xref: /aosp_15_r20/external/libtextclassifier/native/utils/sentencepiece/test_utils.cc (revision 993b0882672172b81d12fad7a7ac0c3e5c824a12)
1*993b0882SAndroid Build Coastguard Worker /*
2*993b0882SAndroid Build Coastguard Worker  * Copyright (C) 2018 The Android Open Source Project
3*993b0882SAndroid Build Coastguard Worker  *
4*993b0882SAndroid Build Coastguard Worker  * Licensed under the Apache License, Version 2.0 (the "License");
5*993b0882SAndroid Build Coastguard Worker  * you may not use this file except in compliance with the License.
6*993b0882SAndroid Build Coastguard Worker  * You may obtain a copy of the License at
7*993b0882SAndroid Build Coastguard Worker  *
8*993b0882SAndroid Build Coastguard Worker  *      http://www.apache.org/licenses/LICENSE-2.0
9*993b0882SAndroid Build Coastguard Worker  *
10*993b0882SAndroid Build Coastguard Worker  * Unless required by applicable law or agreed to in writing, software
11*993b0882SAndroid Build Coastguard Worker  * distributed under the License is distributed on an "AS IS" BASIS,
12*993b0882SAndroid Build Coastguard Worker  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13*993b0882SAndroid Build Coastguard Worker  * See the License for the specific language governing permissions and
14*993b0882SAndroid Build Coastguard Worker  * limitations under the License.
15*993b0882SAndroid Build Coastguard Worker  */
16*993b0882SAndroid Build Coastguard Worker 
17*993b0882SAndroid Build Coastguard Worker #include "utils/sentencepiece/test_utils.h"
18*993b0882SAndroid Build Coastguard Worker 
19*993b0882SAndroid Build Coastguard Worker #include <memory>
20*993b0882SAndroid Build Coastguard Worker 
21*993b0882SAndroid Build Coastguard Worker #include "utils/base/integral_types.h"
22*993b0882SAndroid Build Coastguard Worker #include "utils/container/double-array-trie.h"
23*993b0882SAndroid Build Coastguard Worker #include "utils/strings/stringpiece.h"
24*993b0882SAndroid Build Coastguard Worker 
25*993b0882SAndroid Build Coastguard Worker namespace libtextclassifier3 {
26*993b0882SAndroid Build Coastguard Worker 
NormalizerFromSpec(StringPiece spec,bool add_dummy_prefix,bool remove_extra_whitespaces,bool escape_whitespaces)27*993b0882SAndroid Build Coastguard Worker SentencePieceNormalizer NormalizerFromSpec(StringPiece spec,
28*993b0882SAndroid Build Coastguard Worker                                            bool add_dummy_prefix,
29*993b0882SAndroid Build Coastguard Worker                                            bool remove_extra_whitespaces,
30*993b0882SAndroid Build Coastguard Worker                                            bool escape_whitespaces) {
31*993b0882SAndroid Build Coastguard Worker   const uint32 trie_blob_size = reinterpret_cast<const uint32*>(spec.data())[0];
32*993b0882SAndroid Build Coastguard Worker   spec.RemovePrefix(sizeof(trie_blob_size));
33*993b0882SAndroid Build Coastguard Worker   const TrieNode* trie_blob = reinterpret_cast<const TrieNode*>(spec.data());
34*993b0882SAndroid Build Coastguard Worker   spec.RemovePrefix(trie_blob_size);
35*993b0882SAndroid Build Coastguard Worker   const int num_nodes = trie_blob_size / sizeof(TrieNode);
36*993b0882SAndroid Build Coastguard Worker   return SentencePieceNormalizer(
37*993b0882SAndroid Build Coastguard Worker       DoubleArrayTrie(trie_blob, num_nodes),
38*993b0882SAndroid Build Coastguard Worker       /*charsmap_normalized=*/StringPiece(spec.data(), spec.size()),
39*993b0882SAndroid Build Coastguard Worker       add_dummy_prefix, remove_extra_whitespaces, escape_whitespaces);
40*993b0882SAndroid Build Coastguard Worker }
41*993b0882SAndroid Build Coastguard Worker 
42*993b0882SAndroid Build Coastguard Worker }  // namespace libtextclassifier3
43