mirrors/obs-localvocal
1
0
Fork 0
mirror of https://github.com/occ-ai/obs-localvocal synced 2024-11-07 18:57:14 +00:00
Code Issues Packages Projects Releases Wiki Activity

add text file output

Browse Source
This commit is contained in:
Roy Shilkrot 2023-09-08 12:25:10 -04:00
parent 43416531d8
commit 11a51e7512
2 changed files with 60 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
src/transcription-filter-data.h
View File

@ -58,6 +58,8 @@ struct transcription_filter_data {
std::mutex *text_source_mutex = nullptr;
// Callback to set the text in the output text source (subtitles)
std::function<void(const std::string &str)> setTextCallback;
// Output file path to write the subtitles
std::string output_file_path;
// Use std for thread and mutex
std::thread whisper_thread;

42
src/transcription-filter.cpp
View File

@ -8,6 +8,7 @@
#include "model-utils/model-downloader.h"
#include <algorithm>
#include <fstream>
inline enum speaker_layout convert_speaker_layout(uint8_t channels)
{
@ -176,6 +177,12 @@ void acquire_weak_text_source_ref(struct transcription_filter_data *gf)
void set_text_callback(struct transcription_filter_data *gf, const std::string &str)
{
if (gf->output_file_path != "") {
// Write to file, do not append
std::ofstream output_file(gf->output_file_path, std::ios::out | std::ios::trunc);
output_file << str;
output_file.close();
} else {
if (!gf->text_source_mutex) {
obs_log(LOG_ERROR, "text_source_mutex is null");
return;
@ -201,6 +208,7 @@ void set_text_callback(struct transcription_filter_data *gf, const std::string &
obs_data_set_string(text_settings, "text", str.c_str());
obs_source_update(target, text_settings);
obs_source_release(target);
}
};
void transcription_filter_update(void *data, obs_data_t *s)
@ -219,7 +227,9 @@ void transcription_filter_update(void *data, obs_data_t *s)
obs_weak_source_t *old_weak_text_source = NULL;
if (strcmp(new_text_source_name, "none") == 0 ||
strcmp(new_text_source_name, "(null)") == 0 || strcmp(new_text_source_name, "") == 0) {
strcmp(new_text_source_name, "(null)") == 0 ||
strcmp(new_text_source_name, "text_file") == 0 ||
strlen(new_text_source_name) == 0) {
// new selected text source is not valid, release the old one
if (gf->text_source) {
if (!gf->text_source_mutex) {
@ -234,6 +244,11 @@ void transcription_filter_update(void *data, obs_data_t *s)
bfree(gf->text_source_name);
gf->text_source_name = nullptr;
}
if (strcmp(new_text_source_name, "text_file") == 0) {
gf->output_file_path = obs_data_get_string(s, "subtitle_output_filename");
} else {
gf->output_file_path = "";
}
} else {
// new selected text source is valid, check if it's different from the old one
if (gf->text_source_name == nullptr ||
@ -430,7 +445,7 @@ void transcription_filter_deactivate(void *data)
void transcription_filter_defaults(obs_data_t *s)
{
obs_data_set_default_bool(s, "vad_enabled", true);
obs_data_set_default_int(s, "log_level", LOG_DEBUG);
obs_data_set_default_int(s, "log_level", LOG_INFO);
obs_data_set_default_bool(s, "log_words", true);
obs_data_set_default_string(s, "whisper_model_path", "models/ggml-tiny.en.bin");
obs_data_set_default_string(s, "whisper_language_select", "en");
@ -474,13 +489,28 @@ obs_properties_t *transcription_filter_properties(void *data)
obs_property_list_add_int(list, "WARNING", LOG_WARNING);
obs_properties_add_bool(ppts, "log_words", "Log output words");
obs_property_t *sources =
obs_properties_add_list(ppts, "subtitle_sources", "Subtitles Text Source",
obs_property_t *subs_output =
obs_properties_add_list(ppts, "subtitle_sources", "Subtitles Output",
OBS_COMBO_TYPE_LIST, OBS_COMBO_FORMAT_STRING);
// Add "none" option
obs_property_list_add_string(sources, "None / No output", "none");
obs_property_list_add_string(subs_output, "None / No output", "none");
obs_property_list_add_string(subs_output, "Text File output", "text_file");
// Add text sources
obs_enum_sources(add_sources_to_list, sources);
obs_enum_sources(add_sources_to_list, subs_output);
obs_properties_add_path(ppts, "subtitle_output_filename", "Output filename", OBS_PATH_FILE_SAVE, "Text (*.txt)", NULL);
obs_property_set_modified_callback(subs_output, [](obs_properties_t *props, obs_property_t *property, obs_data_t *settings) {
const char *new_output = obs_data_get_string(settings, "subtitle_sources");
if (strcmp(new_output, "text_file") == 0) {
// Show the output filename selection input
obs_property_set_visible(obs_properties_get(props, "subtitle_output_filename"), true);
} else {
// Hide the output filename selection input
obs_property_set_visible(obs_properties_get(props, "subtitle_output_filename"), false);
}
return true;
});
// Add a list of available whisper models to download
obs_property_t *whisper_models_list =