Merge pull request #241 from MRXLT/general-server-gpu

add web service for gpu server

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 = "0.1.1"
-serving_server_version = "0.1.0"
-module_proto_version = "0.1.0"
+serving_client_version = "0.1.2"
+serving_server_version = "0.1.2"
+module_proto_version = "0.1.2"

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 = "0.1.0"
-serving_server_version = "0.1.0"
-module_proto_version = "0.1.0"
+serving_client_version = "0.1.2"
+serving_server_version = "0.1.2"
+module_proto_version = "0.1.2"

python/paddle_serving_server_gpu/serve.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# 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.
+"""
+Usage:
+    Host a trained paddle model with one line command
+    Example:
+        python -m paddle_serving_server.serve --model ./serving_server_model --port 9292
+"""
+import argparse
+
+
+def parse_args():
+    parser = argparse.ArgumentParser("serve")
+    parser.add_argument(
+        "--thread", type=int, default=10, help="Concurrency of server")
+    parser.add_argument(
+        "--model", type=str, default="", help="Model for serving")
+    parser.add_argument(
+        "--port", type=int, default=9292, help="Port the server")
+    parser.add_argument(
+        "--workdir",
+        type=str,
+        default="workdir",
+        help="Working dir of current service")
+    parser.add_argument(
+        "--device", type=str, default="gpu", help="Type of device")
+    parser.add_argument("--gpuid", type=int, default=0, help="Index of GPU")
+    return parser.parse_args()
+
+
+def start_standard_model():
+    args = parse_args()
+    thread_num = args.thread
+    model = args.model
+    port = args.port
+    workdir = args.workdir
+    device = args.device
+    gpuid = args.gpuid
+
+    if model == "":
+        print("You must specify your serving model")
+        exit(-1)
+
+    import paddle_serving_server_gpu as serving
+    op_maker = serving.OpMaker()
+    read_op = op_maker.create('general_reader')
+    general_infer_op = op_maker.create('general_infer')
+    general_response_op = op_maker.create('general_response')
+
+    op_seq_maker = serving.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 = serving.Server()
+    server.set_op_sequence(op_seq_maker.get_op_sequence())
+    server.set_num_threads(thread_num)
+
+    server.load_model_config(model)
+    server.prepare_server(workdir=workdir, port=port, device=device)
+    server.set_gpuid(gpuid)
+    server.run_server()
+
+
+if __name__ == "__main__":
+    start_standard_model()

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 = "0.1.0"
-serving_server_version = "0.1.0"
-module_proto_version = "0.1.0"
+serving_client_version = "0.1.2"
+serving_server_version = "0.1.2"
+module_proto_version = "0.1.2"

python/paddle_serving_server_gpu/web_serve.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# 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.
+"""
+Usage:
+    Host a trained paddle model with one line command
+    Example:
+        python -m paddle_serving_server.web_serve --model ./serving_server_model --port 9292
+"""
+import argparse
+from multiprocessing import Pool, Process
+from .web_service import WebService
+
+
+def parse_args():
+    parser = argparse.ArgumentParser("web_serve")
+    parser.add_argument(
+        "--thread", type=int, default=10, help="Concurrency of server")
+    parser.add_argument(
+        "--model", type=str, default="", help="Model for serving")
+    parser.add_argument(
+        "--port", type=int, default=9292, help="Port the server")
+    parser.add_argument(
+        "--workdir",
+        type=str,
+        default="workdir",
+        help="Working dir of current service")
+    parser.add_argument(
+        "--device", type=str, default="cpu", help="Type of device")
+    parser.add_argument(
+        "--name", type=str, default="default", help="Default service name")
+    return parser.parse_args()
+
+
+if __name__ == "__main__":
+    args = parse_args()
+    service = WebService(name=args.name)
+    service.load_model_config(args.model)
+    service.prepare_server(
+        workdir=args.workdir, port=args.port, device=args.device)
+    service.run_server()

python/paddle_serving_server_gpu/web_service.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# 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.
+#!flask/bin/python
+from flask import Flask, request, abort
+from multiprocessing import Pool, Process
+from paddle_serving_server_gpu import OpMaker, OpSeqMaker, Server
+from paddle_serving_client import Client
+
+
+class WebService(object):
+    def __init__(self, name="default_service"):
+        self.name = name
+
+    def load_model_config(self, model_config):
+        self.model_config = model_config
+
+    def _launch_rpc_service(self):
+        op_maker = OpMaker()
+        read_op = op_maker.create('general_reader')
+        general_infer_op = op_maker.create('general_infer')
+        general_response_op = op_maker.create('general_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.set_num_threads(16)
+        server.set_gpuid = self.gpuid
+        server.load_model_config(self.model_config)
+        server.prepare_server(
+            workdir=self.workdir, port=self.port + 1, device=self.device)
+        server.run_server()
+
+    def prepare_server(self, workdir="", port=9393, device="gpu", gpuid=0):
+        self.workdir = workdir
+        self.port = port
+        self.device = device
+        self.gpuid = gpuid
+
+    def _launch_web_service(self):
+        app_instance = Flask(__name__)
+        client_service = Client()
+        client_service.load_client_config(
+            "{}/serving_server_conf.prototxt".format(self.model_config))
+        client_service.connect(["127.0.0.1:{}".format(self.port + 1)])
+        service_name = "/" + self.name + "/prediction"
+
+        @app_instance.route(service_name, methods=['POST'])
+        def get_prediction():
+            if not request.json:
+                abort(400)
+            if "fetch" not in request.json:
+                abort(400)
+            feed, fetch = self.preprocess(request.json, request.json["fetch"])
+            fetch_map = client_service.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="127.0.0.1",
+                         port=self.port,
+                         threaded=False,
+                         processes=1)
+
+    def run_server(self):
+        import socket
+        localIP = socket.gethostbyname(socket.gethostname())
+        print("web service address:")
+        print("http://{}:{}/{}/prediction".format(localIP, self.port,
+                                                  self.name))
+        p_rpc = Process(target=self._launch_rpc_service)
+        p_web = Process(target=self._launch_web_service)
+        p_rpc.start()
+        p_web.start()
+        p_web.join()
+        p_rpc.join()
+
+    def preprocess(self, feed={}, fetch=[]):
+        return feed, fetch
+
+    def postprocess(self, feed={}, fetch=[], fetch_map={}):
+        return fetch_map