xref: /aosp_15_r20/external/libtextclassifier/native/utils/grammar/analyzer.h (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 #ifndef LIBTEXTCLASSIFIER_UTILS_GRAMMAR_ANALYZER_H_
18*993b0882SAndroid Build Coastguard Worker #define LIBTEXTCLASSIFIER_UTILS_GRAMMAR_ANALYZER_H_
19*993b0882SAndroid Build Coastguard Worker 
20*993b0882SAndroid Build Coastguard Worker #include "utils/base/arena.h"
21*993b0882SAndroid Build Coastguard Worker #include "utils/base/statusor.h"
22*993b0882SAndroid Build Coastguard Worker #include "utils/grammar/evaluated-derivation.h"
23*993b0882SAndroid Build Coastguard Worker #include "utils/grammar/parsing/parser.h"
24*993b0882SAndroid Build Coastguard Worker #include "utils/grammar/semantics/composer.h"
25*993b0882SAndroid Build Coastguard Worker #include "utils/grammar/text-context.h"
26*993b0882SAndroid Build Coastguard Worker #include "utils/i18n/locale.h"
27*993b0882SAndroid Build Coastguard Worker #include "utils/tokenizer.h"
28*993b0882SAndroid Build Coastguard Worker #include "utils/utf8/unilib.h"
29*993b0882SAndroid Build Coastguard Worker 
30*993b0882SAndroid Build Coastguard Worker namespace libtextclassifier3::grammar {
31*993b0882SAndroid Build Coastguard Worker 
32*993b0882SAndroid Build Coastguard Worker // An analyzer that parses and semantically evaluates an input text with a
33*993b0882SAndroid Build Coastguard Worker // grammar.
34*993b0882SAndroid Build Coastguard Worker class Analyzer {
35*993b0882SAndroid Build Coastguard Worker  public:
36*993b0882SAndroid Build Coastguard Worker   explicit Analyzer(const UniLib* unilib, const RulesSet* rules_set);
37*993b0882SAndroid Build Coastguard Worker   explicit Analyzer(const UniLib* unilib, const RulesSet* rules_set,
38*993b0882SAndroid Build Coastguard Worker                     const Tokenizer* tokenizer);
39*993b0882SAndroid Build Coastguard Worker 
40*993b0882SAndroid Build Coastguard Worker   // Parses and evaluates an input.
41*993b0882SAndroid Build Coastguard Worker   StatusOr<std::vector<EvaluatedDerivation>> Parse(
42*993b0882SAndroid Build Coastguard Worker       const TextContext& input, UnsafeArena* arena,
43*993b0882SAndroid Build Coastguard Worker       bool deduplicate_derivations = true) const;
44*993b0882SAndroid Build Coastguard Worker 
45*993b0882SAndroid Build Coastguard Worker   StatusOr<std::vector<EvaluatedDerivation>> Parse(
46*993b0882SAndroid Build Coastguard Worker       const UnicodeText& text, const std::vector<Locale>& locales,
47*993b0882SAndroid Build Coastguard Worker       UnsafeArena* arena, bool deduplicate_derivations = true) const;
48*993b0882SAndroid Build Coastguard Worker 
49*993b0882SAndroid Build Coastguard Worker   // Pre-processes an input text for parsing.
50*993b0882SAndroid Build Coastguard Worker   TextContext BuildTextContextForInput(
51*993b0882SAndroid Build Coastguard Worker       const UnicodeText& text, const std::vector<Locale>& locales = {}) const;
52*993b0882SAndroid Build Coastguard Worker 
parser()53*993b0882SAndroid Build Coastguard Worker   const Parser& parser() const { return parser_; }
54*993b0882SAndroid Build Coastguard Worker 
55*993b0882SAndroid Build Coastguard Worker  private:
56*993b0882SAndroid Build Coastguard Worker   std::unique_ptr<Tokenizer> owned_tokenizer_;
57*993b0882SAndroid Build Coastguard Worker   const Tokenizer* tokenizer_;
58*993b0882SAndroid Build Coastguard Worker   Parser parser_;
59*993b0882SAndroid Build Coastguard Worker   SemanticComposer semantic_evaluator_;
60*993b0882SAndroid Build Coastguard Worker };
61*993b0882SAndroid Build Coastguard Worker 
62*993b0882SAndroid Build Coastguard Worker }  // namespace libtextclassifier3::grammar
63*993b0882SAndroid Build Coastguard Worker 
64*993b0882SAndroid Build Coastguard Worker #endif  // LIBTEXTCLASSIFIER_UTILS_GRAMMAR_ANALYZER_H_
65