[#158] Add some unit tests for core.comparison.get_comparison_leaves()

The logic in this function that deals with the secondary, threshold_a, and
threshold_b attributes isn't tested or documented, and I'm not sure exactly
what it's supposed to do.

Author: riley-harper

hlink/tests/core/comparison_test.py

@@ -0,0 +1,78 @@
+import pytest
+
+from hlink.linking.core.comparison import get_comparison_leaves
+
+
+def test_get_comparison_leaves_base_case() -> None:
+    """
+    A comparison with no sub-comparisons (comp_a and comp_b) is itself the only leaf.
+    """
+    comparison = {
+        "comparison_type": "threshold",
+        "feature_name": "namefrst_jw",
+        "threshold": 0.79,
+    }
+    leaves = get_comparison_leaves(comparison)
+    assert leaves == [comparison]
+
+
+@pytest.mark.parametrize("operator", ["AND", "OR"])
+def test_get_comparison_leaves_one_level(operator: str) -> None:
+    """
+    When there are comp_a and comp_b subcomparisons, get_comparison_leaves()
+    extracts them as the leaves.
+    """
+    comparison_a = {
+        "comparison_type": "threshold",
+        "feature_name": "namefrst_jw",
+        "threshold": 0.79,
+    }
+    comparison_b = {
+        "comparison_type": "threshold",
+        "feature_name": "namelast_jw",
+        "threshold": 0.84,
+    }
+    comparisons = {
+        "operator": operator,
+        "comp_a": comparison_a,
+        "comp_b": comparison_b,
+    }
+    leaves = get_comparison_leaves(comparisons)
+    assert leaves == [comparison_a, comparison_b]
+
+
+@pytest.mark.parametrize("operator1", ["AND", "OR"])
+@pytest.mark.parametrize("operator2", ["AND", "OR"])
+def test_get_comparison_leaves_nested(operator1: str, operator2: str) -> None:
+    """
+    get_comparison_leaves() recurses through the tree to find leaves when there
+    are multiple nested levels.
+    """
+    comparison_a = {
+        "comparison_type": "threshold",
+        "feature_name": "namefrst_jw",
+        "threshold": 0.79,
+    }
+    comparison_b_a = {
+        "comparison_type": "threshold",
+        "feature_name": "namelast_jw",
+        "threshold": 0.84,
+    }
+    comparison_b_b = {
+        "comparison_type": "threshold",
+        "feature_name": "marst_flag",
+        "threshold_expr": ">0.5",
+    }
+
+    comparisons = {
+        "operator": operator1,
+        "comp_a": comparison_a,
+        "comp_b": {
+            "operator": operator2,
+            "comp_a": comparison_b_a,
+            "comp_b": comparison_b_b,
+        },
+    }
+
+    leaves = get_comparison_leaves(comparisons)
+    assert leaves == [comparison_a, comparison_b_a, comparison_b_b]