diff --git a/tests/unit/helpers/test_jaccard.py b/tests/unit/helpers/test_jaccard.py
index c6a04c3..a5b1d22 100644
--- a/tests/unit/helpers/test_jaccard.py
+++ b/tests/unit/helpers/test_jaccard.py
@@ -1,5 +1,7 @@
 """Unit tests for Jaccard similarity helpers."""
 
+import logging
+
 import numpy as np
 import pandas as pd
 import pytest
@@ -8,6 +10,7 @@
 
 
 def b(*vals: int) -> np.ndarray:
+    """Construct a 1-D boolean ndarray from integer literals."""
     return np.array(vals, dtype=bool)
 
 
@@ -17,17 +20,20 @@ def b(*vals: int) -> np.ndarray:
 
 
 def test_similarity_identical_vectors() -> None:
+    """Identical vectors should have similarity of 1."""
     x = b(1, 1, 0, 1)
     assert jaccard.similarity(x, x) == pytest.approx(1.0)
 
 
 def test_similarity_disjoint_vectors() -> None:
+    """Disjoint vectors should have similarity of 0."""
     x = b(1, 1, 0, 0)
     y = b(0, 0, 1, 1)
     assert jaccard.similarity(x, y) == pytest.approx(0.0)
 
 
 def test_similarity_partial_overlap() -> None:
+    """Vectors with partial overlap should return intersect/union."""
     x = b(1, 1, 1, 0)
     y = b(1, 1, 0, 1)
     # intersect=2, union=4 → 0.5
@@ -35,6 +41,7 @@ def test_similarity_partial_overlap() -> None:
 
 
 def test_similarity_all_zeros_returns_nan() -> None:
+    """All-zero vectors produce a zero denominator; result should be NaN."""
     x = b(0, 0, 0)
     y = b(0, 0, 0)
     result = jaccard.similarity(x, y)
@@ -42,6 +49,7 @@ def test_similarity_all_zeros_returns_nan() -> None:
 
 
 def test_similarity_with_explicit_px_py() -> None:
+    """Explicitly supplied px/py should be used instead of the sample mean."""
     x = b(1, 0, 1, 0)
     y = b(1, 0, 1, 0)
     result = jaccard.similarity(x, y, px=0.5, py=0.5)
@@ -49,6 +57,7 @@ def test_similarity_with_explicit_px_py() -> None:
 
 
 def test_similarity_centered() -> None:
+    """Centered similarity should subtract the expected overlap under independence."""
     x = b(1, 1, 0, 0)
     y = b(1, 1, 0, 0)
     # j=1, px=py=0.5, denominator=0.75, centered = 1 - (0.25/0.75)
@@ -57,6 +66,7 @@ def test_similarity_centered() -> None:
 
 
 def test_similarity_raises_on_2d_input() -> None:
+    """Non-1-D arrays should raise IndexError."""
     x = np.array([[True, False], [True, False]])
     y = np.array([[True, False], [True, False]])
     with pytest.raises(IndexError):
@@ -64,11 +74,13 @@ def test_similarity_raises_on_2d_input() -> None:
 
 
 def test_similarity_raises_on_length_mismatch() -> None:
+    """Arrays of different lengths should raise IndexError."""
     with pytest.raises(IndexError):
         jaccard.similarity(b(1, 0), b(1, 0, 1))
 
 
 def test_similarity_raises_on_non_boolean_dtype() -> None:
+    """Integer arrays should raise TypeError."""
     x = np.array([1, 0, 1], dtype=int)
     y = np.array([1, 0, 1], dtype=int)
     with pytest.raises(TypeError):
@@ -81,23 +93,27 @@ def test_similarity_raises_on_non_boolean_dtype() -> None:
 
 
 def test_distance_identical_vectors() -> None:
+    """Identical vectors should have distance of 0."""
     x = b(1, 0, 1)
     assert jaccard.distance(x, x) == pytest.approx(0.0)
 
 
 def test_distance_disjoint_vectors() -> None:
+    """Disjoint vectors should have distance of 1."""
     x = b(1, 1, 0, 0)
     y = b(0, 0, 1, 1)
     assert jaccard.distance(x, y) == pytest.approx(1.0)
 
 
 def test_distance_partial_overlap() -> None:
+    """Distance should equal 1 minus the Jaccard similarity."""
     x = b(1, 1, 1, 0)
     y = b(1, 1, 0, 1)
     assert jaccard.distance(x, y) == pytest.approx(0.5)
 
 
 def test_distance_with_explicit_px_py() -> None:
+    """Explicit px/py should be forwarded to similarity correctly."""
     x = b(1, 0, 1, 0)
     y = b(1, 0, 1, 0)
     result = jaccard.distance(x, y, px=0.5, py=0.5)
@@ -110,6 +126,7 @@ def test_distance_with_explicit_px_py() -> None:
 
 
 def test_bootstrap_returns_series_with_correct_index() -> None:
+    """Bootstrap should return a Series with J-sim and p-val labels."""
     x = b(1, 0, 1, 0, 1, 0)
     y = b(1, 1, 0, 0, 1, 0)
     result = jaccard.bootstrap(x, y, seed=0)
@@ -118,6 +135,7 @@ def test_bootstrap_returns_series_with_correct_index() -> None:
 
 
 def test_bootstrap_p_value_in_range() -> None:
+    """Bootstrap p-value should be in [0, 1]."""
     x = b(1, 0, 1, 0, 1, 0)
     y = b(1, 1, 0, 0, 1, 0)
     result = jaccard.bootstrap(x, y, n=200, seed=42)
@@ -125,10 +143,9 @@ def test_bootstrap_p_value_in_range() -> None:
 
 
 def test_bootstrap_degenerate_all_ones(caplog: pytest.LogCaptureFixture) -> None:
+    """All-ones vector should trigger the degenerate warning and return p-val=1."""
     x = b(1, 1, 1, 1)
     y = b(1, 0, 1, 0)
-    import logging
-
     with caplog.at_level(logging.WARNING, logger="lta.helpers.jaccard"):
         result = jaccard.bootstrap(x, y)
     assert result["p-val"] == pytest.approx(1.0)
@@ -136,10 +153,9 @@ def test_bootstrap_degenerate_all_ones(caplog: pytest.LogCaptureFixture) -> None
 
 
 def test_bootstrap_degenerate_all_zeros(caplog: pytest.LogCaptureFixture) -> None:
+    """All-zeros vector should trigger the degenerate warning and return p-val=1."""
     x = b(0, 0, 0, 0)
     y = b(1, 0, 1, 0)
-    import logging
-
     with caplog.at_level(logging.WARNING, logger="lta.helpers.jaccard"):
         result = jaccard.bootstrap(x, y)
     assert result["p-val"] == pytest.approx(1.0)
@@ -147,6 +163,7 @@ def test_bootstrap_degenerate_all_zeros(caplog: pytest.LogCaptureFixture) -> Non
 
 
 def test_bootstrap_deterministic_with_seed() -> None:
+    """Same seed should produce identical J-sim and p-val across calls."""
     x = b(1, 0, 1, 0, 1, 0, 1, 0)
     y = b(0, 1, 0, 1, 1, 0, 1, 0)
     r1 = jaccard.bootstrap(x, y, n=500, seed=7)
@@ -156,6 +173,7 @@ def test_bootstrap_deterministic_with_seed() -> None:
 
 
 def test_bootstrap_jsim_matches_similarity() -> None:
+    """J-sim in bootstrap result should match the plain similarity value."""
     x = b(1, 0, 1, 0, 1, 0)
     y = b(1, 1, 0, 0, 1, 0)
     result = jaccard.bootstrap(x, y, seed=42)