Add arguments checking for AddN primitive;

In JIT Fallback, getattr(interpreted_obj, 'attr') should work as interpred_obj.attr;
Support Return InterpretedNode value;
Not throw exception if getattr's default value is InterpretedNode;

mindspore/ccsrc/frontend/optimizer/clean.cc

@@ -33,6 +33,7 @@
 #include "frontend/operator/composite/composite.h"
 #include "ir/anf.h"
 #include "ir/value.h"
+#include "pipeline/jit/fallback.h"
 #include "pipeline/jit/parse/resolve.h"
 #include "utils/hash_map.h"
 #include "utils/anf_utils.h"
@@ -1059,10 +1060,19 @@ class CleanAfterOptARewriter : public BaseRewriter {
     return make_dict_node;
   }
 
+  AnfNodePtr ConvertInterpretedObjectValue(const ValueNodePtr &node, const parse::InterpretedObjectPtr &value) {
+    // Convert InterpretedObject value node to PyExecute CNode.
+    return ConvertInterpretedObjectToPyExecute(root_graph_, value, node);
+  }
+
   AnfNodePtr ConvertValueNode(const ValueNodePtr &value_node, const ValuePtr &value) override {
     const auto support_fallback_runtime = (common::GetEnv("MS_DEV_ENABLE_FALLBACK_RUNTIME") != "0");
-    if (value->isa<ValueDictionary>() && support_fallback_runtime) {
+    if (support_fallback_runtime) {
-      return RebuildValueDict(value_node, value->cast<ValueDictionaryPtr>());
+      if (value->isa<ValueDictionary>()) {
+        return RebuildValueDict(value_node, value->cast<ValueDictionaryPtr>());
+      } else if (value->isa<parse::InterpretedObject>()) {
+        return ConvertInterpretedObjectValue(value_node, value->cast<parse::InterpretedObjectPtr>());
+      }
     }
     bool need_convert = false;
     auto convert_value = ConvertValueSequenceToValueTuple(value, 0, &need_convert);

mindspore/ccsrc/pipeline/jit/static_analysis/prim.cc

@@ -1885,7 +1885,7 @@ EvalResultPtr StaticGetter(const AnalysisEnginePtr &engine, const AbstractBasePt
   }
 
   constexpr auto max_args_size = 3;
-  if (args_abs_list.size() == max_args_size) {
+  if (!support_fallback_runtime && args_abs_list.size() == max_args_size) {
     constexpr size_t default_index = 2;
     auto default_args = args_abs_list[default_index];
     if (default_args->isa<abstract::AbstractScalar>()) {

mindspore/core/ops/addn.cc

@@ -64,11 +64,16 @@ bool AddNDynShapeJoin(ShapeVector *shape1, const ShapeVector *shape2) {
 }
 
 abstract::ShapePtr AddNInferShape(const PrimitivePtr &primitive, const std::vector<AbstractBasePtr> &input_args) {
+  auto prim_name = primitive->name();
+  if (!input_args[0]->isa<abstract::AbstractTuple>() && !input_args[0]->isa<abstract::AbstractList>()) {
+    MS_EXCEPTION(TypeError) << "For '" << prim_name
+                            << "', the input data type must be list or tuple of tensors.But got:"
+                            << input_args[0]->ToString();
+  }
   auto elements = input_args[0]->isa<abstract::AbstractTuple>()
                     ? input_args[0]->cast<abstract::AbstractTuplePtr>()->elements()
                     : input_args[0]->cast<abstract::AbstractListPtr>()->elements();
-  (void)CheckAndConvertUtils::CheckInteger("input num", SizeToLong(elements.size()), kGreaterEqual, 1,
+  (void)CheckAndConvertUtils::CheckInteger("input num", SizeToLong(elements.size()), kGreaterEqual, 1, prim_name);
-                                           primitive->name());
   (void)primitive->AddAttr("N", MakeValue(SizeToLong(elements.size())));
   (void)primitive->AddAttr("n", MakeValue(SizeToLong(elements.size())));
   auto shape_0 = elements[0]->BuildShape();
@@ -91,21 +96,25 @@ abstract::ShapePtr AddNInferShape(const PrimitivePtr &primitive, const std::vect
     }
     // Join input[i] with input[0]
     if (!AddNDynShapeJoin(&output_shape, &shape_vec)) {
-      MS_EXCEPTION(ValueError) << "For '" << primitive->name() << "', input shape must be same, but got shape of input["
+      MS_EXCEPTION(ValueError) << "For '" << prim_name << "', input shape must be same, but got shape of input[" << i
-                               << i << "]: " << shape->ToString() << ", shape of input[0]: " << shape_0->ToString()
+                               << "]: " << shape->ToString() << ", shape of input[0]: " << shape_0->ToString() << ".";
-                               << ".";
     }
   }
   return std::make_shared<abstract::Shape>(output_shape);
 }
 
-TypePtr AddNInferType(const PrimitivePtr &prim, const std::vector<AbstractBasePtr> &input_args) {
+TypePtr AddNInferType(const PrimitivePtr &primitive, const std::vector<AbstractBasePtr> &input_args) {
-  auto prim_name = prim->name();
+  auto prim_name = primitive->name();
+  if (!input_args[0]->isa<abstract::AbstractTuple>() && !input_args[0]->isa<abstract::AbstractList>()) {
+    MS_EXCEPTION(TypeError) << "For '" << prim_name
+                            << "', the input data type must be list or tuple of tensors.But got:"
+                            << input_args[0]->ToString();
+  }
   auto elements = input_args[0]->isa<abstract::AbstractTuple>()
                     ? input_args[0]->cast<abstract::AbstractTuplePtr>()->elements()
                     : input_args[0]->cast<abstract::AbstractListPtr>()->elements();
   (void)CheckAndConvertUtils::CheckInteger("concat element num", SizeToLong(elements.size()), kGreaterEqual, 1,
-                                           prim->name());
+                                           prim_name);
   std::map<std::string, TypePtr> types;
   (void)types.emplace("element_0", elements[0]->BuildType());
   for (size_t i = 0; i < elements.size(); ++i) {

mindspore/python/mindspore/_extends/parse/parser.py

@@ -112,7 +112,7 @@ _unsupported_internal_type = (
 )
 
 _hybrid_type = (
-    print, enumerate, zip, map, filter, abs, round, max, min, sum, hasattr, list, tuple
+    print, enumerate, zip, map, filter, abs, round, max, min, sum, getattr, hasattr, list, tuple
 )
 
 # Unsupported python builtin type in JIT Fallback.

tests/ut/python/graph_syntax/numpy/test_parse_numpy.py

@@ -13,7 +13,6 @@
 # limitations under the License.
 # ============================================================================
 """ test_parse_numpy """
-import pytest
 import numpy as np
 from mindspore import nn, context, jit, Tensor
 
@@ -44,11 +43,7 @@ def test_use_numpy_method():
             return ret
 
     net = Net()
-    # Not raise NotImplementedError('Mindspore not supports to use the numpy ...') any more,
+    net()
-    # but raise RuntimeError('Should not use Python object in runtime...'), after support JIT Fallback.
-    with pytest.raises(RuntimeError) as err:
-        net()
-    assert "Should not use Python object in runtime" in str(err.value)
 
 
 def test_use_numpy_module():
@@ -61,11 +56,7 @@ def test_use_numpy_module():
             return ret
 
     net = Net()
-    # Not raise NotImplementedError('Mindspore not supports to use the numpy ...') any more,
+    net()
-    # but raise RuntimeError('Should not use Python object in runtime...'), after support JIT Fallback.
-    with pytest.raises(RuntimeError) as err:
-        net()
-    assert "Should not use Python object in runtime" in str(err.value)
 
 
 def test_np_calculate():

tests/ut/python/graph_syntax/python_builtin_functions/test_getattr.py

@@ -656,13 +656,10 @@ def test_getattr_numpy_array():
     @jit
     def foo():
         x = np.array([1, 2, 3, 4])
+        # Should work as: return x.shape[0]
         return getattr(x, "shape")[0]
 
-    os.environ['MS_DEV_ENABLE_FALLBACK_RUNTIME'] = '0'
+    foo()
-    with pytest.raises(TypeError) as err:
-        foo()  # Not throw error any more, should move to ST.
-    assert "Do not support to get attribute" in str(err.value)
-    os.environ['MS_DEV_ENABLE_FALLBACK_RUNTIME'] = '1'
 
 
 def test_getattr_numpy_array_2():
@@ -677,9 +674,7 @@ def test_getattr_numpy_array_2():
         x = 1
         return getattr(x, "shape", np.array([0, 1, 2, 3, 4]))
 
-    with pytest.raises(TypeError) as err:
+    foo()
-        foo()
-    assert "For 'getattr', the third input 'default' can not" in str(err.value)
 
 
 def test_getattr_for_fg_object():