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
|
/*
* Copyright (c) 2020 Samsung Electronics Co., Ltd. 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.
*/
#include "tfinfo-v2/TensorInfoLoader.h"
#include "tfinfo-v2/TensorSignature.h"
#include <oops/UserExn.h>
#include <stdex/Memory.h>
#include <tfinfo-v2.pb.h>
#include <google/protobuf/io/zero_copy_stream_impl.h>
#include <google/protobuf/text_format.h>
#include <fstream>
#include <fcntl.h>
namespace
{
// for testing purpose
bool load_text(std::istream *stream, tfinfo_v2_proto::InfoDef &info_def)
{
google::protobuf::io::IstreamInputStream iis(stream);
return google::protobuf::TextFormat::Parse(&iis, &info_def);
}
bool is_num(const std::string &num)
{
for (int i = 0; i < num.length(); i++)
if (!isdigit(num[i]))
return false;
return true;
}
void validate_tensor_name(const std::string &tensor_name, const char *path)
{
// Note that Tensorflow tensor name format is
// operation name ":" index, e.g., "in/placeholder:0"
int pos = tensor_name.find(":");
if (pos == std::string::npos)
throw oops::UserExn("Missing index separator, ':'", "name", tensor_name, "file", path);
if (tensor_name.length() == pos + 1) // ':' is the last char
throw oops::UserExn("Missing tensor index after ':'", "name", tensor_name, "file", path);
// 1. Validating operation name.
// for naming format, refer to https://www.tensorflow.org/api_docs/python/tf/Operation#__init__
// First char is in the form of "[A-Za-z0-9.]"
// and the rest chars are in the form of "[A-Za-z0-9_.\\-/]*"
std::string op_name = tensor_name.substr(0, pos);
// first character
if (!(isalnum(op_name[0]) || op_name[0] == '.'))
throw oops::UserExn("Wrong tensor name format", "name", tensor_name, "file", path);
// and the rest chars
for (int i = 1; i < op_name.length(); i++)
if (!(isalnum(op_name[i]) || std::string("_.\\-/").find(op_name[i]) != std::string::npos))
throw oops::UserExn("Wrong tensor name format", "name", tensor_name, "file", path);
// 2. validating index after ":"
std::string index = tensor_name.substr(pos + 1, op_name.length() - pos - 1);
if (!is_num(index))
throw oops::UserExn("Wrong tensor name format", "name", tensor_name, "file", path);
}
void check_duplicate(tfinfo::v2::TensorSignatures &tensors, const char *path)
{
std::map<std::string, bool> tool;
for (auto &tensor : tensors)
{
if (tool.find(tensor->name()) != tool.end())
throw oops::UserExn("Duplicate tensor definition", "name", tensor->name(), "file", path);
else
tool[tensor->name()] = true;
}
}
void convert(tfinfo_v2_proto::InfoDef &info_def, tfinfo::v2::TensorSignatures &tensors,
const char *path)
{
// processing input. Note that there could be no input.
if (auto input_size = info_def.input_size())
{
for (int i = 0; i < input_size; i++)
{
auto input_def = info_def.input().Get(i);
auto name = input_def.name();
validate_tensor_name(name, path);
auto tensor = stdex::make_unique<tfinfo::v2::TensorSignature>(
tfinfo::v2::TensorSignature::Kind::Input, name);
// when there is dim attribute for unknown shape
if (input_def.dim_size() > 0)
{
for (int d = 0; d < input_def.dim().size(); d++)
{
auto dim = input_def.dim(d);
tensor->shapeHint().add(dim.axis(), dim.size());
}
}
tensors.emplace_back(std::move(tensor));
}
}
// processing output
auto output_size = info_def.output_size();
if (output_size == 0)
throw oops::UserExn("Missing output node. At least 1 output node must exist", "file", path);
if (auto output_node_size = info_def.output_size())
{
for (int i = 0; i < output_node_size; i++)
{
auto name = info_def.output().Get(i).name();
validate_tensor_name(name, path);
auto tensor = stdex::make_unique<tfinfo::v2::TensorSignature>(
tfinfo::v2::TensorSignature::Kind::Output, name);
tensors.emplace_back(std::move(tensor));
}
}
check_duplicate(tensors, path);
}
} // namespace
namespace tfinfo
{
inline namespace v2
{
TensorSignatures load(const char *path)
{
std::ifstream stream(path, std::ios::in | std::ios::binary);
return load(&stream, path);
}
TensorSignatures load(std::istream *stream, const char *path_for_error_msg)
{
tfinfo_v2_proto::InfoDef info_def;
if (!load_text(stream, info_def))
{
throw oops::UserExn("Cannot parse the info file", "path", path_for_error_msg);
}
TensorSignatures tensors;
convert(info_def, tensors, path_for_error_msg);
return tensors;
}
} // namespace v2
} // namespace tfinfo
|
