Files
furumi-ng/furumi-agent/src/ingest/metadata.rs

130 lines
3.9 KiB
Rust

use std::path::Path;
use symphonia::core::{
codecs::CODEC_TYPE_NULL,
formats::FormatOptions,
io::MediaSourceStream,
meta::{MetadataOptions, StandardTagKey},
probe::Hint,
};
#[derive(Debug, Default)]
pub struct RawMetadata {
pub title: Option<String>,
pub artist: Option<String>,
pub album: Option<String>,
pub track_number: Option<u32>,
pub year: Option<u32>,
pub genre: Option<String>,
pub duration_secs: Option<f64>,
}
/// Extract metadata from an audio file using Symphonia.
/// Must be called from a blocking context (spawn_blocking).
pub fn extract(path: &Path) -> anyhow::Result<RawMetadata> {
let file = std::fs::File::open(path)?;
let mss = MediaSourceStream::new(Box::new(file), Default::default());
let mut hint = Hint::new();
if let Some(ext) = path.extension().and_then(|e| e.to_str()) {
hint.with_extension(ext);
}
let mut probed = symphonia::default::get_probe().format(
&hint,
mss,
&FormatOptions { enable_gapless: false, ..Default::default() },
&MetadataOptions::default(),
)?;
let mut meta = RawMetadata::default();
// Check metadata side-data (e.g., ID3 tags probed before format)
if let Some(rev) = probed.metadata.get().as_ref().and_then(|m| m.current()) {
extract_tags(rev.tags(), &mut meta);
}
// Also check format-embedded metadata
if let Some(rev) = probed.format.metadata().current() {
if meta.title.is_none() {
extract_tags(rev.tags(), &mut meta);
}
}
// Duration
meta.duration_secs = probed
.format
.tracks()
.iter()
.find(|t| t.codec_params.codec != CODEC_TYPE_NULL)
.and_then(|t| {
let n_frames = t.codec_params.n_frames?;
let tb = t.codec_params.time_base?;
Some(n_frames as f64 * tb.numer as f64 / tb.denom as f64)
});
Ok(meta)
}
fn extract_tags(tags: &[symphonia::core::meta::Tag], meta: &mut RawMetadata) {
for tag in tags {
let value = fix_encoding(tag.value.to_string());
if let Some(key) = tag.std_key {
match key {
StandardTagKey::TrackTitle => {
if meta.title.is_none() {
meta.title = Some(value);
}
}
StandardTagKey::Artist | StandardTagKey::Performer => {
if meta.artist.is_none() {
meta.artist = Some(value);
}
}
StandardTagKey::Album => {
if meta.album.is_none() {
meta.album = Some(value);
}
}
StandardTagKey::TrackNumber => {
if meta.track_number.is_none() {
meta.track_number = value.parse().ok();
}
}
StandardTagKey::Date | StandardTagKey::OriginalDate => {
if meta.year.is_none() {
meta.year = value[..4.min(value.len())].parse().ok();
}
}
StandardTagKey::Genre => {
if meta.genre.is_none() {
meta.genre = Some(value);
}
}
_ => {}
}
}
}
}
/// Heuristic to fix mojibake (CP1251 bytes interpreted as Latin-1/Windows-1252).
fn fix_encoding(s: String) -> String {
let bytes: Vec<u8> = s.chars().map(|c| c as u32).filter(|&c| c <= 255).map(|c| c as u8).collect();
if bytes.len() != s.chars().count() {
return s;
}
let has_mojibake = bytes.iter().any(|&b| b >= 0xC0);
if !has_mojibake {
return s;
}
let (decoded, _, errors) = encoding_rs::WINDOWS_1251.decode(&bytes);
if errors {
return s;
}
decoded.into_owned()
}