diff --git a/features/__pycache__/__init__.cpython-311.pyc b/features/__pycache__/__init__.cpython-311.pyc
deleted file mode 100644
index f20d9b9..0000000
Binary files a/features/__pycache__/__init__.cpython-311.pyc and /dev/null differ
diff --git a/features/__pycache__/__init__.cpython-312.pyc b/features/__pycache__/__init__.cpython-312.pyc
deleted file mode 100644
index b2bc762..0000000
Binary files a/features/__pycache__/__init__.cpython-312.pyc and /dev/null differ
diff --git a/features/__pycache__/feature_engineering.cpython-311.pyc b/features/__pycache__/feature_engineering.cpython-311.pyc
deleted file mode 100644
index ca342a2..0000000
Binary files a/features/__pycache__/feature_engineering.cpython-311.pyc and /dev/null differ
diff --git a/features/__pycache__/feature_engineering.cpython-312.pyc b/features/__pycache__/feature_engineering.cpython-312.pyc
deleted file mode 100644
index abdc0ef..0000000
Binary files a/features/__pycache__/feature_engineering.cpython-312.pyc and /dev/null differ
diff --git a/features/__pycache__/labeling.cpython-311.pyc b/features/__pycache__/labeling.cpython-311.pyc
deleted file mode 100644
index 2502012..0000000
Binary files a/features/__pycache__/labeling.cpython-311.pyc and /dev/null differ
diff --git a/features/__pycache__/labeling_schemes.cpython-311.pyc b/features/__pycache__/labeling_schemes.cpython-311.pyc
deleted file mode 100644
index a66a95a..0000000
Binary files a/features/__pycache__/labeling_schemes.cpython-311.pyc and /dev/null differ
diff --git a/features/__pycache__/labeling_schemes.cpython-312.pyc b/features/__pycache__/labeling_schemes.cpython-312.pyc
deleted file mode 100644
index e1bc53d..0000000
Binary files a/features/__pycache__/labeling_schemes.cpython-312.pyc and /dev/null differ
diff --git a/tests/test_labeling_schemes.py b/tests/test_labeling_schemes.py
new file mode 100644
index 0000000..4cc4349
--- /dev/null
+++ b/tests/test_labeling_schemes.py
@@ -0,0 +1,90 @@
+import pytest
+import pandas as pd
+import numpy as np
+
+# Adjust import based on where the module actually lives.
+from features.labeling_schemes import calculate_future_returns
+
+def test_calculate_future_returns_default_horizon():
+    """Test default horizon=1 calculation"""
+    # Create sample dataframe
+    df = pd.DataFrame({
+        "close": [100.0, 105.0, 102.9, 110.0]
+    })
+
+    # Calculate returns
+    result_df = calculate_future_returns(df.copy())
+
+    # Check shape - should drop last row due to horizon=1
+    assert len(result_df) == 3
+
+    # Check if 'future_returns' column exists
+    assert "future_returns" in result_df.columns
+
+    # Expected returns
+    # row 0: (105.0 - 100.0) / 100.0 = 0.05
+    # row 1: (102.9 - 105.0) / 105.0 = -0.02
+    # row 2: (110.0 - 102.9) / 102.9 = 0.0690...
+    expected_returns = [0.05, -0.02, (110.0 - 102.9) / 102.9]
+
+    # Assert values are close
+    np.testing.assert_allclose(result_df["future_returns"].values, expected_returns)
+
+def test_calculate_future_returns_custom_horizon():
+    """Test with custom horizon=2"""
+    df = pd.DataFrame({
+        "close": [100.0, 105.0, 110.0, 107.8]
+    })
+
+    result_df = calculate_future_returns(df.copy(), horizon=2)
+
+    # Should drop last 2 rows
+    assert len(result_df) == 2
+
+    # Expected returns for horizon 2
+    # row 0: (110.0 - 100.0) / 100.0 = 0.10
+    # row 1: (107.8 - 105.0) / 105.0 = 0.0266...
+    expected_returns = [0.10, (107.8 - 105.0) / 105.0]
+
+    np.testing.assert_allclose(result_df["future_returns"].values, expected_returns)
+
+def test_calculate_future_returns_missing_close_column():
+    """Test KeyError is raised when 'close' column is missing"""
+    df = pd.DataFrame({
+        "price": [100.0, 105.0]
+    })
+
+    with pytest.raises(KeyError):
+        calculate_future_returns(df)
+
+def test_calculate_future_returns_empty_dataframe():
+    """Test behavior with an empty dataframe"""
+    df = pd.DataFrame(columns=["close"])
+
+    result_df = calculate_future_returns(df.copy())
+
+    assert len(result_df) == 0
+    assert "future_returns" in result_df.columns
+
+def test_calculate_future_returns_all_nans():
+    """Test behavior with a dataframe containing only NaNs"""
+    df = pd.DataFrame({
+        "close": [np.nan, np.nan, np.nan]
+    })
+
+    result_df = calculate_future_returns(df.copy())
+
+    # Dropna should drop all rows
+    assert len(result_df) == 0
+    assert "future_returns" in result_df.columns
+
+def test_calculate_future_returns_horizon_larger_than_data():
+    """Test behavior when horizon is larger than dataframe size"""
+    df = pd.DataFrame({
+        "close": [100.0, 105.0]
+    })
+
+    result_df = calculate_future_returns(df.copy(), horizon=5)
+
+    # Should drop all rows
+    assert len(result_df) == 0