mara/hyperhive
mara/hyperhive
Watch
0
0
Fork
You've already forked hyperhive
 0
Code Activity

swarm-stt: only files that start like an audio container reach ffmpeg

Browse source
This commit is contained in:
damocles 2026-11-11 22:28:32 +01:00 • committed by mara
commit 33e63091d5
1 changed files with 97 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

32
swarm-stt/src/engine.rs
View file

@ -36,6 +36,12 @@ pub struct Whisper {
#[async_trait]
impl Engine for Whisper {
async fn transcribe(&self, audio: &[u8]) -> Result<String, String> {
if !crate::sniff::is_audio(audio) {
return Err(
"the file does not look like audio (Ogg, WAV, FLAC, MP3, AMR, MP4/M4A, WebM)"
.to_owned(),
);
}
let dir = tempfile_in(&self.scratch)
.await
.map_err(|e| format!("making a scratch directory: {e}"))?;
@ -157,8 +163,8 @@ mod tests {
async fn audio_goes_through_ffmpeg_then_whisper_and_the_text_is_trimmed() {
let dir = tempfile::tempdir().unwrap();
let engine = whisper(dir.path(), FFMPEG_OK, WHISPER_ECHO);
let text = engine.transcribe(b" hello there \n").await.unwrap();
assert_eq!(text, "hello there");
let text = engine.transcribe(b"OggS hello there \n").await.unwrap();
assert_eq!(text, "OggS hello there");
assert_eq!(
std::fs::read_dir(dir.path().join("scratch"))
.unwrap()
@ -173,7 +179,7 @@ mod tests {
let dir = tempfile::tempdir().unwrap();
let mut engine = whisper(dir.path(), "sleep 60", WHISPER_ECHO);
engine.timeout = std::time::Duration::from_millis(200);
let err = engine.transcribe(b"x").await.unwrap_err();
let err = engine.transcribe(b"OggSx").await.unwrap_err();
assert!(err.contains("timed out"), "{err}");
assert_eq!(
std::fs::read_dir(dir.path().join("scratch"))
@ -192,7 +198,7 @@ mod tests {
record.display()
);
let engine = whisper(dir.path(), &ffmpeg, WHISPER_ECHO);
engine.transcribe(b"x").await.unwrap();
engine.transcribe(b"OggSx").await.unwrap();
let args = std::fs::read_to_string(record).unwrap();
assert!(
args.contains("-protocol_whitelist pipe -i pipe:0"),
@ -200,11 +206,25 @@ mod tests {
);
}
#[tokio::test]
async fn a_file_that_is_not_audio_never_reaches_ffmpeg() {
let dir = tempfile::tempdir().unwrap();
let record = dir.path().join("ran");
let ffmpeg = format!("touch {}", record.display());
let engine = whisper(dir.path(), &ffmpeg, WHISPER_ECHO);
let err = engine
.transcribe(b"#EXTM3U\nfile:///etc/passwd")
.await
.unwrap_err();
assert!(err.contains("does not look like audio"), "{err}");
assert!(!record.exists(), "ffmpeg was started");
}
#[tokio::test]
async fn an_ffmpeg_failure_is_reported_with_its_stderr() {
let dir = tempfile::tempdir().unwrap();
let engine = whisper(dir.path(), "echo 'bad data' >&2; exit 1", WHISPER_ECHO);
let err = engine.transcribe(b"x").await.unwrap_err();
let err = engine.transcribe(b"OggSx").await.unwrap_err();
assert!(
err.contains("could not decode") && err.contains("bad data"),
"{err}"
@ -215,7 +235,7 @@ mod tests {
async fn a_whisper_failure_is_reported_with_its_stderr() {
let dir = tempfile::tempdir().unwrap();
let engine = whisper(dir.path(), FFMPEG_OK, "echo 'no model' >&2; exit 2");
let err = engine.transcribe(b"x").await.unwrap_err();
let err = engine.transcribe(b"OggSx").await.unwrap_err();
assert!(
err.contains("whisper-cli failed") && err.contains("no model"),
"{err}"