import pandas as pd
import scripts
from preparing_neural import apply_preparing_merge
from cleaning import apply_cleaning
from train_neural_liam_test import split_train_test_data, fit_model, save_model
from preparing_neural import apply_lemmatize
from train_neural_liam_test import get_preprocessor
data = pd.read_csv('../raw_data/working.csv')
data_clean = apply_cleaning(data)
data_merge = apply_preparing_merge(data_clean)
data_lemmatize = apply_lemmatize(data_merge)
x_train, x_test, y_train, y_test = split_train_test_data(data_lemmatize)
x_train_preprocessed, x_test_preprocessed = get_preprocessor(x_train, x_test)
neural_model = fit_model(x_train_preprocessed, y_train)
save_model(neural_model)
print(neural_model.evaluate(x_test_preprocessed, y_test))
if input_method == 'Text':
title = st.text_input('Article title')
text = st.text_area('Article body')
elif input_method == 'Link':
url = st.text_input('Article URL')
analyze_status_logistic = st.button('Analyze_with_Logistic')
analyze_status_neural = st.button('Analyze_with_Neural')
###### Logistic
if input_method == 'Text' and analyze_status_logistic == True:
input_df = convert(title,text)
input_df = apply_cleaning(input_df)
input_df = input_df[['title_clean', 'text_clean','title_length_char','title_Upper_Ratio','text_stop_words_ratio']]
prediction = logistic_model.predict(input_df)
if prediction == 1:
st.write('I think its true')
else:
st.write('I think its fake')
if input_method == 'Link' and analyze_status_logistic == True:
input_df = get_title_text_web(url)
input_df = apply_cleaning(input_df)
input_df = input_df[['title_clean', 'text_clean','title_length_char','title_Upper_Ratio','text_stop_words_ratio']]
import pandas as pd
from preparing_df import apply_preparing
from cleaning import apply_cleaning
from typo import apply_typo_ratio
from train import split_train_test_data, fit_model, save_model
true = pd.read_csv('True.csv')
fake = pd.read_csv('Fake.csv')
data = apply_preparing(true, fake)
apply_cleaning(data)
apply_typo_ratio(data)
x_train, x_test, y_train, y_test = split_train_test_data(data)
pipe = fit_model(x_train, y_train)
save_model(pipe, 'model_test')
print(pipe.score(x_test, y_test))
index0 = st.file_uploader("Choose a file")
if index0 is not None:
st.sidebar.header('Dataset Parameter')
x1 = pd.ExcelFile(index0)
index1 = st.sidebar.selectbox('What Dataset you choose?', x1.sheet_names)
# Load data example (dari functional maupun nonfunctional)
st.header('Dataset parameters')
statement = fulldataset(index0, index1)
# Get text to clean (dari row yang diinginkan)
text_to_clean = list(statement['Requirement Statement'])
# Clean text
print("Loading Original & Cleaned Text...")
cleaned_text = apply_cleaning(text_to_clean)
# Show first example
text_df = pd.DataFrame([text_to_clean, cleaned_text],
index=['ORIGINAL', 'CLEANED'],
columns=statement['ID']).T
st.write(text_df)
st.header('Traceability parameters')
id_requirement = fulldataset(index0, index1)['ID']
genre = st.sidebar.radio(
"What do you choose?",
('Information_Retrieval', 'Ontology', 'IR+LSA', 'IR+LDA'))
if genre == 'Information_Retrieval':
st.subheader("bag of words")