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/zlib/zlib_regex.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/logging.h"
22*993b0882SAndroid Build Coastguard Worker #include "utils/flatbuffers/flatbuffers.h"
23*993b0882SAndroid Build Coastguard Worker #include "utils/utf8/unicodetext.h"
24*993b0882SAndroid Build Coastguard Worker
25*993b0882SAndroid Build Coastguard Worker namespace libtextclassifier3 {
26*993b0882SAndroid Build Coastguard Worker
UncompressMakeRegexPattern(const UniLib & unilib,const flatbuffers::String * uncompressed_pattern,const CompressedBuffer * compressed_pattern,bool lazy_compile_regex,ZlibDecompressor * decompressor,std::string * result_pattern_text)27*993b0882SAndroid Build Coastguard Worker std::unique_ptr<UniLib::RegexPattern> UncompressMakeRegexPattern(
28*993b0882SAndroid Build Coastguard Worker const UniLib& unilib, const flatbuffers::String* uncompressed_pattern,
29*993b0882SAndroid Build Coastguard Worker const CompressedBuffer* compressed_pattern, bool lazy_compile_regex,
30*993b0882SAndroid Build Coastguard Worker ZlibDecompressor* decompressor, std::string* result_pattern_text) {
31*993b0882SAndroid Build Coastguard Worker UnicodeText unicode_regex_pattern;
32*993b0882SAndroid Build Coastguard Worker std::string decompressed_pattern;
33*993b0882SAndroid Build Coastguard Worker if (compressed_pattern != nullptr &&
34*993b0882SAndroid Build Coastguard Worker compressed_pattern->buffer() != nullptr) {
35*993b0882SAndroid Build Coastguard Worker if (decompressor == nullptr ||
36*993b0882SAndroid Build Coastguard Worker !decompressor->MaybeDecompress(compressed_pattern,
37*993b0882SAndroid Build Coastguard Worker &decompressed_pattern)) {
38*993b0882SAndroid Build Coastguard Worker TC3_LOG(ERROR) << "Cannot decompress pattern.";
39*993b0882SAndroid Build Coastguard Worker return nullptr;
40*993b0882SAndroid Build Coastguard Worker }
41*993b0882SAndroid Build Coastguard Worker unicode_regex_pattern =
42*993b0882SAndroid Build Coastguard Worker UTF8ToUnicodeText(decompressed_pattern.data(),
43*993b0882SAndroid Build Coastguard Worker decompressed_pattern.size(), /*do_copy=*/false);
44*993b0882SAndroid Build Coastguard Worker } else {
45*993b0882SAndroid Build Coastguard Worker if (uncompressed_pattern == nullptr) {
46*993b0882SAndroid Build Coastguard Worker TC3_LOG(ERROR) << "Cannot load uncompressed pattern.";
47*993b0882SAndroid Build Coastguard Worker return nullptr;
48*993b0882SAndroid Build Coastguard Worker }
49*993b0882SAndroid Build Coastguard Worker unicode_regex_pattern =
50*993b0882SAndroid Build Coastguard Worker UTF8ToUnicodeText(uncompressed_pattern->c_str(),
51*993b0882SAndroid Build Coastguard Worker uncompressed_pattern->size(), /*do_copy=*/false);
52*993b0882SAndroid Build Coastguard Worker }
53*993b0882SAndroid Build Coastguard Worker
54*993b0882SAndroid Build Coastguard Worker if (result_pattern_text != nullptr) {
55*993b0882SAndroid Build Coastguard Worker *result_pattern_text = unicode_regex_pattern.ToUTF8String();
56*993b0882SAndroid Build Coastguard Worker }
57*993b0882SAndroid Build Coastguard Worker
58*993b0882SAndroid Build Coastguard Worker std::unique_ptr<UniLib::RegexPattern> regex_pattern;
59*993b0882SAndroid Build Coastguard Worker if (lazy_compile_regex) {
60*993b0882SAndroid Build Coastguard Worker regex_pattern = unilib.CreateLazyRegexPattern(unicode_regex_pattern);
61*993b0882SAndroid Build Coastguard Worker } else {
62*993b0882SAndroid Build Coastguard Worker regex_pattern = unilib.CreateRegexPattern(unicode_regex_pattern);
63*993b0882SAndroid Build Coastguard Worker }
64*993b0882SAndroid Build Coastguard Worker
65*993b0882SAndroid Build Coastguard Worker if (!regex_pattern) {
66*993b0882SAndroid Build Coastguard Worker TC3_LOG(ERROR) << "Could not create pattern: "
67*993b0882SAndroid Build Coastguard Worker << unicode_regex_pattern.ToUTF8String();
68*993b0882SAndroid Build Coastguard Worker }
69*993b0882SAndroid Build Coastguard Worker return regex_pattern;
70*993b0882SAndroid Build Coastguard Worker }
71*993b0882SAndroid Build Coastguard Worker
72*993b0882SAndroid Build Coastguard Worker } // namespace libtextclassifier3
73