add odps demo

Author: kuizhiqing

tools/dataset/README.MD

@@ -285,5 +285,13 @@ export KAFKA_GID=xxx
 export KAFKA_TOPICS=wide-and-deep-data
 
 python kafka_reader.py
+
+# test odps reader 
+#
+# create config.py 
+#
+
+python odps_reader.py
+
 ```
 

tools/dataset/odps_reader.py

@@ -0,0 +1,75 @@
+# Copyright (c) 2021 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.
+
+from odps import ODPS
+from odps.models import Schema, Column, Partition
+
+import paddle
+import paddle.distributed.fleet as fleet
+import os
+import sys
+
+cont_min_ = [0, -3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
+cont_max_ = [20, 600, 100, 50, 64000, 500, 100, 50, 500, 10, 10, 10, 50]
+cont_diff_ = [20, 603, 100, 50, 64000, 500, 100, 50, 500, 10, 10, 10, 50]
+hash_dim_ = 1000001
+continuous_range_ = range(1, 14)
+categorical_range_ = range(14, 40)
+
+
+class WideDeepDatasetReader(fleet.MultiSlotDataGenerator):
+    def line_process(self, record):
+        label = [record.label]
+        dense_feature = []
+        sparse_feature = []
+        for idx in continuous_range_:
+            dense_feature.append((float(record[idx]) - cont_min_[idx - 1]) /
+                                 cont_diff_[idx - 1])
+        for idx in categorical_range_:
+            sparse_feature.append([hash(str(idx) + record[idx]) % hash_dim_])
+        return [dense_feature] + sparse_feature + [label]
+
+    def generate_sample(self, line):
+        def wd_reader():
+            for record in table_reader:
+                input_data = self.line_process(record)
+                feature_name = ["dense_input"]
+                for idx in categorical_range_:
+                    feature_name.append("C" + str(idx - 13))
+                feature_name.append("label")
+                yield zip(feature_name, input_data)
+
+        return wd_reader
+
+
+if __name__ == "__main__":
+    my_data_generator = WideDeepDatasetReader()
+    #my_data_generator.set_batch(16)
+
+    from config import *
+    # config should include flowing configuration
+    # access_id
+    # secret_key
+    # project
+    # endpoint
+
+    o = ODPS(access_id, secret_key, project, endpoint=endpoint)
+
+    table_name = 'wide_and_deep'
+
+    table = o.get_table(table_name)  #.to_df()
+
+    table_reader = table.open_reader()
+
+    my_data_generator.run_from_memory()

tools/dataset/utils/odps_writer.py

@@ -0,0 +1,98 @@
+# Copyright (c) 2021 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.
+
+from odps import ODPS
+from odps.models import Schema, Column, Partition
+
+from config import *
+# config should include flowing configuration
+# access_id
+# secret_key
+# project
+# endpoint
+'''
+For more information, ref to
+https://pyodps.readthedocs.io/
+'''
+
+o = ODPS(access_id, secret_key, project, endpoint=endpoint)
+
+label_col = [Column(name='label', type='bigint')]
+dense_col = [
+    Column(
+        name='dense' + str(i), type='double') for i in range(1, 14)
+]
+sparse_col = [Column(name='C' + str(i), type='string') for i in range(14, 40)]
+
+columns = label_col + dense_col + sparse_col
+
+schema = Schema(
+    columns=columns)  # schema = Schema(columns=columns, partitions=partitions)
+
+table_name = 'wide_and_deep'
+
+print(schema.columns)
+
+
+def create_table():
+    table = o.create_table(table_name, schema, if_not_exists=True)
+
+
+#create_table()
+
+table = o.get_table(table_name)  #.to_df()
+print(table.to_df())
+
+
+def write_data():
+    records = []
+
+    # prepare data
+    input_file = './part-0'
+    with open(input_file, 'r') as f:
+        for line in f:
+            example = []
+
+            features = line.rstrip('\n').split('\t')
+            label = int(features[0])
+            example.append(label)
+
+            for idx in range(1, 14):
+                if features[idx] == "":
+                    example.append(0.0)
+                else:
+                    example.append(float(features[idx]))
+            for idx in range(14, 40):
+                example.append(features[idx].encode('utf8'))
+
+            records.append(example)
+
+    with table.open_writer() as writer:
+        writer.write(records)
+
+
+#write_data()
+
+
+def read_data():
+    with table.open_reader() as reader:
+        print("data count in table:", reader.count)
+        for r in reader:
+            print(r.label)
+            #print([i for i in r[1:14]])
+            print([i for i in r])
+            break
+
+
+read_data()