Fix empty tensor issue when reading an non-existing node in sync dump mode

mindspore/ccsrc/debug/debug_services.cc

@@ -629,6 +629,7 @@ void DebugServices::ReadDumpedTensor(std::vector<std::string> backend_name, std:
       d = opendir(specific_dump_dir.c_str());
       if (d != nullptr) {
         struct dirent *dir = nullptr;
+        bool found_file = false;
         while ((dir = readdir(d)) != NULL) {
           if (dir->d_type == DT_REG) {
             std::string file_name = dir->d_name;
@@ -659,8 +660,13 @@ void DebugServices::ReadDumpedTensor(std::vector<std::string> backend_name, std:
             infile.close();
             AddToTensorData(backend_name[i], slot[i], iteration[i], device_id[i], root_graph_id[i], data_size,
                             type_name, shape, buffer, result_list);
+            found_file = true;
           }
         }
+        if (!found_file) {
+          AddToTensorData(backend_name[i], slot[i], iteration[i], device_id[i], root_graph_id[i], 0, type_name, shape,
+                          buffer, result_list);
+        }
       } else {
         MS_LOG(INFO) << "directory does not exist!";
       }

tests/ut/data/dump/gpu_dumps/golden/sync_trans_true_read_tensors_nonexist_node.expected

@@ -0,0 +1,14 @@
+-----------------------------------------------------------
+tensor_info_1 attributes:
+node name =  Default/network-WithLossCell/_backbone-AlexNet/conv3-Conv2d/Conv2D-op318
+slot =  0
+iteration =  2
+device_id =  None
+root_graph_id =  0
+is_parameter =  False
+
+tensor_data_1 attributes:
+data (printed in uint8) =  []
+size in bytes =  0
+debugger dtype =  0
+shape =  []

tests/ut/python/debugger/gpu_tests/test_sync_trans_read_tensors_nonexist_node.py

@@ -0,0 +1,86 @@
+# Copyright 2021 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+"""
+Read tensor test script for offline debugger APIs.
+"""
+
+import mindspore.offline_debug.dbg_services as d
+import numpy as np
+from dump_test_utils import compare_actual_with_expected
+
+GENERATE_GOLDEN = False
+test_name = "sync_trans_true_read_tensors_nonexist_node"
+
+
+def test_sync_trans_read_tensors_nonexist_node():
+    debugger_backend = d.DbgServices(
+        dump_file_path="../data/dump/gpu_dumps/sync_trans_true/alexnet")
+
+    _ = debugger_backend.initialize(
+        net_name="Network Name goes here!", is_sync_mode=True)
+
+    # non-existing tensor with wrong op name
+    info1 = d.TensorInfo(node_name="Default/network-WithLossCell/_backbone-AlexNet/conv3-Conv2d/Conv2D-op318",
+                         slot=0, iteration=2, device_id=0, root_graph_id=0, is_parameter=False)
+
+    tensor_info = [info1]
+
+    tensor_data = debugger_backend.read_tensors(tensor_info)
+
+    # Check the length of tensor list
+    assert len(tensor_info) == 1
+    assert len(tensor_data) == 1
+
+    print_read_tensors(tensor_info, tensor_data)
+    assert compare_actual_with_expected(test_name)
+
+
+def print_read_tensors(tensor_info, tensor_data):
+    """Print read tensors."""
+    if GENERATE_GOLDEN:
+        f_write = open(test_name + ".expected", "w")
+    else:
+        f_write = open(test_name + ".actual", "w")
+
+    for x, _ in enumerate(tensor_info):
+        f_write.write(
+            "-----------------------------------------------------------\n")
+        f_write.write("tensor_info_" + str(x + 1) + " attributes:\n")
+        f_write.write("node name =  " + tensor_info[x].node_name + "\n")
+        f_write.write("slot =  " + str(tensor_info[x].slot) + "\n")
+        f_write.write("iteration =  " + str(tensor_info[x].iteration) + "\n")
+        f_write.write("device_id =  " + str(tensor_info[x].device_id) + "\n")
+        f_write.write("root_graph_id =  " +
+                      str(tensor_info[x].root_graph_id) + "\n")
+        f_write.write("is_parameter =  " +
+                      str(tensor_info[x].is_parameter) + "\n")
+        f_write.write("\n")
+        f_write.write("tensor_data_" + str(x + 1) + " attributes:\n")
+        f_write.write("data (printed in uint8) =  " + str(np.frombuffer(
+            tensor_data[x].data_ptr, np.uint8, tensor_data[x].data_size)) + "\n")
+        py_byte_size = len(tensor_data[x].data_ptr)
+        c_byte_size = tensor_data[x].data_size
+        if c_byte_size != py_byte_size:
+            f_write.write("The python byte size of  " + str(py_byte_size) +
+                          "  does not match the C++ byte size of  " + str(c_byte_size) + "\n")
+        f_write.write("size in bytes =  " +
+                      str(tensor_data[x].data_size) + "\n")
+        f_write.write("debugger dtype =  " + str(tensor_data[x].dtype) + "\n")
+        f_write.write("shape =  " + str(tensor_data[x].shape) + "\n")
+    f_write.close()
+
+
+if __name__ == "__main__":
+    test_sync_trans_read_tensors_nonexist_node()