microsoft/onnxruntime-extensions
Publicmirrored from https://github.com/microsoft/onnxruntime-extensionsAvailable
ocos/kernels/string_regex_replace.cc
73lines · modecode
| 1 | // Copyright (c) Microsoft Corporation. All rights reserved. |
| 2 | // Licensed under the MIT License. |
| 3 | |
| 4 | #include "string_regex_replace.hpp" |
| 5 | #include <vector> |
| 6 | #include <cmath> |
| 7 | #include <algorithm> |
| 8 | #include "re2/re2.h" |
| 9 | |
| 10 | KernelStringRegexReplace::KernelStringRegexReplace(OrtApi api) : BaseKernel(api) { |
| 11 | } |
| 12 | |
| 13 | void KernelStringRegexReplace::Compute(OrtKernelContext* context) { |
| 14 | // Setup inputs |
| 15 | const OrtValue* input = ort_.KernelContext_GetInput(context, 0); |
| 16 | const std::string* str_input = ort_.GetTensorData<std::string>(input); |
| 17 | const OrtValue* pattern = ort_.KernelContext_GetInput(context, 1); |
| 18 | const std::string* str_pattern = ort_.GetTensorData<std::string>(pattern); |
| 19 | const OrtValue* rewrite = ort_.KernelContext_GetInput(context, 2); |
| 20 | const std::string* str_rewrite = ort_.GetTensorData<std::string>(rewrite); |
| 21 | |
| 22 | // Verifications |
| 23 | OrtTensorDimensions pattern_dimensions(ort_, pattern); |
| 24 | OrtTensorDimensions rewrite_dimensions(ort_, rewrite); |
| 25 | if (pattern_dimensions.size() != 1 || pattern_dimensions[0] != 1) |
| 26 | throw std::runtime_error(MakeString( |
| 27 | "pattern (second input) must contain only one element. It has ", |
| 28 | pattern_dimensions.size(), " dimensions.")); |
| 29 | if (rewrite_dimensions.size() != 1 || rewrite_dimensions[0] != 1) |
| 30 | throw std::runtime_error(MakeString( |
| 31 | "rewrite (third input) must contain only one element. It has ", |
| 32 | rewrite_dimensions.size(), " dimensions.")); |
| 33 | |
| 34 | // Setup output |
| 35 | OrtTensorDimensions dimensions(ort_, input); |
| 36 | OrtValue* output = ort_.KernelContext_GetOutput(context, 0, dimensions.data(), dimensions.size()); |
| 37 | std::string* out = ort_.GetTensorMutableData<std::string>(output); |
| 38 | |
| 39 | OrtTensorTypeAndShapeInfo* output_info = ort_.GetTensorTypeAndShape(output); |
| 40 | int64_t size = ort_.GetTensorShapeElementCount(output_info); |
| 41 | ort_.ReleaseTensorTypeAndShapeInfo(output_info); |
| 42 | |
| 43 | re2::StringPiece piece(*str_rewrite); |
| 44 | re2::RE2 reg(*str_pattern); |
| 45 | |
| 46 | // Do computation |
| 47 | for (int64_t i = 0; i < size; i++) { |
| 48 | out[i] = str_input[i]; |
| 49 | re2::RE2::Replace(out + i, reg, piece); |
| 50 | } |
| 51 | } |
| 52 | |
| 53 | void* CustomOpStringRegexReplace::CreateKernel(OrtApi api, const OrtKernelInfo* /* info */) { |
| 54 | return new KernelStringRegexReplace(api); |
| 55 | }; |
| 56 | |
| 57 | const char* CustomOpStringRegexReplace::GetName() const { return "StringRegexReplace"; }; |
| 58 | |
| 59 | size_t CustomOpStringRegexReplace::GetInputTypeCount() const { |
| 60 | return 3; |
| 61 | }; |
| 62 | |
| 63 | ONNXTensorElementDataType CustomOpStringRegexReplace::GetInputType(size_t /*index*/) const { |
| 64 | return ONNX_TENSOR_ELEMENT_DATA_TYPE_STRING; |
| 65 | }; |
| 66 | |
| 67 | size_t CustomOpStringRegexReplace::GetOutputTypeCount() const { |
| 68 | return 1; |
| 69 | }; |
| 70 | |
| 71 | ONNXTensorElementDataType CustomOpStringRegexReplace::GetOutputType(size_t /*index*/) const { |
| 72 | return ONNX_TENSOR_ELEMENT_DATA_TYPE_STRING; |
| 73 | }; |
| 74 | |