Merge pull request #405 from MRXLT/0.2.0-cherry-v2

[cherry-pick to 0.2.0] cherry pick PR 399 400 401 402 403

README.md

@@ -228,14 +228,6 @@ curl -H "Content-Type:application/json" -X POST -d '{"url": "https://paddle-serv
 | Description        | Get click probability from a feature vector of item          |
 
 
-
-| Key                | Value                                                        |
-| :----------------- | :----------------------------------------------------------- |
-| Model Name         | DNN-CTR(with cube)                                           |
-| URL                | https://paddle-serving.bj.bcebos.com/unittest/ctr_cube_unittest.tar.gz                            |
-| Client/Server Code | https://github.com/PaddlePaddle/Serving/tree/develop/python/examples/criteo_ctr_with_cube |
-| Description        | Get click probability from a feature vector of item          |
-
 <h2 align="center">Document</h2>
 
 ### New to Paddle Serving

README_CN.md

@@ -268,14 +268,6 @@ curl -H "Content-Type:application/json" -X POST -d '{"url": "https://paddle-serv
 
 
 
-| Key                | Value                                                        |
-| :----------------- | :----------------------------------------------------------- |
-| 模型名         | DNN-CTR(with cube)                                           |
-| 下载链接               | https://paddle-serving.bj.bcebos.com/unittest/ctr_cube_unittest.tar.gz                            |
-| 客户端/服务端代码 | https://github.com/PaddlePaddle/Serving/tree/develop/python/examples/criteo_ctr_with_cube |
-| 介绍        | 从项目的特征向量中获得点击概率         |
-
-
 <h2 align="center">文档</h2>
 
 ### 新手教程

core/predictor/framework/infer.h

@@ -17,6 +17,7 @@
 #include <sys/types.h>
 #include <unistd.h>
 #include <string>
+#include <utility>
 #include <vector>
 #include "core/predictor/common/inner_common.h"
 #include "core/predictor/framework/bsf.h"
@@ -588,12 +589,14 @@ class VersionedInferEngine : public InferEngine {
       LOG(ERROR) << "Failed generate engine with type:" << engine_type;
       return -1;
     }
-
+    VLOG(2) << "FLGS_logtostderr " << FLAGS_logtostderr;
+    int tmp = FLAGS_logtostderr;
     if (engine->proc_initialize(conf, version) != 0) {
       LOG(ERROR) << "Failed initialize engine, type:" << engine_type;
       return -1;
     }
-
+    VLOG(2) << "FLGS_logtostderr " << FLAGS_logtostderr;
+    FLAGS_logtostderr = tmp;
     auto r = _versions.insert(std::make_pair(engine->version(), engine));
     if (!r.second) {
       LOG(ERROR) << "Failed insert item: " << engine->version()
@@ -760,7 +763,6 @@ class InferManager {
       LOG(ERROR) << "failed load infer config, path: " << path << "/" << file;
       return -1;
     }
-
     size_t engine_num = model_toolkit_conf.engines_size();
     for (size_t ei = 0; ei < engine_num; ++ei) {
       std::string engine_name = model_toolkit_conf.engines(ei).name();
@@ -769,12 +771,10 @@ class InferManager {
         LOG(ERROR) << "Failed generate versioned engine: " << engine_name;
         return -1;
       }
-
       if (engine->proc_initialize(model_toolkit_conf.engines(ei)) != 0) {
         LOG(ERROR) << "Failed initialize version engine, name:" << engine_name;
         return -1;
       }
-
       auto r = _map.insert(std::make_pair(engine_name, engine));
       if (!r.second) {
         LOG(ERROR) << "Failed insert item: " << engine_name;
@@ -782,7 +782,6 @@ class InferManager {
       }
       LOG(WARNING) << "Succ proc initialize engine: " << engine_name;
     }
-
     return 0;
   }
 

doc/IMDB_GO_CLIENT.md

@@ -21,37 +21,10 @@ wget https://paddle-serving.bj.bcebos.com/data/text_classification/imdb_serving_
 tar -xzf imdb_serving_example.tar.gz
 ```
 
-### Server Side Code
-
-```python
-# test_server_go.py
-import os
-import sys
-from paddle_serving_server import OpMaker
-from paddle_serving_server import OpSeqMaker
-from paddle_serving_server import Server
-
-op_maker = OpMaker()
-read_op = op_maker.create('general_text_reader')
-general_infer_op = op_maker.create('general_infer')
-general_response_op = op_maker.create('general_text_response')
-
-op_seq_maker = OpSeqMaker()
-op_seq_maker.add_op(read_op)
-op_seq_maker.add_op(general_infer_op)
-op_seq_maker.add_op(general_response_op)
-
-server = Server()
-server.set_op_sequence(op_seq_maker.get_op_sequence())
-server.load_model_config(sys.argv[1])
-server.prepare_server(workdir="work_dir1", port=9292, device="cpu")
-server.run_server()
-```
-
 ### Start Server
 
 ``` shell
-python test_server_go.py ./serving_server_model/ 9292
+python -m paddle_serving_server.serve --model ./serving_server_model/ --port 9292
 ```
 
 ### Client code example

doc/IMDB_GO_CLIENT_CN.md

@@ -20,106 +20,79 @@ wget https://paddle-serving.bj.bcebos.com/data/text_classification/imdb_serving_
 tar -xzf imdb_serving_example.tar.gz
 ```
 
-### 服务器端代码
-
-```python
-# test_server_go.py
-import os
-import sys
-from paddle_serving_server import OpMaker
-from paddle_serving_server import OpSeqMaker
-from paddle_serving_server import Server
-
-op_maker = OpMaker ()
-read_op = op_maker.create ('general_text_reader')
-general_infer_op = op_maker.create ('general_infer')
-general_response_op = op_maker.create ('general_text_response')
-
-op_seq_maker = OpSeqMaker ()
-op_seq_maker.add_op (read_op)
-op_seq_maker.add_op (general_infer_op)
-op_seq_maker.add_op (general_response_op)
-
-server = Server ()
-server.set_op_sequence (op_seq_maker.get_op_sequence ())
-server.load_model_config (sys.argv [1])
-server.prepare_server (workdir = "work_dir1", port = 9292, device = "cpu")
-server.run_server ()
-```
-
 ### 启动服务器
 
 ```shell
-python test_server_go.py ./serving_server_model/ 9292
+python -m paddle_serving_server.serve --model ./serving_server_model/ --port 9292
 ```
 
 ### 客户端代码示例
 
-```go
+``` go
 // imdb_client.go
 package main
 
 import (
-       "io"
-       "fmt"
-       "strings"
-       "bufio"
-       "strconv"
-       "os"
-       serving_client "github.com/PaddlePaddle/Serving/go/serving_client"
+       "io"
+       "fmt"
+       "strings"
+       "bufio"
+       "strconv"
+       "os"
+       serving_client "github.com/PaddlePaddle/Serving/go/serving_client"
 )
 
-func main () {
-     var config_file_path string
-     config_file_path = os.Args [1]
-     handle: = serving_client.LoadModelConfig (config_file_path)
-     handle = serving_client.Connect ("127.0.0.1", "9292", handle)
+func main() {
+     var config_file_path string
+     config_file_path = os.Args[1]
+     handle := serving_client.LoadModelConfig(config_file_path)
+     handle = serving_client.Connect("127.0.0.1", "9292", handle)
 
-     test_file_path: = os.Args [2]
-     fi, err: = os.Open (test_file_path)
-     if err! = nil {
-     fmt.Print (err)
-     }
+     test_file_path := os.Args[2]
+     fi, err := os.Open(test_file_path)
+     if err != nil {
+     	fmt.Print(err)
+     }
 
-     defer fi.Close ()
-     br: = bufio.NewReader (fi)
+     defer fi.Close()
+     br := bufio.NewReader(fi)
 
-     fetch: = [] string {"cost", "acc", "prediction"}
+     fetch := []string{"cost", "acc", "prediction"}     
 
-     var result map [string] [] float32
+     var result map[string][]float32
 
-     for {
-     line, err: = br.ReadString ('\ n')
-if err == io.EOF {
-break
-}
+     for {
+     	 line, err := br.ReadString('\n')
+	 if err == io.EOF {
+	       break
+	 }
 
-line = strings.Trim (line, "\ n")
+	 line = strings.Trim(line, "\n")
 
-var words = [] int64 {}
+	 var words = []int64{}
 
-s: = strings.Split (line, "")
-value, err: = strconv.Atoi (s [0])
-var feed_int_map map [string] [] int64
-       
-for _, v: = range s [1: value + 1] {
-int_v, _: = strconv.Atoi (v)
-words = append (words, int64 (int_v))
-}
+	 s := strings.Split(line, " ")
+	 value, err := strconv.Atoi(s[0])
+	 var feed_int_map map[string][]int64
        
-label, err: = strconv.Atoi (s [len (s) -1])
+	 for _, v := range s[1:value + 1] {
+	       int_v, _ := strconv.Atoi(v)
+	       words = append(words, int64(int_v))
+	 }
 
-if err! = nil {
-panic (err)
-}
+	 label, err := strconv.Atoi(s[len(s)-1])
 
-feed_int_map = map [string] [] int64 {}
-feed_int_map ["words"] = words
-feed_int_map ["label"] = [] int64 {int64 (label)}
-Ranch
-result = serving_client.Predict (handle, feed_int_map, fetch)
-fmt.Println (result ["prediction"] [1], "\ t", int64 (label))
-    }
+	 if err != nil {
+	       panic(err)
+	 }
+
+	 feed_int_map = map[string][]int64{}
+	 feed_int_map["words"] = words
+	 feed_int_map["label"] = []int64{int64(label)}
+	 
+	 result = serving_client.Predict(handle, feed_int_map, fetch)
+	 fmt.Println(result["prediction"][1], "\t", int64(label))
+    }
 }
 ```
 

doc/Makefile

+# Minimal makefile for Sphinx documentation
+#
+
+# You can set these variables from the command line, and also
+# from the environment for the first two.
+SPHINXOPTS    ?=
+SPHINXBUILD   ?= sphinx-build
+SOURCEDIR     = source
+BUILDDIR      = build
+
+# Put it first so that "make" without argument is like "make help".
+help:
+	@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
+
+.PHONY: help Makefile
+
+# Catch-all target: route all unknown targets to Sphinx using the new
+# "make mode" option.  $(O) is meant as a shortcut for $(SPHINXOPTS).
+%: Makefile
+	@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)

doc/index.rst

+:github_url: https://github.com/PaddlePaddle/Serving
+
+.. mdinclude:: md/logo.md
+
+Paddle Serving
+==============
+.. toctree::
+     :maxdepth: 1
+     :caption: Readme
+     :hidden:
+
+.. _instruction: instruction.html
+

doc/requirements.txt

+sphinx==2.1.0
+mistune
+sphinx_rtd_theme
+paddlepaddle>=1.6

go/client_app/imdb_client.go

@@ -28,7 +28,7 @@ func main() {
      var config_file_path string
      config_file_path = os.Args[1]
      handle := serving_client.LoadModelConfig(config_file_path)
-     handle = serving_client.Connect("127.0.0.1", "9393", handle)
+     handle = serving_client.Connect("127.0.0.1", "9292", handle)
 
      test_file_path := os.Args[2]
      fi, err := os.Open(test_file_path)

python/examples/imdb/benchmark_batch.py

@@ -61,4 +61,7 @@ def single_func(idx, resource):
 
 multi_thread_runner = MultiThreadRunner()
 result = multi_thread_runner.run(single_func, args.thread, {})
-print(result)
+avg_cost = 0
+for cost in result[0]:
+    avg_cost += cost
+print("total cost of each thread".format(avg_cost / args.thread))

python/paddle_serving_client/__init__.py

@@ -93,7 +93,7 @@ class Client(object):
         lib_path = os.path.dirname(paddle_serving_client.__file__)
         client_path = os.path.join(lib_path, 'serving_client.so')
         lib_path = os.path.join(lib_path, 'lib')
-        os.popen('patchelf --set-rpath {} {}'.format(lib_path, client_path))
+        os.system('patchelf --set-rpath {} {}'.format(lib_path, client_path))
 
     def load_client_config(self, path):
         from .serving_client import PredictorClient

python/paddle_serving_client/version.py

@@ -12,6 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 """ Paddle Serving Client version string """
-serving_client_version = "0.1.3"
-serving_server_version = "0.1.3"
-module_proto_version = "0.1.3"
+serving_client_version = "0.2.0"
+serving_server_version = "0.2.0"
+module_proto_version = "0.2.0"

python/paddle_serving_server/__init__.py

@@ -211,6 +211,7 @@ class Server(object):
         tar_name = floder_name + ".tar.gz"
         bin_url = "https://paddle-serving.bj.bcebos.com/bin/" + tar_name
         self.server_path = os.path.join(self.module_path, floder_name)
+
         if not os.path.exists(self.server_path):
             print('Frist time run, downloading PaddleServing components ...')
             r = os.system('wget ' + bin_url + ' --no-check-certificate')

python/paddle_serving_server/version.py

@@ -12,6 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 """ Paddle Serving Client version string """
-serving_client_version = "0.1.3"
-serving_server_version = "0.1.3"
-module_proto_version = "0.1.3"
+serving_client_version = "0.2.0"
+serving_server_version = "0.2.0"
+module_proto_version = "0.2.0"

python/paddle_serving_server_gpu/__init__.py

@@ -20,6 +20,7 @@ import google.protobuf.text_format
 import tarfile
 import socket
 import paddle_serving_server_gpu as paddle_serving_server
+import time
 from .version import serving_server_version
 from contextlib import closing
 import argparse
@@ -115,6 +116,7 @@ class Server(object):
         self.reload_interval_s = 10
         self.module_path = os.path.dirname(paddle_serving_server.__file__)
         self.cur_path = os.getcwd()
+        self.check_cuda
         self.use_local_bin = False
         self.gpuid = 0
 
@@ -141,6 +143,13 @@ class Server(object):
             self.use_local_bin = True
             self.bin_path = os.environ["SERVING_BIN"]
 
+    def check_cuda(self):
+        r = os.system("whereis cuda")
+        if r != 0:
+            raise SystemExit(
+                "CUDA not found, please check your environment or use cpu version by \"pip install paddle_serving_server\""
+            )
+
     def set_gpuid(self, gpuid=0):
         self.gpuid = gpuid
 
@@ -215,11 +224,21 @@ class Server(object):
         os.chdir(self.module_path)
         need_download = False
         device_version = "serving-gpu-"
-        floder_name = device_version + serving_server_version
-        tar_name = floder_name + ".tar.gz"
+        folder_name = device_version + serving_server_version
+        tar_name = folder_name + ".tar.gz"
         bin_url = "https://paddle-serving.bj.bcebos.com/bin/" + tar_name
-        self.server_path = os.path.join(self.module_path, floder_name)
+        self.server_path = os.path.join(self.module_path, folder_name)
+
+        download_flag = "{}/{}.is_download".format(self.module_path,
+                                                   folder_name)
+        if os.path.exists(download_flag):
+            os.chdir(self.cur_path)
+            self.bin_path = self.server_path + "/serving"
+            return
+
         if not os.path.exists(self.server_path):
+            os.system("touch {}/{}.is_download".format(self.module_path,
+                                                       folder_name))
             print('Frist time run, downloading PaddleServing components ...')
             r = os.system('wget ' + bin_url + ' --no-check-certificate')
             if r != 0:
@@ -318,4 +337,8 @@ class Server(object):
                       self.gpuid,)
         print("Going to Run Comand")
         print(command)
+        # wait for other process to download server bin
+        while not os.path.exists(self.server_path):
+            time.sleep(1)
+
         os.system(command)

python/paddle_serving_server_gpu/version.py

@@ -12,6 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 """ Paddle Serving Client version string """
-serving_client_version = "0.1.3"
-serving_server_version = "0.1.3"
-module_proto_version = "0.1.3"
+serving_client_version = "0.2.0"
+serving_server_version = "0.2.0"
+module_proto_version = "0.2.0"

python/paddle_serving_server_gpu/web_service.py

@@ -66,6 +66,12 @@ class WebService(object):
         return server
 
     def _launch_rpc_service(self, service_idx):
+        if service_idx == 0:
+            self.rpc_service_list[service_idx].check_local_bin()
+            if not self.rpc_service_list[service_idx].use_local_bin:
+                self.rpc_service_list[service_idx].download_bin()
+        else:
+            time.sleep(3)
         self.rpc_service_list[service_idx].run_server()
 
     def prepare_server(self, workdir="", port=9393, device="gpu", gpuid=0):
@@ -160,15 +166,6 @@ class WebService(object):
             if not isinstance(result, dict) and result == -1:
                 result = {"result": "Request Value Error"}
             return result
-            '''
-            feed, fetch = self.preprocess(request.json, request.json["fetch"])
-            if "fetch" in feed:
-                del feed["fetch"]
-            fetch_map = client.predict(feed=feed, fetch=fetch)
-            fetch_map = self.postprocess(
-                feed=request.json, fetch=fetch, fetch_map=fetch_map)
-            return fetch_map
-            '''
 
         app_instance.run(host="0.0.0.0",
                          port=self.port,

tools/serving_build.sh

@@ -176,6 +176,7 @@ function python_run_criteo_ctr_with_cube() {
     local TYPE=$1
     yum install -y bc >/dev/null
     cd criteo_ctr_with_cube # pwd: /Serving/python/examples/criteo_ctr_with_cube
+    export SERVING_BIN=${SERVING_WORKDIR}/build-server-${TYPE}/core/general-server/serving
     case $TYPE in
         CPU)
             check_cmd "wget https://paddle-serving.bj.bcebos.com/unittest/ctr_cube_unittest.tar.gz"
@@ -191,6 +192,7 @@ function python_run_criteo_ctr_with_cube() {
             sh cube_prepare.sh &
             check_cmd "mkdir work_dir1 && cp cube/conf/cube.conf ./work_dir1/"    
             python test_server.py ctr_serving_model_kv &
+            sleep 5
             check_cmd "python test_client.py ctr_client_conf/serving_client_conf.prototxt ./ut_data >score"
             tail -n 2 score
             AUC=$(tail -n 2  score | awk 'NR==1')
@@ -217,6 +219,7 @@ function python_run_criteo_ctr_with_cube() {
             sh cube_prepare.sh &
             check_cmd "mkdir work_dir1 && cp cube/conf/cube.conf ./work_dir1/"
             python test_server_gpu.py ctr_serving_model_kv &
+            sleep 5
             check_cmd "python test_client.py ctr_client_conf/serving_client_conf.prototxt ./ut_data >score"
             tail -n 2 score | awk 'NR==1'
             AUC=$(tail -n 2  score | awk 'NR==1')
@@ -235,6 +238,7 @@ function python_run_criteo_ctr_with_cube() {
             exit 1
             ;;
     esac
+    unset SERVING_BIN
     echo "test criteo_ctr_with_cube $TYPE part finished as expected."
     cd .. # pwd: /Serving/python/examples
 }