add design of general blob as data struct between op on server side

core/general-server/op/general_infer_helper.h

@@ -45,7 +45,7 @@ struct GeneralBlob {
     tensor_vector.clear();
   }
   
-  int GetBatchSize() {
+  int GetBatchSize() const {
     if (tensor_vector.size() > 0) {
       if (tensor_vector[0].lod.size() == 1) {
         return tensor_vector[0].lod[0].size() - 1;
@@ -58,7 +58,7 @@ struct GeneralBlob {
   }
 
   std::string ShortDebugString() const { return "Not implemented!"; }
-}
+};
 
 }  // namespace serving
 }  // namespace paddle_serving

core/general-server/op/general_infer_op.cpp

@@ -16,7 +16,6 @@
 #include <iostream>
 #include <memory>
 #include <sstream>
-#include "core/general-server/op/general_infer_helper.h"
 #include "core/general-server/op/general_infer_op.h"
 #include "core/predictor/framework/infer.h"
 #include "core/predictor/framework/memory.h"
@@ -48,7 +47,7 @@ int GeneralInferOp::inference() {
 
   const TensorVector *in = &input_blob->tensor_vector;
   TensorVector *out = butil::get_object<TensorVector>();
-  int batch_size = in->GetBatchSize();
+  int batch_size = input_blob->GetBatchSize();
 
   VLOG(2) << "infer batch size: " << batch_size;
   // infer

core/general-server/op/general_infer_op.h

@@ -25,6 +25,7 @@
 #include "paddle_inference_api.h"  // NOLINT
 #endif
 #include "core/general-server/general_model_service.pb.h"
+#include "core/general-server/op/general_infer_helper.h"
 
 namespace baidu {
 namespace paddle_serving {

core/general-server/op/general_reader_op.cpp

@@ -100,8 +100,8 @@ int GeneralReaderOp::inference() {
   VLOG(2) << "print general model config done.";
 
   // TODO(guru4elephant): how to do conditional check?
-  res->reader_status = conf_check(req, model_config);
-  if (res->reader_status != 0) {
+  int ret = conf_check(req, model_config);
+  if (ret != 0) {
     LOG(INFO) << "model conf of server:";
     resource.print_general_model_config(model_config);
     return 0;

core/general-server/op/general_response_op.cpp

@@ -47,8 +47,7 @@ int GeneralResponseOp::inference() {
   }
 
   const TensorVector *in = &input_blob->tensor_vector;
-  int batch_size = in->GetBatchSize();
-  double infer_time = in->infer_time;
+  int batch_size = input_blob->GetBatchSize();
 
   VLOG(2) << "input batch size: " << batch_size;
 
@@ -72,7 +71,7 @@ int GeneralResponseOp::inference() {
   // response inst with only fetch_var_names
   Response *res = mutable_data<Response>();
 
-  res->set_mean_infer_us(infer_time);
+  // res->set_mean_infer_us(infer_time);
 
   for (int i = 0; i < batch_size; ++i) {
     FetchInst *fetch_inst = res->add_insts();

core/general-server/op/general_text_reader_op.cpp

@@ -17,7 +17,6 @@
 #include <memory>
 #include <sstream>
 #include "core/general-server/op/general_text_reader_op.h"
-#include "core/general-server/op/general_infer_helper.h"
 #include "core/predictor/framework/infer.h"
 #include "core/predictor/framework/memory.h"
 
@@ -51,8 +50,8 @@ int GeneralTextReaderOp::inference() {
   }
 
   if (batch_size <= 0) {
-    res->reader_status = -1;
-    return 0;
+    LOG(ERROR) << "Batch size < 0";
+    return -1;
   }
 
   int var_num = req->insts(0).tensor_array_size();

core/general-server/op/general_text_reader_op.h

@@ -25,6 +25,7 @@
 #endif
 #include <string>
 #include "core/predictor/framework/resource.h"
+#include "core/general-server/op/general_infer_helper.h"
 #include "core/general-server/general_model_service.pb.h"
 #include "core/general-server/load_general_model_service.pb.h"
 

core/general-server/op/general_text_response_op.cpp

@@ -17,7 +17,6 @@
 #include <memory>
 #include <sstream>
 #include "core/general-server/op/general_text_response_op.h"
-#include "core/general-server/op/general_infer_helper.h"
 #include "core/predictor/framework/infer.h"
 #include "core/predictor/framework/memory.h"
 #include "core/predictor/framework/resource.h"
@@ -36,18 +35,18 @@ using baidu::paddle_serving::predictor::general_model::FetchInst;
 using baidu::paddle_serving::predictor::InferManager;
 using baidu::paddle_serving::predictor::PaddleGeneralModelConfig;
 
-int GeneralTextInferOp::inference() {
-  const GeneralBlob *blob_input =
+int GeneralTextResponseOp::inference() {
+  const GeneralBlob *input_blob =
       get_depend_argument<GeneralBlob>(pre_name());
 
-  if (!blob_input) {
+  if (!input_blob) {
     LOG(ERROR) << "Failed mutable depended argument, op: "
                << pre_name();
     return -1;
   }
 
-  const TensorVector *in = &blob_input->tensor_vector;
-  int batch_size = in->GetBatchSize();
+  const TensorVector *in = &input_blob->tensor_vector;
+  int batch_size = input_blob->GetBatchSize();
 
   VLOG(2) << "infer batch size: " << batch_size;
   // infer
@@ -72,7 +71,7 @@ int GeneralTextInferOp::inference() {
   // response inst with only fetch_var_names
   Response *res = mutable_data<Response>();
 
-  res->set_mean_infer_us(infer_time);
+  // res->set_mean_infer_us(infer_time);
 
   for (int i = 0; i < batch_size; ++i) {
     FetchInst *fetch_inst = res->add_insts();

core/general-server/op/general_text_response_op.h

@@ -25,6 +25,7 @@
 #include "paddle_inference_api.h"  // NOLINT
 #endif
 #include "core/general-server/general_model_service.pb.h"
+#include "core/general-server/op/general_infer_helper.h"
 
 namespace baidu {
 namespace paddle_serving {