Watch
0
0
Fork
You've already forked hyperhive
0
hyperhive/swarm-stt/src/sniff.rs

64 lines
2.3 KiB
Rust

//! A check of the first bytes of a file against the audio containers voice
//! memos come in, before anything decodes it.
//!
//! The audio comes from matrix rooms with outside people, so the decoder only
//! ever sees files that start like an audio container. A file that claims to
//! be anything else (a playlist, an image, a script) stops here, which keeps
//! most of `ffmpeg`'s demuxers out of reach.
/// `true` when `bytes` start like Ogg (Opus, Vorbis), WAV, FLAC, MP3 (with or
/// without an `ID3` tag), AMR, `MP4` or `M4A`, Matroska or `WebM`.
#[must_use]
pub fn is_audio(bytes: &[u8]) -> bool {
let starts = |magic: &[u8]| bytes.starts_with(magic);
starts(b"OggS")
|| (starts(b"RIFF") && bytes.get(8..12) == Some(b"WAVE"))
|| starts(b"fLaC")
|| starts(b"ID3")
|| starts(b"#!AMR")
// An MPEG audio frame: eleven set bits of sync.
|| matches!(bytes, [0xFF, second, ..] if second & 0xE0 == 0xE0)
// The `ftyp` box of an ISO base media file (MP4, M4A, 3GP).
|| bytes.get(4..8) == Some(b"ftyp")
// The EBML header of Matroska and WebM.
|| starts(&[0x1A, 0x45, 0xDF, 0xA3])
}
#[cfg(test)]
mod tests {
use super::is_audio;
#[test]
fn the_containers_voice_memos_come_in_pass() {
for (name, head) in [
("ogg", &b"OggS\0\x02"[..]),
("wav", b"RIFF\x24\0\0\0WAVEfmt "),
("flac", b"fLaC\0"),
("mp3 with a tag", b"ID3\x04\0"),
("mp3 frame", &[0xFF, 0xFB, 0x90, 0x00]),
("amr", b"#!AMR\n"),
("m4a", b"\0\0\0\x20ftypM4A "),
("webm", &[0x1A, 0x45, 0xDF, 0xA3, 0x9F]),
] {
assert!(is_audio(head), "{name} was refused");
}
}
#[test]
fn files_that_are_not_audio_containers_are_refused() {
for (name, head) in [
("an HLS playlist", &b"#EXTM3U\n#EXT-X-VERSION:3\n"[..]),
("a png", b"\x89PNG\r\n\x1a\n"),
("a script", b"#!/bin/sh\nid\n"),
(
"a concat list",
b"ffconcat version 1.0\nfile '/etc/passwd'\n",
),
("a wave-less riff (avi)", b"RIFF\x24\0\0\0AVI LIST"),
("empty", b""),
("one byte", &[0xFF]),
] {
assert!(!is_audio(head), "{name} was accepted");
}
}
}