allow empty category in stats (#443)

cogeotiff · Oct 21, 2021 · 1e24c85 · 1e24c85
1 parent d61ed0b
commit 1e24c85
Show file tree

Hide file tree

Showing 3 changed files with 14 additions and 1 deletion.
diff --git a/CHANGES.md b/CHANGES.md
@@ -1,3 +1,7 @@
+# 3.0.0a2 (2021-10-21)
+
+* Allow `rio_tiler.utils.get_array_statistics` to return `0` for unfound category, instead of raising an error (https://github.com/cogeotiff/rio-tiler/pull/443)
+
 # 3.0.0a1 (2021-10-20)
 
 **breaking changes**

diff --git a/rio_tiler/utils.py b/rio_tiler/utils.py
@@ -140,7 +140,7 @@ def get_array_statistics(
                 numpy.array(categories).astype(keys.dtype) if categories else keys
             ).tolist()
             histogram = [
-                [out_dict[x] for x in h_keys],
+                [out_dict.get(x, 0) for x in h_keys],
                 h_keys,
             ]
         else:

diff --git a/tests/test_utils.py b/tests/test_utils.py
@@ -460,6 +460,15 @@ def test_get_array_statistics():
     assert len(stats[0]["histogram"][0]) == stats[0]["unique"]
     assert len(stats[0]["histogram"][1]) == stats[0]["unique"]
 
+    # histogram return only the categories passed
     stats = utils.get_array_statistics(arr, categorical=True, categories=[1, 10, 12])
     assert len(stats[0]["histogram"][0]) == 3
     assert len(stats[0]["histogram"][1]) == 3
+
+    # test if providing a category not in the data (1000000)
+    stats = utils.get_array_statistics(
+        arr, categorical=True, categories=[1, 10, 12, 1000000]
+    )
+    assert len(stats[0]["histogram"][0]) == 4
+    assert len(stats[0]["histogram"][1]) == 4
+    assert stats[0]["histogram"][0][3] == 0.0  # there is no value 1000000