Improved error message for invalid arrays

lib/lightgbm/dataset.rb

@@ -166,6 +166,8 @@ def construct
           ncol = data.first.count
           flat_data = data.flat_map { |v| v.fetch_values(*keys) }
         else
+          data = data.to_a
+          check_2d_array(data)
           nrow = data.count
           ncol = data.first.count
           flat_data = data.flatten

lib/lightgbm/inner_predictor.rb

@@ -49,6 +49,7 @@ def predict(data, start_iteration: 0, num_iteration: -1, raw_score: false, pred_
         data = data.to_a
         singular = !data.first.is_a?(Array)
         data = [data] if singular
+        check_2d_array(data)
       end
 
       preds, nrow =

lib/lightgbm/utils.rb

@@ -24,6 +24,13 @@ def set_verbosity(params)
       end
     end
 
+    def check_2d_array(data)
+      ncol = data.first&.size || 0
+      if !data.all? { |r| r.size == ncol }
+        raise ArgumentError, "Rows have different sizes"
+      end
+    end
+
     # for categorical, NaN and negative value are the same
     def handle_missing(data)
       data.map! { |v| v.nil? ? Float::NAN : v }

test/booster_test.rb

@@ -88,6 +88,14 @@ def test_predict_rover
     end
   end
 
+  def test_predict_array_different_sizes
+    x_test = [[1, 2], [3, 4, 5]]
+    error = assert_raises(ArgumentError) do
+      booster.predict(x_test)
+    end
+    assert_equal "Rows have different sizes", error.message
+  end
+
   def test_predict_raw_score
     x_test = [[3.7, 1.2, 7.2, 9.0], [7.5, 0.5, 7.9, 0.0]]
     expected = [0.9823112229173586, 0.9583143724610858]

test/dataset_test.rb

@@ -113,6 +113,14 @@ def test_rover
     assert_equal ["x0", "x1", "x2", "x3"], dataset.feature_name
   end
 
+  def test_array_different_sizes
+    data = [[1, 2], [3, 4, 5]]
+    error = assert_raises(ArgumentError) do
+      LightGBM::Dataset.new(data)
+    end
+    assert_equal "Rows have different sizes", error.message
+  end
+
   def test_copy
     regression_train.dup
     regression_train.clone