Update README and refactor timed metadata handling for improved trans…

…lation modes

README.md

@@ -31,15 +31,15 @@ Current Features:
 - Support for 100+ languages with dialect recognition
 - Streaming-optimized performance with minimal latency
 - Multiple cloud provider options for transcription and translation
-- Caption output in multiple formats (.txt, .srt, .vtt)
+- Caption output in multiple formats (.txt, .srt)
 - Sync'ed captions with OBS recording timestamps
 - Direct streaming to platforms (YouTube, Twitch) with embedded captions
-- Advanced text filtering and customization options
 - Partial transcriptions for a streaming-captions experience
-- Custom vocabulary and pronunciation support
-- Professional terminology handling for specific industries
 
 Roadmap:
+- Custom vocabulary and pronunciation support
+- Professional terminology handling for specific industries
+- Advanced text filtering and customization options
 - Speaker diarization for multi-speaker environments
 - Advanced profanity filtering options
 - Custom translation glossaries
@@ -86,8 +86,17 @@ $ ./.github/scripts/build-linux
 
 ### Windows
 
+Windows also needs Conan for OpenSSL. Run `conan` to get the dependency (make sure to run `conan` on the `conanfile_win.txt`):
+```powershell
+> pip install conan
+> conan profile detect --force
+> conan install .\conanfile_win.txt --output-folder=./build_conan --build=missing -g CMakeDeps 
+```
+
+Build the plugin:
+
 ```powershell
-> .github/scripts/Build-Windows.ps1 -Configuration Release
+> .\.github\scripts\Build-Windows.ps1 -Configuration Release
 ```
 
 If you're developing the plugin, I find this command to be useful for direct deploymet into OBS after building:

src/cloudvocal-callbacks.cpp

@@ -18,6 +18,7 @@
 #include "cloudvocal-callbacks.h"
 #include "cloudvocal-utils.h"
 #include "plugin-support.h"
+#include "timed-metadata/timed-metadata-utils.h"
 
 void send_caption_to_source(const std::string &target_source_name, const std::string &caption,
 			    struct cloudvocal_data *gf)
@@ -217,6 +218,8 @@ void set_text_callback(struct cloudvocal_data *gf, const DetectionResultWithText
 		}
 	}
 
+	// should translate if translation is enabled and the result is full
+	// or if partial translations are enabled
 	bool should_translate = (gf->translate_only_full_sentences
 					 ? result.result == DETECTION_RESULT_SPEECH
 					 : true) &&
@@ -239,7 +242,18 @@ void set_text_callback(struct cloudvocal_data *gf, const DetectionResultWithText
 									 translated_sentence_cloud,
 									 gf->target_lang);
 				}
+				if (gf->send_timed_metadata) {
+					send_timed_metadata_to_server(gf, SOURCE_AND_TARGET,
+								      result.text, result.language,
+								      translated_sentence_cloud,
+								      gf->target_lang);
+				}
 			});
+	} else {
+		if (gf->send_timed_metadata) {
+			send_timed_metadata_to_server(gf, ONLY_SOURCE, result.text, result.language,
+						      "", "");
+		}
 	}
 
 	// send the original text to the output

src/timed-metadata/timed-metadata-utils.cpp

@@ -130,9 +130,8 @@ void send_timed_metadata_to_ivs_endpoint(struct cloudvocal_data *gf, Translation
 
 	// Construct the inner JSON string
 	nlohmann::json inner_meta_data;
-	if (mode == NON_WHISPER_TRANSLATE) {
-		obs_log(gf->log_level,
-			"send_timed_metadata_to_ivs_endpoint - NON_WHISPER_TRANSLATE");
+	if (mode == SOURCE_AND_TARGET) {
+		obs_log(gf->log_level, "send_timed_metadata_to_ivs_endpoint - SOURCE_AND_TARGET");
 		nlohmann::json array;
 		if (!source_text.empty()) {
 			array.push_back({{"language", source_lang}, {"text", source_text}});
@@ -146,13 +145,13 @@ void send_timed_metadata_to_ivs_endpoint(struct cloudvocal_data *gf, Translation
 			return;
 		}
 		inner_meta_data = {{"captions", array}};
-	} else if (mode == WHISPER_TRANSLATE) {
+	} else if (mode == ONLY_TARGET) {
 		if (target_text.empty()) {
 			obs_log(gf->log_level,
 				"send_timed_metadata_to_ivs_endpoint - target text empty");
 			return;
 		}
-		obs_log(gf->log_level, "send_timed_metadata_to_ivs_endpoint - WHISPER_TRANSLATE");
+		obs_log(gf->log_level, "send_timed_metadata_to_ivs_endpoint - ONLY_TARGET");
 		inner_meta_data = {
 			{"captions", {{{"language", target_lang}, {"text", target_text}}}}};
 	} else {
@@ -161,7 +160,7 @@ void send_timed_metadata_to_ivs_endpoint(struct cloudvocal_data *gf, Translation
 				"send_timed_metadata_to_ivs_endpoint - source text empty");
 			return;
 		}
-		obs_log(gf->log_level, "send_timed_metadata_to_ivs_endpoint - transcription mode");
+		obs_log(gf->log_level, "send_timed_metadata_to_ivs_endpoint - ONLY_SOURCE");
 		inner_meta_data = {
 			{"captions", {{{"language", source_lang}, {"text", source_text}}}}};
 	}

src/timed-metadata/timed-metadata-utils.h

@@ -6,7 +6,7 @@
 
 #include "cloudvocal-data.h"
 
-enum Translation_Mode { WHISPER_TRANSLATE, NON_WHISPER_TRANSLATE, TRANSCRIBE };
+enum Translation_Mode { ONLY_TARGET, SOURCE_AND_TARGET, ONLY_SOURCE };
 
 void send_timed_metadata_to_server(struct cloudvocal_data *gf, Translation_Mode mode,
 				   const std::string &source_text, const std::string &source_lang,