move imdb reader to app

python/examples/imdb/test_client.py

@@ -13,7 +13,7 @@
 # limitations under the License.
 # pylint: disable=doc-string-missing
 from paddle_serving_client import Client
-from imdb_reader import IMDBDataset
+from paddle_serving_app import IMDBReader
 import sys
 
 client = Client()
@@ -23,7 +23,7 @@ client.connect(["127.0.0.1:9292"])
 # you can define any english sentence or dataset here
 # This example reuses imdb reader in training, you
 # can define your own data preprocessing easily.
-imdb_dataset = IMDBDataset()
+imdb_dataset = IMDBReader()
 imdb_dataset.load_resource(sys.argv[2])
 
 for line in sys.stdin:

python/examples/imdb/text_classify_service.py

@@ -14,7 +14,7 @@
 # pylint: disable=doc-string-missing
 
 from paddle_serving_server.web_service import WebService
-from imdb_reader import IMDBDataset
+from paddle_serving_app import IMDBReader
 import sys
 
 
@@ -22,7 +22,7 @@ class IMDBService(WebService):
     def prepare_dict(self, args={}):
         if len(args) == 0:
             exit(-1)
-        self.dataset = IMDBDataset()
+        self.dataset = IMDBReader()
         self.dataset.load_resource(args["dict_file_path"])
 
     def preprocess(self, feed={}, fetch=[]):

python/paddle_serving_app/__init__.py

@@ -15,5 +15,6 @@ from .reader.chinese_bert_reader import ChineseBertReader
 from .reader.image_reader import ImageReader, File2Image, URL2Image, Sequential, Normalize, CenterCrop, Resize
 from .reader.lac_reader import LACReader
 from .reader.senta_reader import SentaReader
+from .reader.imdb_reader import IMDBReader
 from .models import ServingModels
 from .local_predict import Debugger

python/paddle_serving_app/reader/imdb_reader.py

+#   Copyright (c) 2018 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.
+# pylint: disable=doc-string-missing
+
+import sys
+import os
+import paddle
+import re
+import paddle.fluid.incubate.data_generator as dg
+
+py_version = sys.version_info[0]
+
+
+class IMDBReader(dg.MultiSlotDataGenerator):
+    def load_resource(self, dictfile):
+        self._vocab = {}
+        wid = 0
+        if py_version == 2:
+            with open(dictfile) as f:
+                for line in f:
+                    self._vocab[line.strip()] = wid
+                    wid += 1
+        else:
+            with open(dictfile, encoding="utf-8") as f:
+                for line in f:
+                    self._vocab[line.strip()] = wid
+                    wid += 1
+        self._unk_id = len(self._vocab)
+        self._pattern = re.compile(r'(;|,|\.|\?|!|\s|\(|\))')
+        self.return_value = ("words", [1, 2, 3, 4, 5, 6]), ("label", [0])
+
+    def get_words_only(self, line):
+        sent = line.lower().replace("<br />", " ").strip()
+        words = [x for x in self._pattern.split(sent) if x and x != " "]
+        feas = [
+            self._vocab[x] if x in self._vocab else self._unk_id for x in words
+        ]
+        return feas
+
+    def get_words_and_label(self, line):
+        send = '|'.join(line.split('|')[:-1]).lower().replace("<br />",
+                                                              " ").strip()
+        label = [int(line.split('|')[-1])]
+
+        words = [x for x in self._pattern.split(send) if x and x != " "]
+        feas = [
+            self._vocab[x] if x in self._vocab else self._unk_id for x in words
+        ]
+        return feas, label
+
+    def infer_reader(self, infer_filelist, batch, buf_size):
+        def local_iter():
+            for fname in infer_filelist:
+                with open(fname, "r") as fin:
+                    for line in fin:
+                        feas, label = self.get_words_and_label(line)
+                        yield feas, label
+
+        import paddle
+        batch_iter = paddle.batch(
+            paddle.reader.shuffle(
+                local_iter, buf_size=buf_size),
+            batch_size=batch)
+        return batch_iter
+
+    def generate_sample(self, line):
+        def memory_iter():
+            for i in range(1000):
+                yield self.return_value
+
+        def data_iter():
+            feas, label = self.get_words_and_label(line)
+            yield ("words", feas), ("label", label)
+
+        return data_iter
+
+
+if __name__ == "__main__":
+    imdb = IMDBReader()
+    imdb.load_resource("imdb.vocab")
+    imdb.run_from_stdin()