[feat] add TTY-aware progress management with indicatif and file-specific progress bars

This commit is contained in:
2025-08-12 08:11:28 +02:00
parent 2cc5e49131
commit 041e504cb2
6 changed files with 241 additions and 22 deletions

View File

@@ -35,12 +35,14 @@ pub trait TranscribeBackend {
/// - speaker: label to attach to all produced segments.
/// - lang_opt: optional language hint (e.g., "en"); None means auto/multilingual model default.
/// - gpu_layers: optional GPU layer count if applicable (ignored by some backends).
/// - progress_cb: optional callback receiving percentage [0..=100] updates.
fn transcribe(
&self,
audio_path: &Path,
speaker: &str,
lang_opt: Option<&str>,
gpu_layers: Option<u32>,
progress_cb: Option<&(dyn Fn(i32) + Send + Sync)>,
) -> Result<Vec<OutputEntry>>;
}
@@ -148,8 +150,9 @@ impl TranscribeBackend for CpuBackend {
speaker: &str,
lang_opt: Option<&str>,
_gpu_layers: Option<u32>,
progress_cb: Option<&(dyn Fn(i32) + Send + Sync)>,
) -> Result<Vec<OutputEntry>> {
transcribe_with_whisper_rs(audio_path, speaker, lang_opt)
transcribe_with_whisper_rs(audio_path, speaker, lang_opt, progress_cb)
}
}
@@ -163,9 +166,10 @@ impl TranscribeBackend for CudaBackend {
speaker: &str,
lang_opt: Option<&str>,
_gpu_layers: Option<u32>,
progress_cb: Option<&(dyn Fn(i32) + Send + Sync)>,
) -> Result<Vec<OutputEntry>> {
// whisper-rs uses enabled CUDA feature at build time; call same code path
transcribe_with_whisper_rs(audio_path, speaker, lang_opt)
transcribe_with_whisper_rs(audio_path, speaker, lang_opt, progress_cb)
}
}
@@ -179,8 +183,9 @@ impl TranscribeBackend for HipBackend {
speaker: &str,
lang_opt: Option<&str>,
_gpu_layers: Option<u32>,
progress_cb: Option<&(dyn Fn(i32) + Send + Sync)>,
) -> Result<Vec<OutputEntry>> {
transcribe_with_whisper_rs(audio_path, speaker, lang_opt)
transcribe_with_whisper_rs(audio_path, speaker, lang_opt, progress_cb)
}
}
@@ -194,6 +199,7 @@ impl TranscribeBackend for VulkanBackend {
_speaker: &str,
_lang_opt: Option<&str>,
_gpu_layers: Option<u32>,
_progress_cb: Option<&(dyn Fn(i32) + Send + Sync)>,
) -> Result<Vec<OutputEntry>> {
Err(anyhow!(
"Vulkan backend not yet wired to whisper.cpp FFI. Build with --features gpu-vulkan and ensure Vulkan SDK is installed. How to fix: install Vulkan loader (libvulkan), set VULKAN_SDK, and run cargo build --features gpu-vulkan."
@@ -301,8 +307,13 @@ pub(crate) fn transcribe_with_whisper_rs(
audio_path: &Path,
speaker: &str,
lang_opt: Option<&str>,
progress_cb: Option<&(dyn Fn(i32) + Send + Sync)>,
) -> Result<Vec<OutputEntry>> {
if let Some(cb) = progress_cb { cb(0); }
let pcm = decode_audio_to_pcm_f32_ffmpeg(audio_path)?;
if let Some(cb) = progress_cb { cb(5); }
let model = find_model_file()?;
let is_en_only = model
.file_name()
@@ -341,6 +352,7 @@ pub(crate) fn transcribe_with_whisper_rs(
.map_err(|e| anyhow!("Failed to create Whisper state: {:?}", e))?;
Ok::<_, anyhow::Error>((ctx, state))
})?;
if let Some(cb) = progress_cb { cb(20); }
let mut params =
whisper_rs::FullParams::new(whisper_rs::SamplingStrategy::Greedy { best_of: 1 });
@@ -352,13 +364,16 @@ pub(crate) fn transcribe_with_whisper_rs(
if let Some(lang) = lang_opt {
params.set_language(Some(lang));
}
if let Some(cb) = progress_cb { cb(30); }
crate::with_suppressed_stderr(|| {
if let Some(cb) = progress_cb { cb(40); }
state
.full(params, &pcm)
.map_err(|e| anyhow!("Whisper full() failed: {:?}", e))
})?;
if let Some(cb) = progress_cb { cb(90); }
let num_segments = state
.full_n_segments()
.map_err(|e| anyhow!("Failed to get segments: {:?}", e))?;
@@ -383,5 +398,6 @@ pub(crate) fn transcribe_with_whisper_rs(
text: text.trim().to_string(),
});
}
if let Some(cb) = progress_cb { cb(100); }
Ok(items)
}