41 lines
1.6 KiB
C++
41 lines
1.6 KiB
C++
#pragma once
|
|
|
|
#include <string>
|
|
#include <vector>
|
|
#include <cstdint>
|
|
|
|
// Read WAV audio file and store the PCM data into pcmf32
|
|
// fname can be a buffer of WAV data instead of a filename
|
|
// The sample rate of the audio must be equal to COMMON_SAMPLE_RATE
|
|
// If stereo flag is set and the audio has 2 channels, the pcmf32s will contain 2 channel PCM
|
|
bool read_audio_data(
|
|
const std::string & fname,
|
|
std::vector<float> & pcmf32,
|
|
std::vector<std::vector<float>> & pcmf32s,
|
|
bool stereo);
|
|
|
|
// decode audio bytes already held in memory (uploaded file, network buffer)
|
|
bool read_audio_data(
|
|
const char * buffer,
|
|
size_t buffer_size,
|
|
std::vector<float> & pcmf32,
|
|
std::vector<std::vector<float>> & pcmf32s,
|
|
bool stereo);
|
|
|
|
// convert timestamp to string, 6000 -> 01:00.000
|
|
std::string to_timestamp(int64_t t, bool comma = false);
|
|
|
|
// given a timestamp get the sample
|
|
int timestamp_to_sample(int64_t t, int n_samples, int whisper_sample_rate);
|
|
|
|
// Returns the number of trailing bytes still needed for s to end on a complete UTF-8 codepoint.
|
|
int utf8_trailing_bytes_needed(const std::string & s);
|
|
|
|
// Returns s with every byte that is not part of a valid, complete UTF-8 sequence removed.
|
|
// Byte-level decoder tokens can leave a lone UTF-8 lead byte or an orphan continuation byte
|
|
// at a segment boundary; dropping them keeps text output well-formed UTF-8 (issue #3760).
|
|
std::string utf8_sanitize(const std::string & s);
|
|
|
|
// write text to file, and call system("command voice_id file")
|
|
bool speak_with_file(const std::string & command, const std::string & text, const std::string & path, int voice_id);
|