def test_flow(self):
with open(r"..\models\tests.YAML", 'r') as stream:
tests_dict = yaml.load(stream)
tests = tests_dict['tests']
# tests = pass
# Load Data
data = PreprocessTags(True).load_data(r'..\data\train.wtag')
model1 = Model(tests)
# create f_x_y "matrix" for any x,y save indices of non zero tests TODO decide
a = model1.fit(data.x, data.y)
data = PreprocessTags(True).load_data(r'..\data\test.wtag')
try:
y_hat = model1.predict(data.x)
cm = model1.confusion(y_hat=y_hat, y=data.y)
cm.to_csv(r'../training/confusion_matrix.csv')
results = dict(v=model1.v.tolist(),
compare={
test: dict(v_val=v_val, sum=sum)
for test, v_val in zip(tests, model1.v.tolist())
},
acc=model1.accuracy(y_hat=y_hat, y=data.y),
acc_per_tag=model1.acc_per_tag(y_hat=y_hat,
y=data.y))
except:
results = dict(v=model1.v.tolist(),
compare={
test: dict(v_val=v_val, sum=sum)
for test, v_val in zip(tests, model1.v.tolist())
})
t = time.localtime()
with open(
fr"../training/report_{t.tm_hour}+{t.tm_min}_{t.tm_mday}_{t.tm_mon}.YAML",
'w') as stream:
yaml.dump(results, stream, default_flow_style=False)
