def test_eval_input_subsets(self):
""" Test inputs are subsets of the provided DataFrames for eval_pnd()
"""
# Model to make Dataset
md_true = make_pareto_random(twoDim=False)
# Create dataframe
df_data = (
md_true
>> gr.ev_sample(n=2e3, seed=101, df_det="nom")
)
## Select training set
df_train = (
df_data
>> gr.tf_sample(n=10)
)
## select test set
df_test = (
df_data
>> gr.tf_anti_join(
df_train,
by=["x1", "x2"],
)
>> gr.tf_sample(n=200)
)
# Create fitted model
md_fit = (
df_train
>> ft_gp(
var=["x1", "x2", "x3"],
out=["y1", "y2", "y3"],
)
)
# Call eval_pnd w/ only "y1" and "y2"
df_pnd = (
md_fit
>> gr.ev_pnd(
df_train,
df_test,
signs = {"y1":1, "y2":1},
seed = 101
)
)
### how to imply x1 and x2 from y1 and y2?
# Test for correctness by shape
self.assertTrue(len(df_pnd) == df_test.shape[0])
# Test for correctness by # of outputs
self.assertTrue(len(df_pnd.columns.values) == len(df_test.columns.values) + 2)
def test_eval_append(self):
""" Test append parameter on eval_pnd()
"""
# Model to make Dataset
md_true = make_pareto_random(twoDim=False)
# Create dataframe
df_data = (
md_true
>> gr.ev_sample(n=2e3, seed=101, df_det="nom")
)
## Select training set
df_train = (
df_data
>> gr.tf_sample(n=10)
)
## select test set
df_test = (
df_data
>> gr.tf_anti_join(
df_train,
by=["x1", "x2"],
)
>> gr.tf_sample(n=200)
)
# Create fitted model
md_fit = (
df_train
>> ft_gp(
var=["x1", "x2", "x3"],
out=["y1", "y2", "y3"],
)
)
# Call eval_pnd
df_pnd = (
md_fit
>> gr.ev_pnd(
df_train,
df_test,
signs = {"y1":1, "y2":1,"y3":1},
seed = 101,
append = False
)
)
# Test for correctness by shape
self.assertTrue(len(df_pnd) == df_test.shape[0])
# Test for correctness by # of outputs
self.assertTrue(len(df_pnd.columns.values) == 2)
def test_eval_faulty_inputs(self):
""" Test faulty inputs to eval_pnd
"""
# Model to make Dataset
md_true = make_pareto_random()
# Create dataframe
df_data = (
md_true
>> gr.ev_sample(n=2e3, seed=101, df_det="nom")
)
## Select training set
df_train = (
df_data
>> gr.tf_sample(n=10)
)
## select test set
df_test = (
df_data
>> gr.tf_anti_join(
df_train,
by=["x1", "x2"],
)
>> gr.tf_sample(n=200)
)
# Create fitted model
md_fit = (
df_train
>> ft_gp(
var=["x1", "x2"],
out=["y1", "y2"],
)
)
# Call eval_pnd
with self.assertRaises(ValueError):
df_pnd = (
md_fit
>> gr.ev_pnd(
df_train,
df_test,
signs = {"y":1, "y2":1},
seed = 101
)
)
def test_umap(self):
## UMAP executes successfully
df_umap = (data.df_diamonds >> gr.tf_sample(n=100) >>
tran.tf_umap(var=["x", "y", "z"]))
def test_tsne(self):
## t-SNE executes successfully
df_tsne = (data.df_diamonds >> gr.tf_sample(n=100) >>
tran.tf_tsne(var=["x", "y", "z"]))