labels: universal MPLS fallback + bdmt disc metadata + png stub

Three new modules in the labels platform, all layered so framework-
specific parsers (paramount, criterion, pixelogic, ctrm, dbp, deluxe)
always take precedence over the universal layer.

**mpls_universal.rs** (~600 LOC, 9 tests): consumes the already-parsed
`crate::mpls::Playlist::streams` and emits StreamLabel entries with
language + codec_hint per stream. Returns `Confidence::Low` (new
variant) so framework parsers' Medium/High always win the registry's
max-by-confidence tiebreaker; MPLS only gets picked when no framework
matched. Closes the "no BD-J disc" case (HDMV-only navigation) that
previously produced zero labels — language and base codec are
spec-mandated in MPLS STN tables on every Blu-ray ever made.

**bdmt.rs** (~350 LOC, 10 tests): reads `/BDMV/META/DL/bdmt_<lang>.xml`
files into a new `DiscMetadata` struct (localized title names per
ISO 639-2 code, descriptions, optional box-set position). Runs
independently of the parser registry — disc-level metadata, not
per-stream, so the registry's confidence selection doesn't apply.
Surfaced on a new `LabelAnalysis::disc_metadata` field.

**png_filenames.rs** (noop stub): pattern documentation + dead-code
detect/parse for future reactivation. Deferred because MPLS already
delivers per-stream lang/codec/type on every disc; PNG filename
language tokens only add studio variant disambiguation (FRC vs FRP,
LAS vs CSP) — niche enough to not justify the implementation cost
right now.

Wiring changes in `mod.rs`:
- New `Confidence::Low` variant (PartialOrd places it below Medium/High)
- New `ParseResult::low()` constructor
- `mpls_universal` appended last to `PARSERS` registry
- `LabelAnalysis::disc_metadata: Option<DiscMetadata>` field
- `analyze()` runs `bdmt::parse` independently and surfaces result
- `pub use bdmt::DiscMetadata` re-export so the labels-analyze tool
  in freemkv-tools can construct the JSON payload

Total: 151 of 151 labels tests passing (was 132 — added 19 new).

src/labels/bdmt.rs

@@ -0,0 +1,350 @@
+//! BDMV disc-library metadata (`/BDMV/META/DL/bdmt_<lang>.xml`).
+//!
+//! Every commercial Blu-ray carries a disc-library metadata directory
+//! with one XML file per shipped language. The schema is the Blu-ray
+//! "disc library metadata" namespace (`urn:BDA:bdmv;disclibmeta`),
+//! conventionally prefixed `di:`. Fields commonly present:
+//!
+//! - `<di:title>` or `<di:name>` — the title string. Vendor practice
+//!   varies (Paramount discs tend to use `<di:name>`).
+//! - `<di:description>` — optional synopsis (often absent on retail
+//!   discs; common on box sets and special editions).
+//! - `<di:discNumber>` / `<di:numSets>` (or `<di:numberOfSets>`) —
+//!   set position for multi-disc releases.
+//!
+//! This module is intentionally separate from the BD-J `StreamLabel`
+//! parsers under `labels/*.rs`. The XML here is disc-level (title,
+//! description, set position), not per-stream — wiring into the main
+//! parser registry happens elsewhere.
+//!
+//! Real-world XML is irregular: missing description elements, multiple
+//! title elements (first one wins), and occasional malformed content.
+//! Extraction is best-effort — a malformed file is treated as "no
+//! metadata" (returns `None` from the helper), and the caller can
+//! still get metadata from sibling-language XML files.
+
+// The module wiring (registry hook + public re-export) is added
+// separately. Until then the parse/detect entry points have no
+use super::xml;
+use crate::sector::SectorReader;
+use crate::udf::UdfFs;
+use std::collections::BTreeMap;
+
+/// Disc-level metadata extracted from `/BDMV/META/DL/bdmt_*.xml`.
+///
+/// All maps are keyed by 3-char ISO 639-2 language code (e.g.
+/// `"eng"`, `"fra"`, `"jpn"`) — the same key segment used in the
+/// `bdmt_<lang>.xml` filename.
+#[derive(Debug, Clone, serde::Serialize, serde::Deserialize, Default)]
+pub struct DiscMetadata {
+    /// Localized titles, keyed by 3-char ISO 639-2 lang code
+    /// (e.g. "eng" → "Dune Part Two")
+    pub titles: BTreeMap<String, String>,
+    /// First-line / short description, per lang
+    pub descriptions: BTreeMap<String, String>,
+    /// Disc N of M for box sets (None if not a box set)
+    pub disc_number: Option<(u32, u32)>,
+}
+
+/// True if `/BDMV/META/DL/` exists and contains at least one
+/// `bdmt_*.xml` file.
+pub fn detect(udf: &UdfFs) -> bool {
+    let Some(dir) = udf.find_dir("/BDMV/META/DL") else {
+        return false;
+    };
+    dir.entries
+        .iter()
+        .any(|e| !e.is_dir && is_bdmt_filename(&e.name))
+}
+
+/// Read every `bdmt_<lang>.xml` under `/BDMV/META/DL/` and return the
+/// aggregated [`DiscMetadata`]. Returns `None` if no titles could be
+/// extracted from any file.
+pub fn parse(reader: &mut dyn SectorReader, udf: &UdfFs) -> Option<DiscMetadata> {
+    let dir = udf.find_dir("/BDMV/META/DL")?;
+    let mut out = DiscMetadata::default();
+
+    for entry in &dir.entries {
+        if entry.is_dir {
+            continue;
+        }
+        let Some(lang) = lang_code_from_filename(&entry.name) else {
+            continue;
+        };
+        let path = format!("/BDMV/META/DL/{}", entry.name);
+        let Ok(bytes) = udf.read_file(reader, &path) else {
+            continue;
+        };
+        let Ok(text) = std::str::from_utf8(&bytes) else {
+            continue;
+        };
+        let Some((title, description, disc_set)) = parse_bdmt_xml(&lang, text) else {
+            continue;
+        };
+        out.titles.insert(lang.clone(), title);
+        if let Some(desc) = description {
+            out.descriptions.insert(lang.clone(), desc);
+        }
+        // Disc-set position is disc-global; first one we successfully
+        // read wins. (All bdmt_*.xml on a given disc carry the same
+        // value in practice.)
+        if out.disc_number.is_none() {
+            if let Some(ds) = disc_set {
+                out.disc_number = Some(ds);
+            }
+        }
+    }
+
+    if out.titles.is_empty() {
+        None
+    } else {
+        Some(out)
+    }
+}
+
+/// True if `name` matches the `bdmt_<lang>.xml` convention with a
+/// 3-character ISO 639-2 lang code segment. Case-insensitive.
+fn is_bdmt_filename(name: &str) -> bool {
+    lang_code_from_filename(name).is_some()
+}
+
+/// Extract the 3-char language code from a `bdmt_<lang>.xml` filename.
+/// Returns `None` if the filename doesn't match. Lang code is
+/// lowercased so callers always see e.g. `"eng"` not `"ENG"`.
+fn lang_code_from_filename(name: &str) -> Option<String> {
+    let lower = name.to_ascii_lowercase();
+    let stem = lower.strip_suffix(".xml")?;
+    let lang = stem.strip_prefix("bdmt_")?;
+    // ISO 639-2 codes are exactly 3 ASCII letters. Be strict — keeps
+    // us from picking up unrelated `bdmt_foo.xml` siblings.
+    if lang.len() != 3 || !lang.chars().all(|c| c.is_ascii_alphabetic()) {
+        return None;
+    }
+    Some(lang.to_string())
+}
+
+/// Tuple returned by [`parse_bdmt_xml`]: `(title, description?, disc_set?)`.
+/// Aliased so the function signature isn't a clippy::type-complexity offender.
+pub(crate) type BdmtFields = (String, Option<String>, Option<(u32, u32)>);
+
+/// Parse one `bdmt_<lang>.xml` document and return
+/// `(title, description?, disc_set?)`. Returns `None` if no title
+/// could be located — the caller treats this as "skip this file".
+///
+/// Title-element preference: `<di:name>` → `<di:title>` →
+/// `<di:tableOfContents>/<di:titleName>` (first match wins, per the
+/// authoring-tool conventions documented at the module level).
+pub(crate) fn parse_bdmt_xml(_lang_code: &str, xml_text: &str) -> Option<BdmtFields> {
+    let title = extract_title(xml_text)?;
+    let description = xml::text(xml_text, "description").filter(|s| !s.is_empty());
+    let disc_set = extract_disc_set(xml_text);
+    Some((title, description, disc_set))
+}
+
+/// Try title-bearing element variants in priority order. The `xml`
+/// helpers are case- and namespace-insensitive, so callers pass the
+/// bare local name (no `di:` prefix).
+fn extract_title(xml_text: &str) -> Option<String> {
+    // Order matches the module-level convention: <di:name> first
+    // (Paramount-style), then <di:title>, then the nested
+    // tableOfContents/titleName form.
+    for tag in ["name", "title"] {
+        if let Some(s) = xml::text(xml_text, tag) {
+            let trimmed = s.trim();
+            if !trimmed.is_empty() {
+                return Some(trimmed.to_string());
+            }
+        }
+    }
+    // tableOfContents/titleName: search inside the toc block so we
+    // don't accidentally pick a stray <titleName> from elsewhere.
+    if let Some((s, e)) = xml::find_element(xml_text, "tableOfContents", 0) {
+        let block = &xml_text[s..e];
+        if let Some(t) = xml::text(block, "titleName") {
+            let trimmed = t.trim();
+            if !trimmed.is_empty() {
+                return Some(trimmed.to_string());
+            }
+        }
+    }
+    None
+}
+
+/// Extract `(discNumber, numSets)` if both are present and parse as
+/// `u32`. Accepts either `<di:numSets>` or `<di:numberOfSets>` for
+/// the denominator (both forms appear in the wild).
+fn extract_disc_set(xml_text: &str) -> Option<(u32, u32)> {
+    let n = xml::text(xml_text, "discNumber")?
+        .trim()
+        .parse::<u32>()
+        .ok()?;
+    let total = xml::text(xml_text, "numSets")
+        .or_else(|| xml::text(xml_text, "numberOfSets"))?
+        .trim()
+        .parse::<u32>()
+        .ok()?;
+    Some((n, total))
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn extract_simple_title() {
+        // Minimal Paramount-style document: <di:name> as the title
+        // carrier inside a <discInfo> root.
+        let xml = r#"<?xml version="1.0" encoding="UTF-8"?>
+<discInfo xmlns:di="urn:BDA:bdmv;disclibmeta">
+  <di:name>Dune Part Two</di:name>
+</discInfo>"#;
+        let (title, desc, set) = parse_bdmt_xml("eng", xml).expect("title should parse");
+        assert_eq!(title, "Dune Part Two");
+        assert_eq!(desc, None);
+        assert_eq!(set, None);
+    }
+
+    #[test]
+    fn extract_title_element_variant() {
+        // <di:title> is the alternate carrier; should be picked up
+        // when <di:name> is absent.
+        let xml = r#"<discInfo xmlns:di="urn:BDA:bdmv;disclibmeta">
+  <di:title>The Matrix</di:title>
+  <di:description>A film about computers.</di:description>
+</discInfo>"#;
+        let (title, desc, _) = parse_bdmt_xml("eng", xml).unwrap();
+        assert_eq!(title, "The Matrix");
+        assert_eq!(desc.as_deref(), Some("A film about computers."));
+    }
+
+    #[test]
+    fn extract_title_from_table_of_contents_fallback() {
+        // Some authoring tools nest the title under tableOfContents.
+        // No <di:name> or <di:title> at top level → fall back to
+        // titleName inside tableOfContents.
+        let xml = r#"<discInfo xmlns:di="urn:BDA:bdmv;disclibmeta">
+  <di:tableOfContents>
+    <di:titleName>Inside Out 2</di:titleName>
+  </di:tableOfContents>
+</discInfo>"#;
+        let (title, _, _) = parse_bdmt_xml("eng", xml).unwrap();
+        assert_eq!(title, "Inside Out 2");
+    }
+
+    #[test]
+    fn extract_box_set_position() {
+        let xml = r#"<discInfo xmlns:di="urn:BDA:bdmv;disclibmeta">
+  <di:name>LOTR Disc 2</di:name>
+  <di:discNumber>2</di:discNumber>
+  <di:numSets>5</di:numSets>
+</discInfo>"#;
+        let (_, _, set) = parse_bdmt_xml("eng", xml).unwrap();
+        assert_eq!(set, Some((2, 5)));
+    }
+
+    #[test]
+    fn extract_box_set_position_alternate_total_tag() {
+        // <di:numberOfSets> is an alternate spelling we've seen.
+        let xml = r#"<discInfo xmlns:di="urn:BDA:bdmv;disclibmeta">
+  <di:name>X</di:name>
+  <di:discNumber>3</di:discNumber>
+  <di:numberOfSets>6</di:numberOfSets>
+</discInfo>"#;
+        let (_, _, set) = parse_bdmt_xml("eng", xml).unwrap();
+        assert_eq!(set, Some((3, 6)));
+    }
+
+    #[test]
+    fn extract_box_set_requires_both_fields() {
+        // discNumber alone (no total) yields None — we don't fabricate
+        // a denominator.
+        let xml = r#"<discInfo xmlns:di="urn:BDA:bdmv;disclibmeta">
+  <di:name>X</di:name>
+  <di:discNumber>1</di:discNumber>
+</discInfo>"#;
+        let (_, _, set) = parse_bdmt_xml("eng", xml).unwrap();
+        assert_eq!(set, None);
+    }
+
+    #[test]
+    fn multiple_languages_keyed_correctly() {
+        // Simulate driving parse_bdmt_xml from two synthetic XML
+        // blobs and aggregating into DiscMetadata the same way parse()
+        // would. This exercises the BTreeMap key handling without
+        // needing a UdfFs.
+        let eng_xml = r#"<discInfo xmlns:di="urn:BDA:bdmv;disclibmeta">
+  <di:name>Dune Part Two</di:name>
+</discInfo>"#;
+        let fra_xml = r#"<discInfo xmlns:di="urn:BDA:bdmv;disclibmeta">
+  <di:name>Dune Deuxième Partie</di:name>
+  <di:description>Suite du film de 2021.</di:description>
+</discInfo>"#;
+
+        let mut meta = DiscMetadata::default();
+        for (lang, blob) in [("eng", eng_xml), ("fra", fra_xml)] {
+            let (title, desc, ds) = parse_bdmt_xml(lang, blob).unwrap();
+            meta.titles.insert(lang.to_string(), title);
+            if let Some(d) = desc {
+                meta.descriptions.insert(lang.to_string(), d);
+            }
+            if meta.disc_number.is_none() {
+                if let Some(d) = ds {
+                    meta.disc_number = Some(d);
+                }
+            }
+        }
+
+        assert_eq!(
+            meta.titles.get("eng").map(String::as_str),
+            Some("Dune Part Two")
+        );
+        assert_eq!(
+            meta.titles.get("fra").map(String::as_str),
+            Some("Dune Deuxième Partie")
+        );
+        assert!(meta.descriptions.get("eng").is_none());
+        assert_eq!(
+            meta.descriptions.get("fra").map(String::as_str),
+            Some("Suite du film de 2021.")
+        );
+        assert_eq!(meta.disc_number, None);
+    }
+
+    #[test]
+    fn malformed_xml_returns_none() {
+        // Random gibberish has no recognizable title element. We
+        // document the contract: parse_bdmt_xml returns None, and
+        // parse() (the caller) skips the file. Aggregating across
+        // zero files leaves DiscMetadata::default() — which parse()
+        // surfaces as None to its caller. Either is documented as
+        // acceptable per the module spec.
+        let bad = "this is not xml &&& <<< nope";
+        assert!(parse_bdmt_xml("eng", bad).is_none());
+
+        // Half-open tag, no body, no close: also yields no title.
+        let truncated = "<discInfo><di:name>";
+        assert!(parse_bdmt_xml("eng", truncated).is_none());
+    }
+
+    #[test]
+    fn whitespace_in_title_is_trimmed() {
+        let xml = r#"<discInfo><di:name>
+            Dune Part Two
+        </di:name></discInfo>"#;
+        let (title, _, _) = parse_bdmt_xml("eng", xml).unwrap();
+        assert_eq!(title, "Dune Part Two");
+    }
+
+    #[test]
+    fn lang_code_extraction() {
+        assert_eq!(lang_code_from_filename("bdmt_eng.xml"), Some("eng".into()));
+        assert_eq!(lang_code_from_filename("BDMT_FRA.XML"), Some("fra".into()));
+        assert_eq!(lang_code_from_filename("bdmt_jpn.xml"), Some("jpn".into()));
+        // Non-matching cases:
+        assert_eq!(lang_code_from_filename("bdmt_.xml"), None);
+        assert_eq!(lang_code_from_filename("bdmt_engl.xml"), None);
+        assert_eq!(lang_code_from_filename("bdmt_e1g.xml"), None);
+        assert_eq!(lang_code_from_filename("bdmt_eng.txt"), None);
+        assert_eq!(lang_code_from_filename("foo.xml"), None);
+    }
+}

src/labels/mod.rs

@@ -7,14 +7,17 @@
 //!   3. Implement `pub fn parse(reader: &mut dyn SectorReader, udf: &UdfFs) -> Option<Vec<StreamLabel>>`
 //!   4. Add `mod myformat;` below and one line to `PARSERS` array
 
+mod bdmt;
 pub(crate) mod class_reader;
 mod criterion;
 mod ctrm;
 mod dbp;
 mod deluxe;
 pub(crate) mod jar;
+mod mpls_universal;
 mod paramount;
 mod pixelogic;
+mod png_filenames;
 pub(crate) mod text;
 pub mod vocab;
 pub(crate) mod xml;
@@ -23,6 +26,11 @@ use crate::disc::{DiscTitle, Stream};
 use crate::sector::SectorReader;
 use crate::udf::UdfFs;
 
+// Re-export bdmt's public type so callers can construct/inspect
+// disc-level metadata via `labels::DiscMetadata`. The module itself
+// stays private — analyze() drives the parse path.
+pub use bdmt::DiscMetadata;
+
 // Re-exported via crate::disc — the public API surfaces these next to
 // AudioStream/SubtitleStream so callers can map purpose/qualifier to display
 // text in their own locale.
@@ -88,10 +96,14 @@ type ParseFn = fn(&mut dyn SectorReader, &UdfFs) -> Option<ParseResult>;
 /// A parser SHOULD return `High` only when its full schema was
 /// extracted with no fallback or guessing. `Medium` is for matched-
 /// but-degraded outputs (some streams missing fields, fingerprint
-/// matched but a sub-table couldn't be decoded, etc.). The registry
+/// matched but a sub-table couldn't be decoded, etc.). `Low` is for
-/// prefers `High` over `Medium`; ties fall to array order.
+/// the universal MPLS fallback — spec-mandated stream metadata
+/// (language + base codec) that's correct but lacks editorial labels
+/// (commentary, SDH, etc.). The registry prefers `High > Medium > Low`;
+/// ties fall to array order.
 #[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord)]
 pub enum Confidence {
+    Low,
     Medium,
     High,
 }
@@ -124,6 +136,16 @@ impl ParseResult {
             confidence: Confidence::Medium,
         }
     }
+
+    /// Convenience for the universal MPLS fallback: spec-derived
+    /// stream language + codec, but no editorial labels (commentary,
+    /// SDH, etc.). Framework parsers always win over `low`.
+    pub fn low(labels: Vec<StreamLabel>) -> Self {
+        ParseResult {
+            labels,
+            confidence: Confidence::Low,
+        }
+    }
 }
 
 const PARSERS: &[(&str, DetectFn, ParseFn)] = &[
@@ -138,6 +160,17 @@ const PARSERS: &[(&str, DetectFn, ParseFn)] = &[
     // (constant-pool iteration vs. deluxe's bytecode walking).
     ("dbp", dbp::detect, dbp::parse),
     ("deluxe", deluxe::detect, deluxe::parse),
+    // Universal MPLS fallback. Returns Confidence::Low so framework
+    // parsers always win when they match. Closes the "no framework
+    // matched" gap (e.g. HDMV-only discs) with spec-derived language
+    // + base codec for every stream the playlist references. Runs
+    // last in registry order so it's only the chosen parser when
+    // nothing else fired.
+    (
+        "mpls_universal",
+        mpls_universal::detect,
+        mpls_universal::parse,
+    ),
 ];
 
 /// Search disc for config files, extract labels, apply to streams.
@@ -426,12 +459,23 @@ pub fn analyze(reader: &mut dyn SectorReader, udf: &UdfFs) -> LabelAnalysis {
         );
     }
 
+    // bdmt runs independently of the parser registry: it's disc-level
+    // metadata (localized titles, box-set position), not per-stream
+    // labels, so the "highest confidence wins" logic doesn't apply.
+    // Always run if detected; surface result as a separate field.
+    let disc_metadata = if bdmt::detect(udf) {
+        bdmt::parse(reader, udf)
+    } else {
+        None
+    };
+
     LabelAnalysis {
         parser,
         parsers_detected,
         confidence,
         jar_inventory: inventory,
         labels,
+        disc_metadata,
     }
 }
 
@@ -461,6 +505,11 @@ pub struct LabelAnalysis {
     /// Raw labels emitted by the selected parser (empty if `parser`
     /// is `None`).
     pub labels: Vec<StreamLabel>,
+    /// Disc-level metadata from `/BDMV/META/DL/bdmt_*.xml` if present.
+    /// Localized title names, descriptions, box-set position. Orthogonal
+    /// to per-stream labels; populated independently from the parser
+    /// registry.
+    pub disc_metadata: Option<bdmt::DiscMetadata>,
 }
 
 /// List filenames found under any `/BDMV/JAR/<x>/` subdirectory of
@@ -544,12 +593,14 @@ mod registry_tests {
                 "pixelogic",
                 "ctrm",
                 "dbp",
-                "deluxe"
+                "deluxe",
+                "mpls_universal",
             ],
-            "PARSERS array order changed — confirm dbp + deluxe stay last \
+            "PARSERS array order changed — confirm dbp + deluxe stay just \
-             (loose detect, real check in parse), and stricter parsers \
+             before mpls_universal (loose detect, real check in parse), \
-             (paramount/criterion/pixelogic/ctrm — all file-presence \
+             stricter parsers (paramount/criterion/pixelogic/ctrm — all \
-             gated detect) stay first."
+             file-presence gated detect) stay first, and mpls_universal \
+             stays LAST as the universal Low-confidence fallback."
         );
     }
 

src/labels/mpls_universal.rs

@@ -0,0 +1,595 @@
+//! Universal MPLS-based stream labels.
+//!
+//! Unlike the framework-specific parsers in this directory (dbp,
+//! pixelogic, ctrm, criterion, ...), this module is the *floor*:
+//! every Blu-ray ships with MPLS playlists under `/BDMV/PLAYLIST/`,
+//! and every MPLS file has an STN table with per-stream ISO 639-2
+//! language codes plus coding-type / channel-layout / sample-rate
+//! bytes from the BD spec.
+//!
+//! The framework parsers extract richer editorial labels ("English
+//! Dolby Atmos", "Director's Commentary") when the disc was authored
+//! with a recognized tool. When none of them match (e.g. a "no BD-J"
+//! disc, or an authoring framework we haven't catalogued), MPLS still
+//! gives us language + codec on every stream — enough to render
+//! something more useful than the bare PID.
+//!
+//! Output confidence is Low: MPLS carries language + codec but
+//! never purpose/qualifier info (no way to tell "Commentary" from
+//! "Normal" from the STN table alone). Higher-confidence framework
+//! parsers, when present, always win on the registry's max-by-confidence
+//! tiebreaker — MPLS is only chosen when nothing else matched.
+
+use super::{
+    LabelPurpose, LabelQualifier, ParseResult, StreamLabel, StreamLabelType,
+    vocab::{self, LangInfo},
+};
+use crate::sector::SectorReader;
+use crate::udf::UdfFs;
+
+/// True iff `/BDMV/PLAYLIST/` exists and contains at least one
+/// `.mpls` file. Cheap directory walk only — no sector reads.
+pub fn detect(udf: &UdfFs) -> bool {
+    let Some(dir) = udf.find_dir("/BDMV/PLAYLIST") else {
+        return false;
+    };
+    dir.entries
+        .iter()
+        .any(|e| !e.is_dir && has_mpls_extension(&e.name))
+}
+
+/// Walk every `*.mpls` in `/BDMV/PLAYLIST/`, parse it, and convert
+/// each StreamEntry to a [`StreamLabel`]. Streams shared across
+/// playlists (same PID) are deduped.
+///
+/// Returns `None` if no labels could be produced (e.g. no .mpls files
+/// parsed successfully, or every parsed stream was a type we skip
+/// like IG / DV EL).
+pub fn parse(reader: &mut dyn SectorReader, udf: &UdfFs) -> Option<ParseResult> {
+    let playlist_dir = udf.find_dir("/BDMV/PLAYLIST")?;
+
+    // Collect mpls filenames first so we don't hold a borrow on udf
+    // while we call udf.read_file (which takes &self).
+    let mpls_names: Vec<String> = playlist_dir
+        .entries
+        .iter()
+        .filter(|e| !e.is_dir && has_mpls_extension(&e.name))
+        .map(|e| e.name.clone())
+        .collect();
+
+    if mpls_names.is_empty() {
+        return None;
+    }
+
+    let mut labels: Vec<StreamLabel> = Vec::new();
+    // (stream_type_tag, language, codec_hint, pid) — PID is the
+    // canonical "same physical stream" key; type+lang+codec round
+    // out the rare case where two distinct logical streams happen
+    // to share a PID across playlists with different metadata.
+    let mut seen: Vec<(u8, String, String, u16)> = Vec::new();
+
+    for name in &mpls_names {
+        let path = format!("/BDMV/PLAYLIST/{}", name);
+        let Ok(data) = udf.read_file(reader, &path) else {
+            continue;
+        };
+        let Ok(playlist) = crate::mpls::parse(&data) else {
+            continue;
+        };
+
+        // Per-MPLS-file 1-based counters keyed by StreamLabelType.
+        // The dedup pass below removes duplicates across files; the
+        // numbering of the *surviving* entries comes from whichever
+        // playlist contributed each PID first.
+        let mut audio_idx: u16 = 0;
+        let mut sub_idx: u16 = 0;
+
+        for entry in &playlist.streams {
+            let label_type = match entry.stream_type {
+                2 | 5 => StreamLabelType::Audio, // primary + secondary audio
+                3 => StreamLabelType::Subtitle,  // PG subtitle
+                // 1 = primary video, 6 = secondary video, 7 = DV EL
+                //     → no StreamLabelType variant for video, skip.
+                // 4 = IG (interactive graphics) — not a user-facing
+                //     stream, skip.
+                _ => continue,
+            };
+
+            let stream_number = match label_type {
+                StreamLabelType::Audio => {
+                    audio_idx += 1;
+                    audio_idx
+                }
+                StreamLabelType::Subtitle => {
+                    sub_idx += 1;
+                    sub_idx
+                }
+            };
+
+            let language = normalize_language(&entry.language);
+            let name = language_display_name(&language);
+            let codec_hint = build_codec_hint(label_type, entry);
+
+            let type_tag = type_tag(label_type);
+            let key = (type_tag, language.clone(), codec_hint.clone(), entry.pid);
+            if seen.contains(&key) {
+                continue;
+            }
+            seen.push(key);
+
+            labels.push(StreamLabel {
+                stream_number,
+                stream_type: label_type,
+                language,
+                name,
+                purpose: LabelPurpose::Normal,
+                qualifier: LabelQualifier::None,
+                codec_hint,
+                variant: String::new(),
+            });
+        }
+    }
+
+    if labels.is_empty() {
+        return None;
+    }
+
+    // MPLS gives language + codec but never editorial info (no
+    // commentary/SDH/director's cut). Low confidence means framework
+    // parsers (paramount, criterion, pixelogic, ctrm, dbp, deluxe) always
+    // win when they match. MPLS only gets chosen as the parser when
+    // nothing else fired — exactly the universal-fallback role we want.
+    Some(ParseResult::low(labels))
+}
+
+fn has_mpls_extension(name: &str) -> bool {
+    // Case-insensitive ".mpls" suffix. Some discs use uppercase,
+    // some lowercase; UDF filenames preserve case but we don't.
+    let n = name.len();
+    if n < 5 {
+        return false;
+    }
+    name[n - 5..].eq_ignore_ascii_case(".mpls")
+}
+
+/// Lowercase + trim the raw 3-char ISO 639-2 code. If the lowered
+/// string maps via [`vocab::lang`] (it won't for plain "eng" — that
+/// matcher is for English-name fragments, not codes) use its
+/// canonical code; otherwise return the trimmed lowercase string.
+fn normalize_language(raw: &str) -> String {
+    let trimmed = raw.trim().to_ascii_lowercase();
+    if trimmed.is_empty() {
+        return String::new();
+    }
+    // vocab::lang() matches free-form English names, not ISO 639-2
+    // codes — so for the typical MPLS payload ("eng", "fra", ...)
+    // it returns None and we keep the trimmed code.
+    if let Some(LangInfo { code, .. }) = vocab::lang(&trimmed) {
+        return code.to_string();
+    }
+    trimmed
+}
+
+/// Human-readable English name for an ISO 639-2 code, or empty if
+/// the code is unknown. Kept inline rather than in vocab because
+/// vocab is the *reverse* mapping (name → code).
+fn language_display_name(iso: &str) -> String {
+    match iso {
+        "eng" => "English",
+        "fra" | "fre" => "French",
+        "spa" => "Spanish",
+        "deu" | "ger" => "German",
+        "ita" => "Italian",
+        "jpn" => "Japanese",
+        "zho" | "chi" => "Chinese",
+        "kor" => "Korean",
+        "por" => "Portuguese",
+        "pol" => "Polish",
+        "ces" | "cze" => "Czech",
+        "hun" => "Hungarian",
+        "nld" | "dut" => "Dutch",
+        "ara" => "Arabic",
+        "hin" => "Hindi",
+        "tur" => "Turkish",
+        "tha" => "Thai",
+        "swe" => "Swedish",
+        "nor" => "Norwegian",
+        "dan" => "Danish",
+        "fin" => "Finnish",
+        "heb" => "Hebrew",
+        "rus" => "Russian",
+        "ell" | "gre" => "Greek",
+        "vie" => "Vietnamese",
+        "ind" => "Indonesian",
+        "msa" | "may" => "Malay",
+        "ukr" => "Ukrainian",
+        "ron" | "rum" => "Romanian",
+        "bul" => "Bulgarian",
+        "hrv" => "Croatian",
+        "srp" => "Serbian",
+        "slk" | "slo" => "Slovak",
+        "slv" => "Slovenian",
+        "est" => "Estonian",
+        "lav" => "Latvian",
+        "lit" => "Lithuanian",
+        "isl" | "ice" => "Icelandic",
+        "eus" | "baq" => "Basque",
+        "cat" => "Catalan",
+        "glg" => "Galician",
+        _ => "",
+    }
+    .to_string()
+}
+
+/// Map BD coding_type byte → codec name. Returns empty for unknown
+/// bytes (the table covers everything the spec defines, but unknown
+/// values are still possible on malformed discs).
+fn codec_name(coding_type: u8) -> &'static str {
+    match coding_type {
+        0x02 => "MPEG-2",
+        0x1B => "H.264",
+        0x24 => "HEVC",
+        0x80 => "LPCM",
+        0x81 => "AC-3",
+        0x82 => "DTS",
+        0x83 => "TrueHD",
+        0x84 => "AC-3+",
+        0x85 => "DTS-HD",
+        0x86 => "DTS-HD MA",
+        0x90 => "PG",
+        0x91 => "IG",
+        0xA1 => "AC-3+ Secondary",
+        0xA2 => "DTS-HD Secondary",
+        _ => "",
+    }
+}
+
+/// Build the final `codec_hint`. For audio streams, optionally
+/// append " <channels>" and/or " <rate>" suffixes. Sample rate is
+/// only spelled out for non-48k (the universal default).
+fn build_codec_hint(label_type: StreamLabelType, entry: &crate::mpls::StreamEntry) -> String {
+    let base = codec_name(entry.coding_type);
+    if base.is_empty() {
+        return String::new();
+    }
+
+    if label_type != StreamLabelType::Audio {
+        return base.to_string();
+    }
+
+    let mut out = base.to_string();
+
+    let channels = match entry.audio_format {
+        1 => Some("mono"),
+        3 => Some("2.0"),
+        6 => Some("5.1"),
+        12 => Some("7.1"),
+        _ => None,
+    };
+    if let Some(ch) = channels {
+        out.push(' ');
+        out.push_str(ch);
+    }
+
+    // 1 = 48 kHz (universal default, omit). Only call out higher rates.
+    let rate = match entry.audio_rate {
+        4 => Some("96kHz"),
+        5 => Some("192kHz"),
+        _ => None,
+    };
+    if let Some(r) = rate {
+        out.push(' ');
+        out.push_str(r);
+    }
+
+    out
+}
+
+/// Dedup-tag for the label type. `u8` instead of `StreamLabelType`
+/// itself because the enum does not derive `Hash` / `Eq`-by-discriminant
+/// in a way that we want to couple to (and `==` works fine for the
+/// linear `Vec::contains` lookup we do).
+fn type_tag(t: StreamLabelType) -> u8 {
+    match t {
+        StreamLabelType::Audio => 1,
+        StreamLabelType::Subtitle => 2,
+    }
+}
+
+// ── Tests ────────────────────────────────────────────────────────────
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::mpls::{Playlist, StreamEntry};
+
+    fn audio_entry(pid: u16, coding: u8, fmt: u8, rate: u8, lang: &str) -> StreamEntry {
+        StreamEntry {
+            stream_type: 2,
+            pid,
+            coding_type: coding,
+            video_format: 0,
+            video_rate: 0,
+            audio_format: fmt,
+            audio_rate: rate,
+            language: lang.to_string(),
+            dynamic_range: 0,
+            color_space: 0,
+            secondary: false,
+        }
+    }
+
+    fn pg_entry(pid: u16, lang: &str) -> StreamEntry {
+        StreamEntry {
+            stream_type: 3,
+            pid,
+            coding_type: 0x90,
+            video_format: 0,
+            video_rate: 0,
+            audio_format: 0,
+            audio_rate: 0,
+            language: lang.to_string(),
+            dynamic_range: 0,
+            color_space: 0,
+            secondary: false,
+        }
+    }
+
+    fn playlist_with(streams: Vec<StreamEntry>) -> Playlist {
+        Playlist {
+            version: "0200".to_string(),
+            play_items: Vec::new(),
+            streams,
+            marks: Vec::new(),
+        }
+    }
+
+    /// Drive the same conversion logic that `parse()` runs on real
+    /// disc data, but starting from already-parsed Playlists so we
+    /// don't have to synthesize valid MPLS bytes.
+    fn labels_from_playlists(playlists: &[Playlist]) -> Vec<StreamLabel> {
+        let mut labels: Vec<StreamLabel> = Vec::new();
+        let mut seen: Vec<(u8, String, String, u16)> = Vec::new();
+
+        for playlist in playlists {
+            let mut audio_idx: u16 = 0;
+            let mut sub_idx: u16 = 0;
+
+            for entry in &playlist.streams {
+                let label_type = match entry.stream_type {
+                    2 | 5 => StreamLabelType::Audio,
+                    3 => StreamLabelType::Subtitle,
+                    _ => continue,
+                };
+                let stream_number = match label_type {
+                    StreamLabelType::Audio => {
+                        audio_idx += 1;
+                        audio_idx
+                    }
+                    StreamLabelType::Subtitle => {
+                        sub_idx += 1;
+                        sub_idx
+                    }
+                };
+                let language = normalize_language(&entry.language);
+                let name = language_display_name(&language);
+                let codec_hint = build_codec_hint(label_type, entry);
+                let key = (
+                    type_tag(label_type),
+                    language.clone(),
+                    codec_hint.clone(),
+                    entry.pid,
+                );
+                if seen.contains(&key) {
+                    continue;
+                }
+                seen.push(key);
+                labels.push(StreamLabel {
+                    stream_number,
+                    stream_type: label_type,
+                    language,
+                    name,
+                    purpose: LabelPurpose::Normal,
+                    qualifier: LabelQualifier::None,
+                    codec_hint,
+                    variant: String::new(),
+                });
+            }
+        }
+        labels
+    }
+
+    #[test]
+    fn mpls_audio_streams_become_labels() {
+        // Two audio streams: English TrueHD 7.1 48k, French AC-3 5.1 48k.
+        let pl = playlist_with(vec![
+            audio_entry(0x1100, 0x83, 12, 1, "eng"),
+            audio_entry(0x1101, 0x81, 6, 1, "fra"),
+        ]);
+        let labels = labels_from_playlists(&[pl]);
+        assert_eq!(labels.len(), 2);
+
+        // English TrueHD 7.1
+        let a = &labels[0];
+        assert_eq!(a.stream_type, StreamLabelType::Audio);
+        assert_eq!(a.stream_number, 1);
+        assert_eq!(a.language, "eng");
+        assert_eq!(a.name, "English");
+        assert_eq!(a.codec_hint, "TrueHD 7.1");
+        assert_eq!(a.purpose, LabelPurpose::Normal);
+        assert_eq!(a.qualifier, LabelQualifier::None);
+        assert_eq!(a.variant, "");
+
+        // French AC-3 5.1
+        let b = &labels[1];
+        assert_eq!(b.stream_type, StreamLabelType::Audio);
+        assert_eq!(b.stream_number, 2);
+        assert_eq!(b.language, "fra");
+        assert_eq!(b.name, "French");
+        assert_eq!(b.codec_hint, "AC-3 5.1");
+    }
+
+    #[test]
+    fn mpls_pg_streams_become_subtitle_labels() {
+        let pl = playlist_with(vec![
+            pg_entry(0x1200, "eng"),
+            pg_entry(0x1201, "spa"),
+            pg_entry(0x1202, "fra"),
+        ]);
+        let labels = labels_from_playlists(&[pl]);
+        assert_eq!(labels.len(), 3);
+        for label in &labels {
+            assert_eq!(label.stream_type, StreamLabelType::Subtitle);
+            assert_eq!(label.codec_hint, "PG");
+        }
+        assert_eq!(labels[0].stream_number, 1);
+        assert_eq!(labels[0].language, "eng");
+        assert_eq!(labels[0].name, "English");
+        assert_eq!(labels[1].stream_number, 2);
+        assert_eq!(labels[1].language, "spa");
+        assert_eq!(labels[1].name, "Spanish");
+        assert_eq!(labels[2].stream_number, 3);
+        assert_eq!(labels[2].language, "fra");
+    }
+
+    #[test]
+    fn dedup_streams_across_playlists() {
+        // Two playlists, same English TrueHD 7.1 PID 0x1100 in both.
+        // Expect one Audio label, not two.
+        let pl1 = playlist_with(vec![
+            audio_entry(0x1100, 0x83, 12, 1, "eng"),
+            audio_entry(0x1101, 0x81, 6, 1, "fra"),
+        ]);
+        let pl2 = playlist_with(vec![
+            audio_entry(0x1100, 0x83, 12, 1, "eng"), // duplicate
+            audio_entry(0x1102, 0x82, 6, 1, "deu"),  // new
+        ]);
+        let labels = labels_from_playlists(&[pl1, pl2]);
+        // Expected: eng@0x1100, fra@0x1101, deu@0x1102 — three uniques.
+        assert_eq!(labels.len(), 3);
+        // PID isn't stored on StreamLabel, so assert on the surviving
+        // language set instead.
+        let mut langs: Vec<String> = labels.iter().map(|l| l.language.clone()).collect();
+        langs.sort();
+        assert_eq!(langs, vec!["deu", "eng", "fra"]);
+    }
+
+    #[test]
+    fn coding_type_to_codec_hint_table() {
+        // Spot-check every entry in the spec table. Audio entries
+        // come back bare (no channels/rate set) so codec_hint is the
+        // codec name alone.
+        let cases: &[(u8, &str)] = &[
+            (0x02, "MPEG-2"),
+            (0x1B, "H.264"),
+            (0x24, "HEVC"),
+            (0x80, "LPCM"),
+            (0x81, "AC-3"),
+            (0x82, "DTS"),
+            (0x83, "TrueHD"),
+            (0x84, "AC-3+"),
+            (0x85, "DTS-HD"),
+            (0x86, "DTS-HD MA"),
+            (0x90, "PG"),
+            (0x91, "IG"),
+            (0xA1, "AC-3+ Secondary"),
+            (0xA2, "DTS-HD Secondary"),
+        ];
+        for (ct, expected) in cases {
+            assert_eq!(
+                codec_name(*ct),
+                *expected,
+                "coding_type 0x{:02X} should map to {}",
+                ct,
+                expected
+            );
+        }
+        // Unknown bytes return empty.
+        assert_eq!(codec_name(0x00), "");
+        assert_eq!(codec_name(0xFF), "");
+    }
+
+    #[test]
+    fn audio_format_appends_channel_layout() {
+        let mono = audio_entry(1, 0x83, 1, 1, "eng");
+        let stereo = audio_entry(2, 0x83, 3, 1, "eng");
+        let surround_51 = audio_entry(3, 0x83, 6, 1, "eng");
+        let surround_71 = audio_entry(4, 0x83, 12, 1, "eng");
+        let unknown = audio_entry(5, 0x83, 0, 1, "eng");
+        assert_eq!(
+            build_codec_hint(StreamLabelType::Audio, &mono),
+            "TrueHD mono"
+        );
+        assert_eq!(
+            build_codec_hint(StreamLabelType::Audio, &stereo),
+            "TrueHD 2.0"
+        );
+        assert_eq!(
+            build_codec_hint(StreamLabelType::Audio, &surround_51),
+            "TrueHD 5.1"
+        );
+        assert_eq!(
+            build_codec_hint(StreamLabelType::Audio, &surround_71),
+            "TrueHD 7.1"
+        );
+        assert_eq!(build_codec_hint(StreamLabelType::Audio, &unknown), "TrueHD");
+    }
+
+    #[test]
+    fn audio_rate_only_shows_above_48k() {
+        // 48 kHz (rate=1) is the universal default → not surfaced.
+        let r48 = audio_entry(1, 0x83, 6, 1, "eng");
+        // 96 kHz (rate=4) → surfaced.
+        let r96 = audio_entry(2, 0x83, 6, 4, "eng");
+        // 192 kHz (rate=5) → surfaced.
+        let r192 = audio_entry(3, 0x83, 6, 5, "eng");
+        assert_eq!(build_codec_hint(StreamLabelType::Audio, &r48), "TrueHD 5.1");
+        assert_eq!(
+            build_codec_hint(StreamLabelType::Audio, &r96),
+            "TrueHD 5.1 96kHz"
+        );
+        assert_eq!(
+            build_codec_hint(StreamLabelType::Audio, &r192),
+            "TrueHD 5.1 192kHz"
+        );
+    }
+
+    #[test]
+    fn unknown_iso_code_passes_through_without_display_name() {
+        // Made-up code: keep the raw lowercase code as `language`,
+        // but `name` is empty because we don't know it.
+        let pl = playlist_with(vec![audio_entry(0x1100, 0x83, 6, 1, "xyz")]);
+        let labels = labels_from_playlists(&[pl]);
+        assert_eq!(labels.len(), 1);
+        assert_eq!(labels[0].language, "xyz");
+        assert_eq!(labels[0].name, "");
+    }
+
+    #[test]
+    fn ig_and_dv_streams_are_skipped() {
+        // stream_type 4 = IG, 7 = DV EL — both must not surface.
+        let mut ig = pg_entry(0x1400, "eng");
+        ig.stream_type = 4;
+        let mut dv = audio_entry(0x1011, 0x24, 0, 0, "");
+        dv.stream_type = 7;
+        let pl = playlist_with(vec![ig, dv]);
+        let labels = labels_from_playlists(&[pl]);
+        assert!(labels.is_empty());
+    }
+
+    #[test]
+    fn secondary_audio_becomes_audio_label() {
+        // stream_type 5 = secondary audio. The conversion should
+        // still produce an Audio label (the registry's apply path
+        // can ignore secondary if it wants — this module just
+        // surfaces what's there).
+        let mut sec = audio_entry(0x1A00, 0x83, 3, 1, "eng");
+        sec.stream_type = 5;
+        sec.secondary = true;
+        let pl = playlist_with(vec![sec]);
+        let labels = labels_from_playlists(&[pl]);
+        assert_eq!(labels.len(), 1);
+        assert_eq!(labels[0].stream_type, StreamLabelType::Audio);
+        assert_eq!(labels[0].codec_hint, "TrueHD 2.0");
+    }
+}

src/labels/png_filenames.rs

@@ -0,0 +1,72 @@
+//! PNG-filename language token parser — stubbed (noop) pending need.
+//!
+//! ## What this would do
+//!
+//! Some discs encode per-language menu localization as pre-rendered PNG
+//! menu buttons, one per language, with the language token embedded in
+//! the filename. Examples observed in the 2026-05-10 corpus:
+//!
+//! - **disc-01 (The Amateur)** — `<region>_<lang>_<context>_<format>.png`
+//!   Region prefix: `USA` / `UK` / `JPN` / etc.
+//!   Lang tokens (3-char, uppercase): `ENG`, `FRC`, `FRP`, `DEU`, `DUT`,
+//!   `ITA`, `JPN`, `LAS`, `CSP`, `POL`, `CZE` (11 languages)
+//!
+//! - **disc-09 (Dune orig)** — `<title>_<variant>_<lang>_Composite<N>.png`
+//!   Lang tokens (3-char, mixed-case): `Eng`, `Ger` (2 languages)
+//!
+//! ## Why stubbed
+//!
+//! MPLS already gives per-stream `language` + `coding_type` + stream-type
+//! (audio vs subtitle) on every disc. For the 2 unknown-framework discs
+//! that PNG filenames would close (disc-01, disc-09), MPLS will produce
+//! a strict superset of what filenames could give us, because MPLS knows
+//! per-stream attribution while filenames only know "the disc offers
+//! these N language buttons."
+//!
+//! The **only** thing PNG filenames give us that MPLS doesn't is **studio
+//! variant disambiguation**:
+//! - `FRC` (French Canadian) vs `FRP` (French Parisian) — MPLS just says `fra`
+//! - `LAS` (Latin American Spanish) vs `CSP` (Castilian Spanish) — MPLS just says `spa`
+//!
+//! That's niche enough that it doesn't justify implementing right now.
+//! Reactivate this parser only when:
+//! 1. We hit a disc where MPLS is malformed/empty AND PNG filenames are
+//!    the only language hint, OR
+//! 2. A downstream consumer needs the studio variant suffix for output
+//!    naming (e.g. `Title (French Canadian).mkv` vs `Title (French).mkv`).
+//!
+//! ## When reactivating
+//!
+//! Implement `parse` to:
+//! 1. Iterate top-level PNG paths in `/BDMV/JAR/` (and `<id>/` subdirs).
+//! 2. Tokenize each filename on `_` / `-` / `.`
+//! 3. Match each token against an alias table:
+//!    - ISO 639-1 / 639-2 standard codes
+//!    - Studio variants: `FRC`/`FRP` → `fra-CA`/`fra-FR`,
+//!      `LAS`/`CSP` → `spa-419`/`spa-ES`,
+//!      mixed-case shortforms `Eng`/`Ger`/`Fra`/`Spa`/`Jpn` → ISO 639-2
+//!    - Country prefix filter: drop `USA`/`UK`/`JPN`/`AUS`/`GER`/`FR` when
+//!      they appear in position 0 (those are region markers, not langs).
+//! 4. Deduplicate. Confidence stays `Low` because we still don't know
+//!    per-stream codec or audio/subtitle attribution.
+//!
+//! Wire as ENRICHMENT after MPLS in `mod.rs::analyze`, not as a primary
+//! parser: PNG filenames upgrade `lang=fra` to `lang=fra-CA` when both
+//! sources agree on the disc; they should never overwrite MPLS data.
+
+use super::ParseResult;
+use crate::sector::SectorReader;
+use crate::udf::UdfFs;
+
+/// Stub: returns false so the dispatcher never calls `parse`. Reactivate
+/// by checking for the patterns described in the module docs.
+#[allow(dead_code)] // module-level noop, not wired into PARSERS until needed
+pub fn detect(_udf: &UdfFs) -> bool {
+    false
+}
+
+/// Stub: returns None. See module docs for the implementation sketch.
+#[allow(dead_code)] // module-level noop, not wired into PARSERS until needed
+pub fn parse(_reader: &mut dyn SectorReader, _udf: &UdfFs) -> Option<ParseResult> {
+    None
+}