A free, open source, and extensible speech-to-text application that works completely offline.
remove some debug logging, plus some general minor cleanup
CJ Pais6e8128defd84
diff --git a/src-tauri/src/actions.rs b/src-tauri/src/actions.rsindex ac40c765..2045db44 100644--- a/src-tauri/src/actions.rs+++ b/src-tauri/src/actions.rs@@ -41,7 +41,7 @@ impl ShortcutAction for TranscribeAction {if let Some(samples) = rm.stop_recording(&binding_id) {match tm.transcribe(samples) {Ok(transcription) => {- println!("Global Shortcut Transcription: {}", transcription);+ println!("Transcription Result: {}", transcription);if !transcription.is_empty() {utils::paste(transcription, ah);}diff --git a/src-tauri/src/lib.rs b/src-tauri/src/lib.rsindex 26a714cb..bbad689b 100644--- a/src-tauri/src/lib.rs+++ b/src-tauri/src/lib.rs@@ -85,13 +85,6 @@ pub fn run() {let autostart_manager = app.autolaunch();// Enable autostartlet _ = autostart_manager.enable();- // Check enable state- println!(- "registered for autostart? {}",- autostart_manager.is_enabled().unwrap()- );- // Disable autostart- let _ = autostart_manager.disable();let recording_manager = Arc::new(AudioRecordingManager::new(app).expect("Failed to initialize recording manager"),@@ -121,9 +114,6 @@ pub fn run() {if let Err(e) = res {println!("Failed to set activation policy: {}", e);}-- // TODO may be different on windows, this works for macos- // tauri::AppHandle::hide(window.app_handle()).unwrap();}}_ => {}diff --git a/src-tauri/src/managers/audio.rs b/src-tauri/src/managers/audio.rsindex fb6b8b01..f49a823f 100644--- a/src-tauri/src/managers/audio.rs+++ b/src-tauri/src/managers/audio.rs@@ -137,14 +137,9 @@ impl AudioRecordingManager {if let Ok(mut vad) = vad_clone.lock() {match vad.compute(&chunk) {Ok(result) => {- println!("VAD PROB {}", result.prob);if result.prob > 0.1 {let mut buffer = buffer_clone.lock().unwrap();buffer.extend_from_slice(&chunk);- println!(- "EXTENDED BUFFER. BUFFER LENGTH {}",- buffer.len()- );}}Err(error) => {@@ -284,14 +279,10 @@ impl AudioRecordingManager {println!("Stopped recording for binding {}", binding_id);let mut buffer = self.buffer.lock().unwrap();- println!("READING BUFFER");let audio_data: Vec<f32> = buffer.drain(..).collect();- println!("READ BUFFER");let samples = audio_data.len();- println!("SAMPLE AFTER VAD {}, DURATION {}", samples, samples / 1000);-if samples < WHISPER_SAMPLE_RATE && samples > 1000 {let target_samples = WHISPER_SAMPLE_RATE * 5 / 4; // sample rate * 1.25let mut padded_audio = audio_data;diff --git a/src-tauri/src/managers/transcription.old b/src-tauri/src/managers/transcription.olddeleted file mode 100644index 7970d487..00000000--- a/src-tauri/src/managers/transcription.old+++ /dev/null@@ -1,67 +0,0 @@-// use kalosm::sound::rodio::buffer::SamplesBuffer;-// use kalosm::sound::{ModelLoadingProgress, TextStream, Whisper, WhisperBuilder, WhisperSource};--// pub struct TranscriptionManager {-// model: Whisper,-// }--// impl TranscriptionManager {-// pub fn new() -> Result<Self, Box<dyn std::error::Error>> {-// let model = tauri::async_runtime::block_on(async {-// WhisperBuilder::default()-// .with_source(WhisperSource::Small)-// .build_with_loading_handler(|progress| match progress {-// ModelLoadingProgress::Downloading {-// source,-// start_time,-// progress,-// } => {-// let progress = (progress * 100.0) as u32;-// let elapsed = start_time.elapsed().as_secs_f32();-// println!("Downloading file {source} {progress}% ({elapsed}s)");-// }-// ModelLoadingProgress::Loading { progress } => {-// let progress = (progress * 100.0) as u32;-// println!("Loading model {progress}%");-// }-// })-// .await-// })?;--// Ok(Self { model })-// }--// pub fn transcribe(&self, audio: SamplesBuffer<f32>) -> Result<String, anyhow::Error> {-// let mut result = String::new();--// println!("lot line transcribing");-// let mut transcribed = self.model.transcribe(audio)?;-// println!("Sync transcribing finished");--// use std::time::Instant;--// let start = Instant::now();-// tauri::async_runtime::block_on(async {-// transcribed.to_std_out().await.unwrap();-// });-// let duration = start.elapsed();-// println!("Time elapsed: {:?}", duration);--// // let segments = tauri::async_runtime::block_on(async {-// // let mut segments = Vec::new();-// // let mut stream = transcribed;-// // while let Some(segment) = stream.next().await {-// // segments.push(segment);-// // }-// // segments-// // });--// // for segment in segments {-// // if segment.probability_of_no_speech() < 0.10 {-// // result.push_str(segment.text());-// // }-// // }--// Ok(result)-// }-// }\ No newline at end of filediff --git a/src-tauri/src/managers/transcription.rs b/src-tauri/src/managers/transcription.rsindex 9757c664..4ed0f6fa 100644--- a/src-tauri/src/managers/transcription.rs+++ b/src-tauri/src/managers/transcription.rs@@ -71,7 +71,7 @@ impl TranscriptionManager {}let et = std::time::Instant::now();- println!("\n\ntook {}ms", (et - st).as_millis());+ println!("\ntook {}ms", (et - st).as_millis());Ok(result.trim().to_string())}