')
df['predictions'] = model.predict(df['sentiment'])
score = model.predict_proba(df['sentiment'])
sentiment_score = [i[1] for i in score]
df['sentiment_score'] = sentiment_score
return df
except ValueError as e:
return f'Value error is {e}'
def load_model(file_path):
"""Loads a model that was saved with joblib."""
return joblib.load(file_path)
def predict():
pred = joblib.load('JUSL_RF_model.sav')
result = pred.predict(
[[33, 27, 18, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]])
return result