testcases for community detection (#3163)

* testcases for community detection

* ruff lint

* Run 'pre-commit run --all'

* Rename test cases to clarify that they're comm. detect. related

---------

Co-authored-by: Tom Aarsen <[email protected]>

tests/test_util.py

@@ -1,10 +1,12 @@
 from __future__ import annotations
 
 import numpy as np
+import pytest
 import sklearn
 import torch
 
 from sentence_transformers import SentenceTransformer, util
+from sentence_transformers.util import community_detection
 
 
 def test_normalize_embeddings() -> None:
@@ -145,3 +147,131 @@ def test_dot_score_cos_sim() -> None:
 
     assert np.allclose(cosine_calculated, dot_and_cosine_expected)
     assert np.allclose(dot_calculated, dot_and_cosine_expected)
+
+
+def test_community_detection_two_clear_communities():
+    """Test case with two clear communities."""
+    embeddings = torch.tensor(
+        [
+            [1.0, 0.0, 0.0],  # Point 0
+            [0.9, 0.1, 0.0],  # Point 1
+            [0.8, 0.2, 0.0],  # Point 2
+            [0.1, 0.9, 0.0],  # Point 3
+            [0.0, 1.0, 0.0],  # Point 4
+            [0.2, 0.8, 0.0],  # Point 5
+        ]
+    )
+    expected = [
+        [0, 1, 2],  # Community 1
+        [3, 4, 5],  # Community 2
+    ]
+    result = community_detection(embeddings, threshold=0.8, min_community_size=2)
+    assert sorted([sorted(community) for community in result]) == sorted([sorted(community) for community in expected])
+
+
+def test_community_detection_no_communities_high_threshold():
+    """Test case where no communities are found due to a high threshold."""
+    embeddings = torch.tensor(
+        [
+            [1.0, 0.0, 0.0],
+            [0.0, 1.0, 0.0],
+            [0.0, 0.0, 1.0],
+        ]
+    )
+    expected = []
+    result = community_detection(embeddings, threshold=0.99, min_community_size=2)
+    assert result == expected
+
+
+def test_community_detection_all_points_in_one_community():
+    """Test case where all points form a single community due to a low threshold."""
+    embeddings = torch.tensor(
+        [
+            [1.0, 0.0, 0.0],
+            [0.9, 0.1, 0.0],
+            [0.8, 0.2, 0.0],
+        ]
+    )
+    expected = [
+        [0, 1, 2],  # Single community
+    ]
+    result = community_detection(embeddings, threshold=0.5, min_community_size=2)
+    assert sorted([sorted(community) for community in result]) == sorted([sorted(community) for community in expected])
+
+
+def test_community_detection_min_community_size_filtering():
+    """Test case where communities are filtered based on minimum size."""
+    embeddings = torch.tensor(
+        [
+            [1.0, 0.0, 0.0],
+            [0.9, 0.1, 0.0],
+            [0.8, 0.2, 0.0],
+            [0.1, 0.9, 0.0],
+        ]
+    )
+    expected = [
+        [0, 1, 2],  # Only one community meets the min size requirement
+    ]
+    result = community_detection(embeddings, threshold=0.8, min_community_size=3)
+    assert sorted([sorted(community) for community in result]) == sorted([sorted(community) for community in expected])
+
+
+def test_community_detection_overlapping_communities():
+    """Test case with overlapping communities (resolved by the function)."""
+    embeddings = torch.tensor(
+        [
+            [1.0, 0.0, 0.0],  # Point 0
+            [0.9, 0.1, 0.0],  # Point 1
+            [0.8, 0.2, 0.0],  # Point 2
+            [0.7, 0.3, 0.0],  # Point 3 (overlaps with both communities)
+            [0.1, 0.9, 0.0],  # Point 4
+            [0.0, 1.0, 0.0],  # Point 5
+        ]
+    )
+    expected = [
+        [0, 1, 2, 3],  # Community 1 (includes overlapping point 3)
+        [4, 5],  # Community 2
+    ]
+    result = community_detection(embeddings, threshold=0.8, min_community_size=2)
+    assert sorted([sorted(community) for community in result]) == sorted([sorted(community) for community in expected])
+
+
+def test_community_detection_numpy_input():
+    """Test case where input is a numpy array instead of a torch tensor."""
+    embeddings = np.array(
+        [
+            [1.0, 0.0, 0.0],
+            [0.9, 0.1, 0.0],
+            [0.8, 0.2, 0.0],
+        ]
+    )
+    expected = [
+        [0, 1, 2],  # Single community
+    ]
+    result = community_detection(embeddings, threshold=0.8, min_community_size=2)
+    assert sorted([sorted(community) for community in result]) == sorted([sorted(community) for community in expected])
+
+
+def test_community_detection_large_batch_size():
+    """Test case with a large dataset and batching."""
+    embeddings = torch.rand(1000, 128)  # Random embeddings
+    result = community_detection(embeddings, threshold=0.8, min_community_size=10, batch_size=256)
+    # Check that all communities meet the minimum size requirement
+    assert all(len(community) >= 10 for community in result)
+
+
+@pytest.mark.skipif(not torch.cuda.is_available(), reason="GPU not available")
+def test_community_detection_gpu_support():
+    """Test case for GPU support (if available)."""
+    embeddings = torch.tensor(
+        [
+            [1.0, 0.0, 0.0],
+            [0.9, 0.1, 0.0],
+            [0.8, 0.2, 0.0],
+        ]
+    ).cuda()
+    expected = [
+        [0, 1, 2],  # Single community
+    ]
+    result = community_detection(embeddings, threshold=0.8, min_community_size=2)
+    assert sorted([sorted(community) for community in result]) == sorted([sorted(community) for community in expected])