napari-tmidas 0.2.2__py3-none-any.whl → 0.2.4__py3-none-any.whl

napari_tmidas/_tests/test_regionprops_analysis.py

@@ -0,0 +1,547 @@
+"""
+Tests for regionprops_analysis processing function
+"""
+
+import os
+import tempfile
+from pathlib import Path
+
+import numpy as np
+import pandas as pd
+import pytest
+
+from napari_tmidas.processing_functions.regionprops_analysis import (
+    analyze_folder_regionprops,
+    extract_regionprops_folder,
+    extract_regionprops_recursive,
+    find_label_images,
+    load_label_image,
+    parse_dimensions_from_shape,
+    reset_regionprops_cache,
+)
+
+
+@pytest.fixture
+def temp_label_folder():
+    """Create a temporary folder with test label images."""
+    with tempfile.TemporaryDirectory() as tmpdir:
+        folder = Path(tmpdir) / "labels"
+        folder.mkdir()
+
+        # Create 2D label image
+        label_2d = np.zeros((100, 100), dtype=np.uint16)
+        label_2d[20:40, 20:40] = 1
+        label_2d[60:80, 60:80] = 2
+        np.save(folder / "image_2d.npy", label_2d)
+
+        # Create 3D label image (ZYX)
+        label_3d = np.zeros((10, 100, 100), dtype=np.uint16)
+        label_3d[2:5, 20:40, 20:40] = 1
+        label_3d[5:8, 60:80, 60:80] = 2
+        np.save(folder / "image_3d.npy", label_3d)
+
+        # Create 4D label image (TZYX) - 3 timepoints
+        label_4d = np.zeros((3, 10, 100, 100), dtype=np.uint16)
+        for t in range(3):
+            label_4d[t, 2:5, 20:40, 20:40] = 1
+            label_4d[t, 5:8, 60:80, 60:80] = 2
+        np.save(folder / "image_4d.npy", label_4d)
+
+        yield folder
+
+
+def test_parse_dimensions_from_shape():
+    """Test dimension parsing from image shape."""
+    # Test 2D
+    dims = parse_dimensions_from_shape((100, 100), 2)
+    assert "Y" in dims and "X" in dims
+    assert dims["Y"] == 100 and dims["X"] == 100
+
+    # Test 3D
+    dims = parse_dimensions_from_shape((10, 100, 100), 3)
+    assert "Z" in dims and "Y" in dims and "X" in dims
+    assert dims["Z"] == 10
+
+    # Test 4D
+    dims = parse_dimensions_from_shape((5, 10, 100, 100), 4)
+    assert "T" in dims and "Z" in dims
+    assert dims["T"] == 5 and dims["Z"] == 10
+
+
+def test_find_label_images(temp_label_folder):
+    """Test finding label images in a folder."""
+    files = find_label_images(str(temp_label_folder))
+    assert len(files) == 3
+    assert all(f.endswith(".npy") for f in files)
+
+
+def test_load_label_image(temp_label_folder):
+    """Test loading label images."""
+    files = find_label_images(str(temp_label_folder))
+
+    for filepath in files:
+        img = load_label_image(filepath)
+        assert isinstance(img, np.ndarray)
+        assert img.dtype in [np.uint16, np.int32, np.int64]
+
+
+def test_extract_regionprops_recursive_2d(temp_label_folder):
+    """Test extracting regionprops from 2D label image."""
+    files = find_label_images(str(temp_label_folder))
+    file_2d = [f for f in files if "2d" in f][0]
+
+    img = load_label_image(file_2d)
+    results = extract_regionprops_recursive(
+        img,
+        prefix_dims={"filename": os.path.basename(file_2d)},
+        max_spatial_dims=3,
+    )
+
+    # Should have 2 regions
+    assert len(results) == 2
+
+    # Check that results have expected keys
+    for result in results:
+        assert "filename" in result
+        assert "label" in result
+        assert "size" in result  # area is renamed to size
+        assert "centroid_y" in result
+        assert "centroid_x" in result
+        assert "bbox_min_y" in result
+        assert "bbox_max_x" in result
+
+    # Check areas (20x20 = 400 pixels each)
+    assert results[0]["size"] == 400
+    assert results[1]["size"] == 400
+
+
+def test_extract_regionprops_recursive_3d(temp_label_folder):
+    """Test extracting regionprops from 3D label image."""
+    files = find_label_images(str(temp_label_folder))
+    file_3d = [f for f in files if "3d" in f][0]
+
+    img = load_label_image(file_3d)
+    results = extract_regionprops_recursive(
+        img,
+        prefix_dims={"filename": os.path.basename(file_3d)},
+        max_spatial_dims=3,
+    )
+
+    # Should have 2 regions
+    assert len(results) == 2
+
+    # Check that results have expected keys including Z coordinates
+    for result in results:
+        assert "filename" in result
+        assert "label" in result
+        assert "size" in result  # area is renamed to size
+        assert "centroid_z" in result
+        assert "centroid_y" in result
+        assert "centroid_x" in result
+        assert "bbox_min_z" in result
+        assert "bbox_max_z" in result
+
+    # Check volumes (3 z-slices x 20x20 = 1200 voxels each)
+    assert results[0]["size"] == 1200
+    assert results[1]["size"] == 1200
+
+
+def test_extract_regionprops_recursive_4d(temp_label_folder):
+    """Test extracting regionprops from 4D label image with time dimension."""
+    files = find_label_images(str(temp_label_folder))
+    file_4d = [f for f in files if "4d" in f][0]
+
+    img = load_label_image(file_4d)
+    results = extract_regionprops_recursive(
+        img,
+        prefix_dims={"filename": os.path.basename(file_4d)},
+        max_spatial_dims=3,
+    )
+
+    # Should have 2 regions x 3 timepoints = 6 results
+    assert len(results) == 6
+
+    # Check that results have time dimension
+    for result in results:
+        assert "T" in result
+        assert "label" in result
+        assert result["T"] in [0, 1, 2]
+
+    # Check that we have 2 labels per timepoint
+    for t in range(3):
+        t_results = [r for r in results if r["T"] == t]
+        assert len(t_results) == 2
+        labels = [r["label"] for r in t_results]
+        assert sorted(labels) == [1, 2]
+
+
+def test_analyze_folder_regionprops(temp_label_folder):
+    """Test analyzing entire folder and saving to CSV."""
+    output_csv = temp_label_folder.parent / "test_regionprops.csv"
+
+    df = analyze_folder_regionprops(
+        folder_path=str(temp_label_folder),
+        output_csv=str(output_csv),
+        max_spatial_dims=3,
+        dimension_order="Auto",
+    )
+
+    # Check that CSV was created
+    assert output_csv.exists()
+
+    # Check DataFrame structure
+    assert isinstance(df, pd.DataFrame)
+    assert len(df) > 0
+    assert "filename" in df.columns
+    assert "label" in df.columns
+    assert "size" in df.columns  # area is renamed to size
+
+    # Check that we have results from all files
+    filenames = df["filename"].unique()
+    assert len(filenames) == 3
+
+    # Load CSV and verify it matches the DataFrame
+    df_loaded = pd.read_csv(output_csv)
+    assert len(df_loaded) == len(df)
+
+
+def test_extract_regionprops_folder_integration(temp_label_folder):
+    """Test the full batch processing function."""
+    # Reset cache first
+    reset_regionprops_cache()
+
+    # Load one of the images
+    files = find_label_images(str(temp_label_folder))
+    img = load_label_image(files[0])
+
+    # Mock the filepath in the call stack by calling from a function
+    # that has filepath in its locals
+    def mock_process(filepath, image):
+        return extract_regionprops_folder(
+            image,
+            max_spatial_dims=3,
+            dimension_order="Auto",
+            overwrite_existing=True,
+        )
+
+    result = mock_process(files[0], img)
+
+    # Function should return None (only generates CSV)
+    assert result is None
+
+    # Check that CSV was created
+    folder_name = temp_label_folder.name
+    parent_dir = temp_label_folder.parent
+    output_csv = parent_dir / f"{folder_name}_regionprops.csv"
+    assert output_csv.exists()
+
+    # Load and verify CSV
+    df = pd.read_csv(output_csv)
+    assert len(df) > 0
+    assert "filename" in df.columns
+
+
+def test_extract_regionprops_folder_no_overwrite(temp_label_folder):
+    """Test that function respects overwrite_existing flag."""
+    reset_regionprops_cache()
+
+    # Create a CSV file first
+    folder_name = temp_label_folder.name
+    parent_dir = temp_label_folder.parent
+    output_csv = parent_dir / f"{folder_name}_regionprops.csv"
+
+    # Create dummy CSV
+    pd.DataFrame({"test": [1, 2, 3]}).to_csv(output_csv, index=False)
+    initial_mtime = output_csv.stat().st_mtime
+
+    # Load image
+    files = find_label_images(str(temp_label_folder))
+    img = load_label_image(files[0])
+
+    # Process with overwrite_existing=False
+    def mock_process(filepath, image):
+        return extract_regionprops_folder(
+            image,
+            max_spatial_dims=3,
+            dimension_order="Auto",
+            overwrite_existing=False,
+        )
+
+    _ = mock_process(files[0], img)  # result unused, just checking behavior
+
+    # File should not have been modified
+    assert output_csv.stat().st_mtime == initial_mtime
+
+    # Now test with overwrite=True
+    reset_regionprops_cache()
+
+    # Create a mock process that properly sets filepath in locals
+    def mock_process_with_overwrite(filepath, image):
+        return extract_regionprops_folder(
+            image,
+            max_spatial_dims=3,
+            dimension_order="Auto",
+            overwrite_existing=True,
+        )
+
+    _ = mock_process_with_overwrite(files[0], img)  # result unused
+
+    # Verify the CSV was updated properly
+    if output_csv.exists():
+        df = pd.read_csv(output_csv)
+        # Either it should have the proper structure, or it should have been skipped
+        # (because we can't determine filepath in test context)
+        if len(df.columns) > 1:
+            assert "filename" in df.columns
+
+
+def test_reset_cache():
+    """Test that cache reset works."""
+    from napari_tmidas.processing_functions.regionprops_analysis import (
+        _REGIONPROPS_CSV_FILES,
+    )
+
+    # First clear any existing cache
+    reset_regionprops_cache()
+    assert len(_REGIONPROPS_CSV_FILES) == 0
+
+    # Add a CSV file to the cache
+    _REGIONPROPS_CSV_FILES["/test/folder/test_regionprops.csv"] = True
+    assert len(_REGIONPROPS_CSV_FILES) == 1
+
+    # Reset cache
+    reset_regionprops_cache()
+    assert len(_REGIONPROPS_CSV_FILES) == 0
+
+
+def test_empty_label_image():
+    """Test handling of empty label images."""
+    empty_img = np.zeros((100, 100), dtype=np.uint16)
+
+    results = extract_regionprops_recursive(
+        empty_img,
+        prefix_dims={"filename": "empty.npy"},
+        max_spatial_dims=3,
+    )
+
+    # Should return empty list
+    assert len(results) == 0
+
+
+def test_single_region():
+    """Test handling of label image with single region."""
+    img = np.zeros((50, 50), dtype=np.uint16)
+    img[10:30, 10:30] = 5  # Single region with label 5
+
+    results = extract_regionprops_recursive(
+        img,
+        prefix_dims={"filename": "single.npy"},
+        max_spatial_dims=3,
+    )
+
+    assert len(results) == 1
+    assert results[0]["label"] == 5
+    assert results[0]["size"] == 400  # 20x20
+
+
+def test_dimension_order_tzyx():
+    """Test that dimension_order correctly identifies T and Z dimensions."""
+    # Create TZYX image (3 timepoints, 5 z-slices, 50x50)
+    label_tzyx = np.zeros((3, 5, 50, 50), dtype=np.uint16)
+    for t in range(3):
+        label_tzyx[t, 1:4, 10:20, 10:20] = 1
+        label_tzyx[t, 2:4, 30:40, 30:40] = 2
+
+    # Extract with explicit dimension order
+    results = extract_regionprops_recursive(
+        label_tzyx,
+        prefix_dims={"filename": "test_tzyx.npy"},
+        current_dim=0,
+        max_spatial_dims=3,
+        dimension_order="TZYX",
+    )
+
+    # Should have 2 regions x 3 timepoints = 6 results
+    assert len(results) == 6
+
+    # Check that T dimension is properly identified
+    for result in results:
+        assert "T" in result
+        assert result["T"] in [0, 1, 2]
+        assert "filename" in result
+        assert "label" in result
+
+
+def test_dimension_order_czyx():
+    """Test that dimension_order correctly identifies C and Z dimensions."""
+    # Create CZYX image (2 channels, 5 z-slices, 50x50)
+    label_czyx = np.zeros((2, 5, 50, 50), dtype=np.uint16)
+    for c in range(2):
+        label_czyx[c, 1:4, 10:20, 10:20] = 1
+        label_czyx[c, 2:4, 30:40, 30:40] = 2
+
+    # Extract with explicit dimension order
+    results = extract_regionprops_recursive(
+        label_czyx,
+        prefix_dims={"filename": "test_czyx.npy"},
+        current_dim=0,
+        max_spatial_dims=3,
+        dimension_order="CZYX",
+    )
+
+    # Should have 2 regions x 2 channels = 4 results
+    assert len(results) == 4
+
+    # Check that C dimension is properly identified
+    for result in results:
+        assert "C" in result
+        assert result["C"] in [0, 1]
+        assert "filename" in result
+        assert "label" in result
+
+
+def test_dimension_order_tczyx():
+    """Test that dimension_order correctly identifies T, C, and Z dimensions."""
+    # Create TCZYX image (2 timepoints, 2 channels, 3 z-slices, 30x30)
+    label_tczyx = np.zeros((2, 2, 3, 30, 30), dtype=np.uint16)
+    for t in range(2):
+        for c in range(2):
+            label_tczyx[t, c, 1:3, 5:15, 5:15] = 1
+            label_tczyx[t, c, 1:2, 20:25, 20:25] = 2
+
+    # Extract with explicit dimension order
+    results = extract_regionprops_recursive(
+        label_tczyx,
+        prefix_dims={"filename": "test_tczyx.npy"},
+        current_dim=0,
+        max_spatial_dims=3,
+        dimension_order="TCZYX",
+    )
+
+    # Should have 2 regions x 2 timepoints x 2 channels = 8 results
+    assert len(results) == 8
+
+    # Check that T and C dimensions are properly identified
+    for result in results:
+        assert "T" in result
+        assert result["T"] in [0, 1]
+        assert "C" in result
+        assert result["C"] in [0, 1]
+        assert "filename" in result
+        assert "label" in result
+
+
+def test_extract_regionprops_summary_basic():
+    """Test summary statistics extraction without grouping by dimensions."""
+    # Create a simple 2D label image
+    label_2d = np.zeros((100, 100), dtype=np.uint16)
+    label_2d[10:20, 10:20] = 1  # 100 pixels
+    label_2d[30:50, 30:50] = 2  # 400 pixels
+    label_2d[60:70, 60:80] = 3  # 200 pixels
+
+    # Create intensity image
+    intensity_2d = np.random.rand(100, 100) * 100
+
+    # Extract regionprops first to get individual values
+    results = extract_regionprops_recursive(
+        label_2d,
+        intensity_image=intensity_2d,
+        prefix_dims={"filename": "test.npy"},
+        max_spatial_dims=2,
+        properties=[
+            "label",
+            "area",
+            "mean_intensity",
+            "median_intensity",
+            "std_intensity",
+        ],
+    )
+
+    # Convert to DataFrame
+    df = pd.DataFrame(results)
+
+    # Calculate expected summary statistics (note: 'area' is renamed to 'size')
+    expected_count = len(df)
+    expected_size_sum = df["size"].sum()
+    expected_size_mean = df["size"].mean()
+    expected_size_median = df["size"].median()
+
+    # Verify we have 3 labels
+    assert expected_count == 3
+    assert expected_size_sum == 700  # 100 + 400 + 200
+    assert expected_size_mean == pytest.approx(233.333, rel=0.01)
+    assert expected_size_median == 200
+
+
+def test_extract_regionprops_summary_with_grouping():
+    """Test summary statistics extraction with grouping by dimensions."""
+    # Create 4D label image (T=2, Z=3, Y=50, X=50)
+    label_4d = np.zeros((2, 3, 50, 50), dtype=np.uint16)
+
+    # T=0: 2 labels
+    label_4d[0, 1:3, 10:20, 10:20] = 1  # 200 pixels
+    label_4d[0, 1:2, 30:40, 30:40] = 2  # 100 pixels
+
+    # T=1: 3 labels
+    label_4d[1, 1:3, 10:15, 10:15] = 1  # 50 pixels
+    label_4d[1, 1:2, 20:30, 20:30] = 2  # 100 pixels
+    label_4d[1, 1:2, 35:45, 35:45] = 3  # 100 pixels
+
+    # Extract with dimension order
+    results = extract_regionprops_recursive(
+        label_4d,
+        intensity_image=None,
+        prefix_dims={"filename": "test_4d.npy"},
+        current_dim=0,
+        max_spatial_dims=3,
+        dimension_order="TZYX",
+        properties=["label", "area"],
+    )
+
+    # Convert to DataFrame
+    df = pd.DataFrame(results)
+
+    # Group by T dimension (note: 'area' is renamed to 'size')
+    summary_stats = []
+    for t, group in df.groupby("T"):
+        summary_stats.append(
+            {
+                "T": t,
+                "label_count": len(group),
+                "size_sum": group["size"].sum(),
+                "size_mean": group["size"].mean(),
+            }
+        )
+
+    # Verify T=0 has 2 labels
+    t0_stats = [s for s in summary_stats if s["T"] == 0][0]
+    assert t0_stats["label_count"] == 2
+    assert t0_stats["size_sum"] == 300  # 200 + 100
+
+    # Verify T=1 has 3 labels
+    t1_stats = [s for s in summary_stats if s["T"] == 1][0]
+    assert t1_stats["label_count"] == 3
+    assert t1_stats["size_sum"] == 250  # 50 + 100 + 100
+
+
+def test_summary_statistics_calculations():
+    """Test that summary statistics are calculated correctly."""
+    # Simple test data
+    data = np.array([10, 20, 30, 40, 50])
+
+    # Expected values
+    expected_sum = 150
+    expected_mean = 30.0
+    expected_median = 30.0
+    expected_std = np.std(data, ddof=1)  # pandas uses ddof=1
+
+    # Verify using pandas
+    df = pd.DataFrame({"values": data})
+    assert df["values"].sum() == expected_sum
+    assert df["values"].mean() == expected_mean
+    assert df["values"].median() == expected_median
+    assert df["values"].std() == pytest.approx(expected_std, rel=0.01)
+
+
+if __name__ == "__main__":
+    pytest.main([__file__, "-v"])

napari_tmidas/_tests/test_registry.py

@@ -63,5 +63,73 @@ class TestBatchProcessingRegistry:
         for t in threads:
             t.join()
 
-        assert len(results) == 10
-        assert len(BatchProcessingRegistry.list_functions()) == 10
+    def test_get_function_info_nonexistent(self):
+        """Test getting info for non-existent function"""
+        info = BatchProcessingRegistry.get_function_info("NonExistent")
+        assert info is None
+
+    def test_register_with_none_parameters(self):
+        """Test registering function with None parameters (should convert to empty dict)"""
+
+        @BatchProcessingRegistry.register(
+            name="None Params Function",
+            suffix="_none",
+            description="Function with None parameters",
+        )
+        def none_params_func(image):
+            return image
+
+        info = BatchProcessingRegistry.get_function_info(
+            "None Params Function"
+        )
+        assert info["parameters"] == {}
+        assert info["suffix"] == "_none"
+        assert info["description"] == "Function with None parameters"
+
+    def test_register_minimal(self):
+        """Test registering function with minimal parameters"""
+
+        @BatchProcessingRegistry.register(name="Minimal Function")
+        def minimal_func(image):
+            return image
+
+        info = BatchProcessingRegistry.get_function_info("Minimal Function")
+        assert info is not None
+        assert callable(info["func"])
+
+    def test_register_with_complex_parameters(self):
+        """Test registering function with complex parameter metadata"""
+
+        @BatchProcessingRegistry.register(
+            name="Complex Params Function",
+            suffix="_complex",
+            description="Function with complex parameters",
+            parameters={
+                "param1": {
+                    "type": int,
+                    "default": 5,
+                    "min": 1,
+                    "max": 10,
+                    "description": "Parameter description",
+                },
+                "param2": {
+                    "type": str,
+                    "default": "default_value",
+                    "description": "String parameter",
+                },
+            },
+        )
+        def complex_func(image, param1=5, param2="default"):
+            return image
+
+        info = BatchProcessingRegistry.get_function_info(
+            "Complex Params Function"
+        )
+        assert info["parameters"]["param1"]["type"] is int
+        assert info["parameters"]["param1"]["default"] == 5
+        assert info["parameters"]["param1"]["min"] == 1
+        assert info["parameters"]["param1"]["max"] == 10
+        assert (
+            info["parameters"]["param1"]["description"]
+            == "Parameter description"
+        )