def test_purified_file(self):
data_df = get_sms_dataset(raw=False)
assert data_df is not None
# confirm purified
for rule_name, matched in TextPurifier(
texts=data_df['message']).show_iter():
print(matched)
assert matched == []
def kf_test_with_datas(begin,
end,
test_num,
name_scoring="neg_mean_squared_error"):
data_df = get_sms_dataset(noStopwords=True, overwrite=True)
X, y = data_df['message'], data_df['target']
X_train, X_test, y_train, y_test = train_test_split(X,
y,
random_state=0,
test_size=0.2)
cv = CountVectorizer()
X_train = cv.fit_transform(X_train)
kf_test_and_draw(X_train, y_train, begin, end, test_num, name_scoring)
def auto_test_for_other_model(model, param_grid, name_scoring):
data_df = get_sms_dataset(noStopwords=True, overwrite=True)
X, y = data_df['message'], data_df['target']
X_train, X_test, y_train, y_test = train_test_split(X,
y,
random_state=0,
test_size=0.2)
cv = CountVectorizer()
X_train = cv.fit_transform(X_train)
grid = gsc(model, param_grid, scoring=name_scoring)
grid.fit(X_train, y_train)
print(grid.best_params_, grid.best_score_)
return grid.best_params_, grid.best_score_
def kf_test_with_datas_and_draw_alot(begin, end, test_num):
data_df = get_sms_dataset(noStopwords=True, overwrite=True)
X, y = data_df['message'], data_df['target']
X_train, X_test, y_train, y_test = train_test_split(X,
y,
random_state=0,
test_size=0.2)
cv = CountVectorizer()
X_train = cv.fit_transform(X_train)
kf_test_and_draw(X_train, y_train, begin, end, test_num, 'accuracy')
kf_test_and_draw(X_train, y_train, begin, end, test_num, 'precision')
kf_test_and_draw(X_train, y_train, begin, end, test_num, 'f1')
kf_test_and_draw(X_train, y_train, begin, end, test_num, 'recall')
def __init__(self,
sequence_length,
transform,
cate_type,
file_path='data/spam.csv'):
# get data and transform
data_df = get_sms_dataset(SMS_DATASET=file_path,
type=cate_type,
noStopwords=False)
texts, targets = data_df.message.to_list(), data_df.target.to_list()
self.texts, self.targets, self.word_dict, self.index_dict = transform(
texts, targets)
self.sequence_length = sequence_length
self.words = " ".join(texts).split()
self.words_indexes = [self.word_dict[w] for w in self.words]
def generate_model_no_length(scoring="accuracy"):
data_df = get_sms_dataset(noStopwords=True, overwrite=True)
X, y = data_df['message'], data_df['target']
cv = CountVectorizer()
X = cv.fit_transform(X)
X_train, X_test, y_train, y_test = train_test_split(X,
y,
random_state=0,
test_size=0.2)
bp = test_tool.auto_test(X_train, y_train, 0.1, 1.5, 200, scoring)
nb = MultiNB_Wrapper.train_once(X_train, y_train, bp)
MultiNB_Wrapper.save_model(nb,
cv,
model_name='old_model',
cv_name="old_cv")
pred = nb.predict(X_test)
print(metrics.confusion_matrix(y_test, pred))
return nb
def generate_model(scoring="precision"):
data_df = get_sms_dataset(noStopwords=True, overwrite=True)
X, y = data_df['message'], data_df['target']
cv = CountVectorizer()
X = cv.fit_transform(X)
lens = data_df['length']
new_len = csr_matrix(lens)
X = hstack((X, new_len.reshape(-1, 1)))
X_train, X_test, y_train, y_test = train_test_split(X,
y,
random_state=0,
test_size=0.2)
bp = test_tool.auto_test(X_train, y_train, 0.1, 1.5, 200, scoring)
nb = MultiNB_Wrapper.train_once(X_train, y_train, bp)
MultiNB_Wrapper.save_model(nb, cv)
pred = nb.predict(X_test)
print(metrics.confusion_matrix(y_test, pred))
return nb
"""
@Author: peviroy
@Date: 2020-09-03 20:54
@Last Modified by: peviroy
@Last Modified time: 2020-09-06 21:07
"""
import os
os.chdir(os.path.split(os.path.realpath(__file__))[0])
import sys
sys.path.append(os.path.abspath(".."))
from dataset import get_sms_dataset
from utils.preprocessing import TextPurifier
raw_texts = get_sms_dataset(raw=True)['v2'].tolist()
tp = TextPurifier(raw_texts)
class TestTextPurifierInstance:
def test_str(self):
assert tp.__str__().startswith('Name:')
def test_iter(self):
for rule_name, matched_strings in tp.show_iter():
assert matched_strings != []
def test_purify(self):
purifed_texts = tp.purify()
tp.set_texts(purifed_texts)
for rule_name, matched_strings in tp.show_iter():
def __init__(self, transform=SMSTransform(), file_path='data/spam.csv'):
data_df = get_sms_dataset(SMS_DATASET=file_path)
texts, targets = data_df.message.to_list(), data_df.target.to_list()
self.texts, self.targets, self.word_dict, self.index_dict = transform(
texts, targets)
def test_raw_file(self):
assert get_sms_dataset(raw=True) is not None
def test_null_raw_file(self):
try:
get_sms_dataset(raw=True, SMS_DATASET='null')
except Exception as e:
assert isinstance(e, FileNotFoundError)
