enable transcript rename for non-srt files (#179)

* enable transcript rename for non-srt files

* Update transcription-filter-callbacks.cpp

fix edit mistake

* tweak formatting

* Refactor file renaming logic in transcription-filter-callbacks.cpp

* Refactor file renaming logic and handle recording state changes in transcription-filter-callbacks.cpp

* Refactor file renaming logic and handle recording state changes in transcription-filter-callbacks.cpp

---------

Co-authored-by: Stephen <stephen@lenovo-thinkpad.lan>
Co-authored-by: Stephen Schrauger <6665521+schrauger@users.noreply.github.com>
Co-authored-by: schrauger <schrauger@users.noreply.github.com>

src/transcription-filter-callbacks.cpp

@@ -12,6 +12,7 @@
 #include <regex>
 #include <string>
 #include <vector>
+#include <filesystem>
 
 #include "transcription-filter-callbacks.h"
 #include "transcription-utils.h"
@@ -297,6 +298,23 @@ void set_text_callback(struct transcription_filter_data *gf,
 	}
 };
 
+/**
+ * @brief Callback function to handle recording state changes in OBS.
+ *
+ * This function is triggered by OBS frontend events related to recording state changes.
+ * It performs actions based on whether the recording is starting or stopping.
+ *
+ * @param event The OBS frontend event indicating the recording state change.
+ * @param data Pointer to user data, expected to be a struct transcription_filter_data.
+ *
+ * When the recording is starting:
+ * - If saving SRT files and saving only while recording is enabled, it resets the SRT file,
+ *   truncates the existing file, and initializes the sentence number and start timestamp.
+ *
+ * When the recording is stopping:
+ * - If saving only while recording or renaming the file to match the recording is not enabled, it returns immediately.
+ * - Otherwise, it renames the output file to match the recording file name with the appropriate extension.
+ */
 void recording_state_callback(enum obs_frontend_event event, void *data)
 {
 	struct transcription_filter_data *gf_ =
@@ -315,19 +333,38 @@ void recording_state_callback(enum obs_frontend_event event, void *data)
 			gf_->start_timestamp_ms = now_ms();
 		}
 	} else if (event == OBS_FRONTEND_EVENT_RECORDING_STOPPED) {
-		if (gf_->save_srt && gf_->save_only_while_recording &&
-		    gf_->rename_file_to_match_recording) {
+		if (!gf_->save_only_while_recording || !gf_->rename_file_to_match_recording) {
+			return;
+		}
+
+		namespace fs = std::filesystem;
+
+		char *recordingFileName = obs_frontend_get_last_recording();
+		std::string recordingFileNameStr(recordingFileName);
+		bfree(recordingFileName);
+		fs::path recordingPath(recordingFileName);
+		fs::path outputPath(gf_->output_file_path);
+
+		fs::path newPath = recordingPath.stem();
+
+		if (gf_->save_srt) {
 			obs_log(gf_->log_level, "Recording stopped. Rename srt file.");
-			// rename file to match the recording file name with .srt extension
-			// use obs_frontend_get_last_recording to get the last recording file name
-			std::string recording_file_name = obs_frontend_get_last_recording();
-			// remove the extension
-			recording_file_name = recording_file_name.substr(
-				0, recording_file_name.find_last_of("."));
-			std::string srt_file_name = recording_file_name + ".srt";
-			// rename the file
-			std::rename(gf_->output_file_path.c_str(), srt_file_name.c_str());
+			newPath.replace_extension(".srt");
+		} else {
+			obs_log(gf_->log_level, "Recording stopped. Rename transcript file.");
+			std::string newExtension = outputPath.extension().string();
+
+			if (newExtension == recordingPath.extension().string()) {
+				newExtension += ".txt";
+			}
+
+			newPath.replace_extension(newExtension);
 		}
+
+		// make sure newPath is next to the recording file
+		newPath = recordingPath.parent_path() / newPath.filename();
+
+		fs::rename(outputPath, newPath);
 	}
 }
 

src/transcription-filter-properties.cpp

@@ -594,7 +594,7 @@ void transcription_filter_defaults(obs_data_t *s)
 	obs_data_set_default_string(s, "translation_model_path_external", "");
 	obs_data_set_default_int(s, "translate_input_tokenization_style", INPUT_TOKENIZAION_M2M100);
 	obs_data_set_default_double(s, "sentence_psum_accept_thresh", 0.4);
-	obs_data_set_default_bool(s, "partial_group", false);
+	obs_data_set_default_bool(s, "partial_group", true);
 	obs_data_set_default_int(s, "partial_latency", 1100);
 
 	// translation options

src/transcription-filter-utils.cpp

@@ -28,7 +28,12 @@ void create_obs_text_source_if_needed()
 		// set source settings
 		obs_data_t *source_settings = obs_source_get_settings(source);
 		obs_data_set_bool(source_settings, "word_wrap", true);
-		obs_data_set_int(source_settings, "custom_width", 1760);
+		obs_data_set_bool(source_settings, "extents", true);
+		obs_data_set_bool(source_settings, "outline", true);
+		obs_data_set_int(source_settings, "outline_color", 4278190080);
+		obs_data_set_int(source_settings, "outline_size", 7);
+		obs_data_set_int(source_settings, "extents_cx", 1500);
+		obs_data_set_int(source_settings, "extents_cy", 230);
 		obs_data_t *font_data = obs_data_create();
 		obs_data_set_string(font_data, "face", "Arial");
 		obs_data_set_string(font_data, "style", "Regular");
@@ -62,7 +67,7 @@ void create_obs_text_source_if_needed()
 
 bool add_sources_to_list(void *list_property, obs_source_t *source)
 {
-	auto source_id = obs_source_get_id(source);
+	const char *source_id = obs_source_get_id(source);
 	if (strcmp(source_id, "text_ft2_source_v2") != 0 &&
 	    strcmp(source_id, "text_gdiplus_v3") != 0 &&
 	    strcmp(source_id, "text_gdiplus_v2") != 0) {

src/transcription-filter.cpp

@@ -404,7 +404,7 @@ void transcription_filter_update(void *data, obs_data_t *s)
 		}
 	}
 
-	if (gf->context != nullptr && obs_source_enabled(gf->context)) {
+	if (gf->context != nullptr && (obs_source_enabled(gf->context) || gf->initial_creation)) {
 		if (gf->initial_creation) {
 			obs_log(LOG_INFO, "Initial filter creation and source enabled");
 
@@ -423,6 +423,8 @@ void transcription_filter_update(void *data, obs_data_t *s)
 				update_whisper_model(gf);
 			}
 		}
+	} else {
+		obs_log(LOG_INFO, "Filter not enabled, not updating whisper model.");
 	}
 }
 
@@ -449,6 +451,7 @@ void *transcription_filter_create(obs_data_t *settings, obs_source_t *filter)
 		obs_data_get_bool(settings, "rename_file_to_match_recording");
 	gf->process_while_muted = obs_data_get_bool(settings, "process_while_muted");
 	gf->buffered_output = obs_data_get_bool(settings, "buffered_output");
+	gf->initial_creation = true;
 
 	for (size_t i = 0; i < gf->channels; i++) {
 		circlebuf_init(&gf->input_buffers[i]);