1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
|
#
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements. See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You 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.
#
import sys
import uuid
from functools import wraps
if sys.version > '3':
basestring = str
from pyspark import SparkContext, since
from pyspark.mllib.common import inherit_doc
def _jvm():
"""
Returns the JVM view associated with SparkContext. Must be called
after SparkContext is initialized.
"""
jvm = SparkContext._jvm
if jvm:
return jvm
else:
raise AttributeError("Cannot load _jvm from SparkContext. Is SparkContext initialized?")
def keyword_only(func):
"""
A decorator that forces keyword arguments in the wrapped method
and saves actual input keyword arguments in `_input_kwargs`.
"""
@wraps(func)
def wrapper(*args, **kwargs):
if len(args) > 1:
raise TypeError("Method %s forces keyword arguments." % func.__name__)
wrapper._input_kwargs = kwargs
return func(*args, **kwargs)
return wrapper
class Identifiable(object):
"""
Object with a unique ID.
"""
def __init__(self):
#: A unique id for the object.
self.uid = self._randomUID()
def __repr__(self):
return self.uid
@classmethod
def _randomUID(cls):
"""
Generate a unique id for the object. The default implementation
concatenates the class name, "_", and 12 random hex chars.
"""
return cls.__name__ + "_" + uuid.uuid4().hex[12:]
@inherit_doc
class MLWriter(object):
"""
.. note:: Experimental
Utility class that can save ML instances.
.. versionadded:: 2.0.0
"""
def save(self, path):
"""Save the ML instance to the input path."""
raise NotImplementedError("MLWriter is not yet implemented for type: %s" % type(self))
def overwrite(self):
"""Overwrites if the output path already exists."""
raise NotImplementedError("MLWriter is not yet implemented for type: %s" % type(self))
def context(self, sqlContext):
"""Sets the SQL context to use for saving."""
raise NotImplementedError("MLWriter is not yet implemented for type: %s" % type(self))
@inherit_doc
class JavaMLWriter(MLWriter):
"""
(Private) Specialization of :py:class:`MLWriter` for :py:class:`JavaWrapper` types
"""
def __init__(self, instance):
super(JavaMLWriter, self).__init__()
_java_obj = instance._to_java()
self._jwrite = _java_obj.write()
def save(self, path):
"""Save the ML instance to the input path."""
if not isinstance(path, basestring):
raise TypeError("path should be a basestring, got type %s" % type(path))
self._jwrite.save(path)
def overwrite(self):
"""Overwrites if the output path already exists."""
self._jwrite.overwrite()
return self
def context(self, sqlContext):
"""Sets the SQL context to use for saving."""
self._jwrite.context(sqlContext._ssql_ctx)
return self
@inherit_doc
class MLWritable(object):
"""
.. note:: Experimental
Mixin for ML instances that provide :py:class:`MLWriter`.
.. versionadded:: 2.0.0
"""
def write(self):
"""Returns an JavaMLWriter instance for this ML instance."""
raise NotImplementedError("MLWritable is not yet implemented for type: %r" % type(self))
def save(self, path):
"""Save this ML instance to the given path, a shortcut of `write().save(path)`."""
self.write().save(path)
@inherit_doc
class JavaMLWritable(MLWritable):
"""
(Private) Mixin for ML instances that provide :py:class:`JavaMLWriter`.
"""
def write(self):
"""Returns an JavaMLWriter instance for this ML instance."""
return JavaMLWriter(self)
@inherit_doc
class MLReader(object):
"""
.. note:: Experimental
Utility class that can load ML instances.
.. versionadded:: 2.0.0
"""
def load(self, path):
"""Load the ML instance from the input path."""
raise NotImplementedError("MLReader is not yet implemented for type: %s" % type(self))
def context(self, sqlContext):
"""Sets the SQL context to use for loading."""
raise NotImplementedError("MLReader is not yet implemented for type: %s" % type(self))
@inherit_doc
class JavaMLReader(MLReader):
"""
(Private) Specialization of :py:class:`MLReader` for :py:class:`JavaWrapper` types
"""
def __init__(self, clazz):
self._clazz = clazz
self._jread = self._load_java_obj(clazz).read()
def load(self, path):
"""Load the ML instance from the input path."""
if not isinstance(path, basestring):
raise TypeError("path should be a basestring, got type %s" % type(path))
java_obj = self._jread.load(path)
if not hasattr(self._clazz, "_from_java"):
raise NotImplementedError("This Java ML type cannot be loaded into Python currently: %r"
% self._clazz)
return self._clazz._from_java(java_obj)
def context(self, sqlContext):
"""Sets the SQL context to use for loading."""
self._jread.context(sqlContext._ssql_ctx)
return self
@classmethod
def _java_loader_class(cls, clazz):
"""
Returns the full class name of the Java ML instance. The default
implementation replaces "pyspark" by "org.apache.spark" in
the Python full class name.
"""
java_package = clazz.__module__.replace("pyspark", "org.apache.spark")
if clazz.__name__ in ("Pipeline", "PipelineModel"):
# Remove the last package name "pipeline" for Pipeline and PipelineModel.
java_package = ".".join(java_package.split(".")[0:-1])
return java_package + "." + clazz.__name__
@classmethod
def _load_java_obj(cls, clazz):
"""Load the peer Java object of the ML instance."""
java_class = cls._java_loader_class(clazz)
java_obj = _jvm()
for name in java_class.split("."):
java_obj = getattr(java_obj, name)
return java_obj
@inherit_doc
class MLReadable(object):
"""
.. note:: Experimental
Mixin for instances that provide :py:class:`MLReader`.
.. versionadded:: 2.0.0
"""
@classmethod
def read(cls):
"""Returns an JavaMLReader instance for this class."""
raise NotImplementedError("MLReadable.read() not implemented for type: %r" % cls)
@classmethod
def load(cls, path):
"""Reads an ML instance from the input path, a shortcut of `read().load(path)`."""
return cls.read().load(path)
@inherit_doc
class JavaMLReadable(MLReadable):
"""
(Private) Mixin for instances that provide JavaMLReader.
"""
@classmethod
def read(cls):
"""Returns an JavaMLReader instance for this class."""
return JavaMLReader(cls)
|
