lexicon.h
1.8 KB
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
// sherpa-onnx/csrc/lexicon.h
//
// Copyright (c) 2022-2023 Xiaomi Corporation
#ifndef SHERPA_ONNX_CSRC_LEXICON_H_
#define SHERPA_ONNX_CSRC_LEXICON_H_
#include <cstdint>
#include <istream>
#include <memory>
#include <string>
#include <unordered_map>
#include <unordered_set>
#include <vector>
#include "sherpa-onnx/csrc/offline-tts-frontend.h"
namespace sherpa_onnx {
class Lexicon : public OfflineTtsFrontend {
public:
Lexicon() = default; // for subclasses
//
// Note: for models from piper, we won't use this class.
Lexicon(const std::string &lexicon, const std::string &tokens,
const std::string &punctuations, const std::string &language,
bool debug = false);
template <typename Manager>
Lexicon(Manager *mgr, const std::string &lexicon, const std::string &tokens,
const std::string &punctuations, const std::string &language,
bool debug = false);
std::vector<TokenIDs> ConvertTextToTokenIds(
const std::string &text, const std::string &voice = "") const override;
private:
std::vector<TokenIDs> ConvertTextToTokenIdsNotChinese(
const std::string &text) const;
std::vector<TokenIDs> ConvertTextToTokenIdsChinese(
const std::string &text) const;
void InitLanguage(const std::string &lang);
void InitTokens(std::istream &is);
void InitLexicon(std::istream &is);
void InitPunctuations(const std::string &punctuations);
private:
enum class Language {
kNotChinese,
kChinese,
kUnknown,
};
private:
std::unordered_map<std::string, std::vector<int32_t>> word2ids_;
std::unordered_set<std::string> punctuations_;
std::unordered_map<std::string, int32_t> token2id_;
Language language_ = Language::kUnknown;
bool debug_ = false;
};
} // namespace sherpa_onnx
#endif // SHERPA_ONNX_CSRC_LEXICON_H_