Python* API Reference for Intel® Data Analytics Acceleration Library 2018 Update 2

datastructures_aos.py

1 # file: datastructures_aos.py
2 #===============================================================================
3 # Copyright 2014-2018 Intel Corporation
4 # All Rights Reserved.
5 #
6 # If this software was obtained under the Intel Simplified Software License,
7 # the following terms apply:
8 #
9 # The source code, information and material ("Material") contained herein is
10 # owned by Intel Corporation or its suppliers or licensors, and title to such
11 # Material remains with Intel Corporation or its suppliers or licensors. The
12 # Material contains proprietary information of Intel or its suppliers and
13 # licensors. The Material is protected by worldwide copyright laws and treaty
14 # provisions. No part of the Material may be used, copied, reproduced,
15 # modified, published, uploaded, posted, transmitted, distributed or disclosed
16 # in any way without Intel's prior express written permission. No license under
17 # any patent, copyright or other intellectual property rights in the Material
18 # is granted to or conferred upon you, either expressly, by implication,
19 # inducement, estoppel or otherwise. Any license under such intellectual
20 # property rights must be express and approved by Intel in writing.
21 #
22 # Unless otherwise agreed by Intel in writing, you may not remove or alter this
23 # notice or any other notice embedded in Materials by Intel or Intel's
24 # suppliers or licensors in any way.
25 #
26 #
27 # If this software was obtained under the Apache License, Version 2.0 (the
28 # "License"), the following terms apply:
29 #
30 # You may not use this file except in compliance with the License. You may
31 # obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0
32 #
33 #
34 # Unless required by applicable law or agreed to in writing, software
35 # distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
36 # WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
37 #
38 # See the License for the specific language governing permissions and
39 # limitations under the License.
40 #===============================================================================
41 
42 
43 
44 
45 import os
46 import sys
47 
48 import numpy as np
49 
50 from daal.data_management import data_feature_utils, AOSNumericTable, BlockDescriptor, readOnly, readWrite
51 
52 utils_folder = os.path.realpath(os.path.abspath(os.path.dirname(os.path.dirname(__file__))))
53 if utils_folder not in sys.path:
54  sys.path.insert(0, utils_folder)
55 from utils import printArray
56 
57 
58 if __name__ == "__main__":
59 
60  print("Array of structures (AOS) numeric table example\n")
61 
62  points = np.array([(0.5, -1.3, 1, 100.1),
63  (2.5, -3.3, 2, 200.2),
64  (4.5, -5.3, 2, 350.3),
65  (6.5, -7.3, 0, 470.4),
66  (8.5, -9.3, 1, 270.5)],
67  dtype=[('x','f4'), ('y','f4'), ('categ','i4'), ('value','f8')])
68 
69  nObservations = len(points)
70  nFeatures = len(points[0])
71 
72  # Construct AOS numericTable for a data array with nFeatures fields and nObservations elements
73  # Dictionary will be initialized with type information from ndarray
74  dataTable = AOSNumericTable(points)
75 
76  # Get the dictionary and update it with additional information about data
77  dict = dataTable.getDictionary()
78 
79  # Add a feature type to the dictionary
80  dict[0].featureType = data_feature_utils.DAAL_CONTINUOUS
81  dict[1].featureType = data_feature_utils.DAAL_CONTINUOUS
82  dict[2].featureType = data_feature_utils.DAAL_CATEGORICAL
83  dict[3].featureType = data_feature_utils.DAAL_CONTINUOUS
84 
85  # Set the number of categories for a categorical feature
86  dict[2].categoryNumber = 3
87 
88  # Read a block of rows
89  firstReadRow = 0
90  doubleBlock = BlockDescriptor()
91  dataTable.getBlockOfRows(firstReadRow, nObservations, readWrite, doubleBlock)
92  printArray(
93  doubleBlock.getArray(), nFeatures, doubleBlock.getNumberOfRows(),
94  doubleBlock.getNumberOfColumns(),"Print AOS data structures as double:"
95  )
96  dataTable.releaseBlockOfRows(doubleBlock)
97 
98  # Read a feature (column)
99  readFeatureIdx = 2
100 
101  intBlock = BlockDescriptor(ntype=np.intc)
102  dataTable.getBlockOfColumnValues(readFeatureIdx, firstReadRow, nObservations, readOnly, intBlock)
103  printArray(
104  intBlock.getArray(), 1, intBlock.getNumberOfRows(), intBlock.getNumberOfColumns(),
105  "Print the third feature of AOS:", flt64=False
106  )
107  dataTable.releaseBlockOfColumnValues(intBlock)

For more complete information about compiler optimizations, see our Optimization Notice.