microsoft/onnxruntime-extensions
Publicmirrored fromhttps://github.com/microsoft/onnxruntime-extensionsAvailable
operators/string_utils.h
70lines · modecode
| 1 | // Copyright (c) Microsoft Corporation. All rights reserved. |
| 2 | // Licensed under the MIT License. |
| 3 | #pragma once |
| 4 | #include <iostream> |
| 5 | #include <sstream> |
| 6 | #include <vector> |
| 7 | #include "ocos.h" |
| 8 | |
| 9 | template <typename T> |
| 10 | inline void MakeStringInternal(std::ostringstream& ss, const T& t) noexcept { |
| 11 | ss << t; |
| 12 | } |
| 13 | |
| 14 | template <> |
| 15 | inline void MakeStringInternal(std::ostringstream& ss, const std::vector<int64_t>& t) noexcept { |
| 16 | ss << "["; |
| 17 | for (int i = 0; i < t.size(); i++) { |
| 18 | if (i != 0) { |
| 19 | ss << ", "; |
| 20 | } |
| 21 | ss << t[i]; |
| 22 | } |
| 23 | ss << "]"; |
| 24 | } |
| 25 | |
| 26 | template <> |
| 27 | inline void MakeStringInternal(std::ostringstream& ss, const OrtTensorDimensions& t) noexcept { |
| 28 | MakeStringInternal(ss, static_cast<const std::vector<int64_t>&>(t)); |
| 29 | } |
| 30 | |
| 31 | template <> |
| 32 | inline void MakeStringInternal(std::ostringstream& ss, const std::vector<std::string>& t) noexcept { |
| 33 | ss << "["; |
| 34 | for (int i = 0; i < t.size(); i++) { |
| 35 | if (i != 0) { |
| 36 | ss << ", "; |
| 37 | } |
| 38 | ss << t[i]; |
| 39 | } |
| 40 | ss << "]"; |
| 41 | } |
| 42 | |
| 43 | template <typename T, typename... Args> |
| 44 | void MakeStringInternal(std::ostringstream& ss, const T& t, const Args&... args) noexcept { |
| 45 | MakeStringInternal(ss, t); |
| 46 | MakeStringInternal(ss, args...); |
| 47 | } |
| 48 | |
| 49 | template <typename... Args> |
| 50 | std::string MakeString(const Args&... args) { |
| 51 | std::ostringstream ss; |
| 52 | MakeStringInternal(ss, args...); |
| 53 | return std::string(ss.str()); |
| 54 | } |
| 55 | |
| 56 | std::vector<std::string_view> SplitString(const std::string_view& str, const std::string_view& seps, bool remove_empty_entries = false); |
| 57 | |
| 58 | bool IsCJK(char32_t c); |
| 59 | |
| 60 | bool IsAccent(char32_t c); |
| 61 | |
| 62 | char32_t StripAccent(char32_t c); |
| 63 | |
| 64 | uint64_t Hash64(const char* data, size_t n, uint64_t seed); |
| 65 | |
| 66 | inline uint64_t Hash64(const char* data, size_t n) { |
| 67 | return Hash64(data, n, 0xDECAFCAFFE); |
| 68 | } |
| 69 | |
| 70 | uint64_t Hash64Fast(const char* data, size_t n); |
| 71 | |