Add empty mask handling within pcv.visualize.obj_size_ecdf

plantcv/plantcv/visualize/obj_size_ecdf.py

@@ -27,6 +27,16 @@ def obj_size_ecdf(mask):
     # Remove objects with areas < 1px
     areas = [i for i in areas if i >= 1.0]
 
+    # If the mask is empty, return empty chart
+    if len(areas) == 0:
+        ecdf_df = pd.DataFrame({'object_area': [], 'cumulative_probability': []})
+        chart = alt.Chart(ecdf_df).mark_circle(size=10).encode(
+            x=alt.X("object area:Q").scale(type='log'),
+            y="cumulative probability:Q",
+            tooltip=['object area', 'cumulative probability']
+        )
+        return chart.interactive()
+
     ecdf = ECDF(areas, side='right')
 
     ecdf_df = pd.DataFrame({'object area': ecdf.x[1:], 'cumulative probability': ecdf.y[1:]})

tests/plantcv/visualize/test_obj_size_ecdf.py

@@ -1,5 +1,7 @@
 """Tests for pcv.visualize.obj_size_ecdf."""
 import cv2
+import numpy as np
+import pandas as pd
 from altair.vegalite.v5.api import Chart
 from plantcv.plantcv.visualize import obj_size_ecdf
 
@@ -9,3 +11,11 @@ def test_obj_size_ecdf(visualize_test_data):
     mask = cv2.imread(visualize_test_data.small_bin_img, -1)
     fig_ecdf = obj_size_ecdf(mask=mask)
     assert isinstance(fig_ecdf, Chart)
+
+
+def test_obj_size_ecdf_empty_mask():
+    """Test for empty mask"""
+    mask = np.zeros((100, 100), dtype=np.uint8)
+    chart = obj_size_ecdf(mask=mask)
+    assert isinstance(chart.data, pd.DataFrame)
+    assert chart.data.shape[0] == 0