Or Hiltch orcaman

## main.go
	ctx := context.Background()
	cli, err := client.NewClientWithOpts(client.FromEnv, client.WithAPIVersionNegotiation())
	if err != nil {
		return err
	}

	resp, err := cli.ContainerCreate(ctx, &container.Config{
		Image:        image,
		Env:          env,
		Tty:          true,

## iris.csv

          
            sepal_length
            sepal_width
            petal_length
            petal_width
            species

            
              5.1
              3.5
              1.4
              0.2
              setosa

            
              4.9
              3.0
              1.4
              0.2
              setosa

            
              4.7
              3.2
              1.3
              0.2
              setosa

            
              4.6
              3.1
              1.5
              0.2
              setosa

            
              5.0
              3.6
              1.4
              0.2
              setosa

            
              5.4
              3.9
              1.7
              0.4
              setosa

            
              4.6
              3.4
              1.4
              0.3
              setosa

            
              5.0
              3.4
              1.5
              0.2
              setosa

            
              4.4
              2.9
              1.4
              0.2
              setosa

## gmmroc.py
# results on training set
y_pred = xgb_test.predict(dtrain, ntree_limit=xgb_test.best_iteration+1)
y_true = train_df['class'].values
print(roc_auc(y_pred, dtrain))

# results on test set
y_pred = xgb_test.predict(dtest, ntree_limit=xgb_test.best_iteration+1)
y_true = test_df['class'].values
print(roc_auc(y_pred, dtest))

## gmm_xgb_test.predict.py
dsyn = xgb.DMatrix(syn[X_col], syn[y_col], feature_names=X_col)
y_pred = xgb_test.predict(dsyn, ntree_limit=xgb_test.best_iteration+1)
y_true = syn['class'].values
print(recall(y_pred, dsyn))
print(precision(y_pred, dsyn))

## gmm.sample.py
t1 = gmm.sample(len(real_data))
data_new = t1[0]

points_new = pca.inverse_transform(data_new)
syn = pd.DataFrame(points_new, columns=list(real_data))

## gmm.py
gmm = GaussianMixture(20, covariance_type='full', random_state=0)
gmm.fit(data)

t1 = gmm.sample(len(data_real))
data_new = t1[0]

points_new = pca.inverse_transform(data_new)
df_fake = pd.DataFrame(points_new, columns=list(data_real)[:-2])

## pca.py
from sklearn.decomposition import PCA
pca = PCA(0.99, whiten=True)
data = pca.fit_transform(df_real)

## find_components.py
n_components = np.arange(5, 30, 5)
models = [GaussianMixture(n, covariance_type='full', random_state=0)
          for n in n_components]
aics = [model.fit(data).aic(data) for model in models]
plt.plot(n_components, aics)
plt.show()

## train.py
results_dict = {}

xgb_params = {
    'objective': 'binary:logistic',
    'random_state': 0,
    'eval_metric': 'auc', # auc, error
}

xgb_test = xgb.train(xgb_params, dtrain, num_boost_round=100,
                     verbose_eval=False,

## setup.py
np.random.seed(0)

n_real = np.sum(data.Class==0)
n_test = np.sum(data.Class==1)
train_fraction = 0.7
fn_real = int(n_real * train_fraction)
fn_test = int(n_test * train_fraction)

real_samples = data.loc[ data.Class==0, test_cols].sample(n_real, replace=False).reset_index(drop=True)
test_samples = data.loc[ data.Class==1, test_cols].sample(n_test, replace=False).reset_index(drop=True)
	ctx := context.Background()
	cli, err := client.NewClientWithOpts(client.FromEnv, client.WithAPIVersionNegotiation())
	if err != nil {
	return err
	}

	resp, err := cli.ContainerCreate(ctx, &container.Config{
	Image: image,
	Env: env,
	Tty: true,
sepal_length	sepal_width	petal_length	petal_width	species
5.1	3.5	1.4	0.2	setosa
4.9	3.0	1.4	0.2	setosa
4.7	3.2	1.3	0.2	setosa
4.6	3.1	1.5	0.2	setosa
5.0	3.6	1.4	0.2	setosa
5.4	3.9	1.7	0.4	setosa
4.6	3.4	1.4	0.3	setosa
5.0	3.4	1.5	0.2	setosa
4.4	2.9	1.4	0.2	setosa
	# results on training set
	y_pred = xgb_test.predict(dtrain, ntree_limit=xgb_test.best_iteration+1)
	y_true = train_df['class'].values
	print(roc_auc(y_pred, dtrain))

	# results on test set
	y_pred = xgb_test.predict(dtest, ntree_limit=xgb_test.best_iteration+1)
	y_true = test_df['class'].values
	print(roc_auc(y_pred, dtest))
	dsyn = xgb.DMatrix(syn[X_col], syn[y_col], feature_names=X_col)
	y_pred = xgb_test.predict(dsyn, ntree_limit=xgb_test.best_iteration+1)
	y_true = syn['class'].values
	print(recall(y_pred, dsyn))
	print(precision(y_pred, dsyn))
	t1 = gmm.sample(len(real_data))
	data_new = t1[0]

	points_new = pca.inverse_transform(data_new)
	syn = pd.DataFrame(points_new, columns=list(real_data))
	gmm = GaussianMixture(20, covariance_type='full', random_state=0)
	gmm.fit(data)

	t1 = gmm.sample(len(data_real))
	data_new = t1[0]

	points_new = pca.inverse_transform(data_new)
	df_fake = pd.DataFrame(points_new, columns=list(data_real)[:-2])
	from sklearn.decomposition import PCA
	pca = PCA(0.99, whiten=True)
	data = pca.fit_transform(df_real)
	n_components = np.arange(5, 30, 5)
	models = [GaussianMixture(n, covariance_type='full', random_state=0)
	for n in n_components]
	aics = [model.fit(data).aic(data) for model in models]
	plt.plot(n_components, aics)
	plt.show()
	results_dict = {}

	xgb_params = {
	'objective': 'binary:logistic',
	'random_state': 0,
	'eval_metric': 'auc', # auc, error
	}

	xgb_test = xgb.train(xgb_params, dtrain, num_boost_round=100,
	verbose_eval=False,
	np.random.seed(0)

	n_real = np.sum(data.Class==0)
	n_test = np.sum(data.Class==1)
	train_fraction = 0.7
	fn_real = int(n_real * train_fraction)
	fn_test = int(n_test * train_fraction)

	real_samples = data.loc[ data.Class==0, test_cols].sample(n_real, replace=False).reset_index(drop=True)
	test_samples = data.loc[ data.Class==1, test_cols].sample(n_test, replace=False).reset_index(drop=True)