microsoft/onnxruntime-extensions
Publicmirrored fromhttps://github.com/microsoft/onnxruntime-extensionsAvailable
include/ortx_extractor.h
91lines · modecode
| 1 | // C ABI header file for the onnxruntime-extensions tokenization module |
| 2 | |
| 3 | #pragma once |
| 4 | |
| 5 | #include "ortx_utils.h" |
| 6 | |
| 7 | typedef OrtxObject OrtxFeatureExtractor; |
| 8 | typedef OrtxObject OrtxRawAudios; |
| 9 | |
| 10 | #ifdef __cplusplus |
| 11 | extern "C" { |
| 12 | #endif |
| 13 | |
| 14 | /** |
| 15 | * @brief Creates a feature extractor object. |
| 16 | * |
| 17 | * This function creates a feature extractor object based on the provided feature definition. |
| 18 | * |
| 19 | * @param[out] extractor Pointer to a pointer to the created feature extractor object. |
| 20 | * @param[in] fe_def The feature definition used to create the feature extractor. |
| 21 | * |
| 22 | * @return An error code indicating the result of the operation. |
| 23 | */ |
| 24 | extError_t ORTX_API_CALL OrtxCreateSpeechFeatureExtractor(OrtxFeatureExtractor** extractor, const char* fe_def); |
| 25 | |
| 26 | /** |
| 27 | * Loads a collection of audio files into memory. |
| 28 | * |
| 29 | * This function loads a collection of audio files specified by the `audio_paths` array |
| 30 | * into memory and returns a pointer to the loaded audio data in the `audios` parameter. |
| 31 | * |
| 32 | * @param audios A pointer to a pointer that will be updated with the loaded audio data. |
| 33 | * The caller is responsible for freeing the memory allocated for the audio data. |
| 34 | * @param audio_paths An array of strings representing the paths to the audio files to be loaded. |
| 35 | * @param num_audios The number of audio files to be loaded. |
| 36 | * |
| 37 | * @return An `extError_t` value indicating the success or failure of the operation. |
| 38 | */ |
| 39 | extError_t ORTX_API_CALL OrtxLoadAudios(OrtxRawAudios** audios, const char* const* audio_paths, size_t num_audios); |
| 40 | |
| 41 | /** |
| 42 | * @brief Creates an array of raw audio objects, which refers to the audio data and sizes provided. |
| 43 | * |
| 44 | * This function creates an array of raw audio objects based on the provided data and sizes. The data will be stored in |
| 45 | * the `audios` parameter. |
| 46 | * |
| 47 | * @param audios Pointer to the variable that will hold the created raw audio objects. |
| 48 | * @param data Array of pointers to the audio data. |
| 49 | * @param sizes Array of pointers to the sizes of the audio data. |
| 50 | * @param num_audios Number of audio objects to create. |
| 51 | * |
| 52 | * @return extError_t Error code indicating the success or failure of the operation. |
| 53 | */ |
| 54 | extError_t ORTX_API_CALL OrtxCreateRawAudios(OrtxRawAudios** audios, const void* data[], const int64_t sizes[], size_t num_audios); |
| 55 | |
| 56 | /** |
| 57 | * @brief Calculates the log mel spectrogram for a given audio using the specified feature extractor. |
| 58 | * |
| 59 | * This function takes an instance of the OrtxFeatureExtractor struct, an instance of the OrtxRawAudios struct, |
| 60 | * and a pointer to an OrtxTensorResult pointer. It calculates the log mel spectrogram for the given audio using |
| 61 | * the specified feature extractor and stores the result in the provided log_mel pointer. |
| 62 | * |
| 63 | * @param extractor The feature extractor to use for calculating the log mel spectrogram. |
| 64 | * @param audio The raw audio data to process. |
| 65 | * @param log_mel A pointer to an OrtxTensorResult pointer where the result will be stored. |
| 66 | * @return An extError_t value indicating the success or failure of the operation. |
| 67 | */ |
| 68 | extError_t ORTX_API_CALL OrtxSpeechLogMel(OrtxFeatureExtractor* extractor, OrtxRawAudios* audio, OrtxTensorResult** log_mel); |
| 69 | |
| 70 | |
| 71 | /** |
| 72 | * @brief Extracts log-mel features from raw audio data using a feature extractor. |
| 73 | * |
| 74 | * This function processes the input audio buffers through the provided feature extractor, |
| 75 | * producing log-mel spectrogram outputs suitable for inference or further signal analysis. |
| 76 | * |
| 77 | * @param extractor A pointer to an OrtxFeatureExtractor object that defines the feature |
| 78 | * extraction pipeline and processing parameters. |
| 79 | * @param audio A pointer to an OrtxRawAudios structure containing raw audio data buffers |
| 80 | * and associated metadata (e.g., sampling rate, channels). |
| 81 | * @param result A pointer to an OrtxTensorResult pointer that will be allocated and set to |
| 82 | * hold the resulting log-mel spectrogram data and other outputs based on json configuration. |
| 83 | * |
| 84 | * @return An extError_t value indicating success or error status. Returns |
| 85 | * EXT_SUCCESS on success, or an appropriate error code if extraction fails. |
| 86 | */ |
| 87 | extError_t ORTX_API_CALL OrtxFeatureExtraction(OrtxFeatureExtractor* extractor, OrtxRawAudios* audio, OrtxTensorResult** result); |
| 88 | |
| 89 | #ifdef __cplusplus |
| 90 | } |
| 91 | #endif |