def test_datamation_groupby_multiple():
df = small_salary().df
df = DatamationFrame(df)
# Group by Degree, Work
mean = df.groupby(['Degree', 'Work']).mean()
assert "groupby" in mean.operations
assert "mean" in mean.operations
assert len(mean.states) == 2
assert df.equals(mean.states[0])
assert mean.Salary.Masters.Academia == 84.0298831968801
assert mean.Salary.Masters.Industry == 91.22576155606282
assert mean.Salary.PhD.Academia == 85.55796571969728
assert mean.Salary.PhD.Industry == 93.08335885824636
# sum
sum = df.groupby(['Degree', 'Work']).sum()
assert "groupby" in sum.operations
assert "sum" in sum.operations
assert len(sum.states) == 2
assert df.equals(sum.states[0])
assert sum.Salary.Masters.Academia == 840.2988319688011
assert sum.Salary.Masters.Industry == 5655.997216475895
assert sum.Salary.PhD.Academia == 1540.043382954551
assert sum.Salary.PhD.Industry == 930.8335885824636
# product
product = df.groupby(['Degree', 'Work']).prod()
assert "groupby" in product.operations
assert "product" in product.operations
assert len(product.states) == 2
assert df.equals(product.states[0])
assert product.Salary.Masters.Academia == 1.753532557780977e+19
assert product.Salary.Masters.Industry == 3.3602152421057308e+121
assert product.Salary.PhD.Academia == 6.027761935702164e+34
assert product.Salary.PhD.Industry == 4.8818435443657834e+19
# Group by species, island, sex
df = DatamationFrame(load_penguins())
mean = df.groupby(['species', 'island', 'sex']).mean()
assert "groupby" in mean.operations
assert "mean" in mean.operations
assert len(mean.states) == 2
assert df.equals(mean.states[0])
assert mean.bill_length_mm.Adelie.Biscoe.male == approx(40.5909090909091)
assert mean.bill_length_mm.Adelie.Biscoe.female == approx(
37.35909090909092)
def _load_penguins():
print("🐧 loading penguins...")
from palmerpenguins import load_penguins
X, y = load_penguins(return_X_y=True)
print(X.head())
return X, y
def train():
data, target = load_penguins(return_X_y=True, drop_na=True)
models = {
"rf": RandomForestClassifier(max_depth=3, random_state=0),
"dt": DecisionTreeClassifier(max_depth=3, random_state=0),
}
for name, model in models.items():
model.fit(data, target)
joblib.dump(model, f"model/{name}.pkl")
print(f"save model: {name}")
def data_sourcing():
"""
This template function uses the Palmer Peguins dataset as a place holder.
Replace it by your own code to import your project's data.
"""
df = palmerpenguins.load_penguins()
cols = [
"bill_length_mm",
"bill_depth_mm",
"flipper_length_mm",
"body_mass_g",
"sex",
"species",
]
return df[cols]
def input_dataframe():
penguin_df = load_penguins()
return penguin_df
def penguins_data():
return simplify(load_penguins())