More single-file splits from the re-index

Releases of one file whose counter announces many (910k of 3M after the
sixth reset's first minutes), by cause:
- An unquoted name with double spaces ("lara_bath  15of19 larabath_2.5.jpg",
  "Luisa:     (068 of 160)"): the subject is collapsed, the name wasn't,
  so it stayed in the remainder; it's taken out collapsed too.
- Audiobook and album tracks keyed by their own name ("RP - File 268 of
  295 - \"… Seeing A Large Cat - D11.16-22.mp3\"", "CANDYD :: \"Candy
  Dulfer - Candy Store - 11 - everytime.mp3\""): by the name without its
  numbers, as image sets.
- A size without a unit before the closing yEnc ("…mp3\" - 8123 yEnc").
- Random posters with separators ("SOFZRw1-OTEd@…", "…_OOW@…").
Both implementations; tests; fixtures identical (models: 267 -> 172).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

edge/src/names.rs

@@ -328,7 +328,32 @@ fn remove_counters_and_sizes(s: &str) -> String {
     let s = replace_counters(&SLASH_COUNTER, &s);
     let s = replace_counters(&ZERO_COUNTER, &s);
     let s = replace(&SIZES, &s, " ", true);
-    replace(&BRACKET_SIZES, &s, " ", true)
+    let s = replace(&BRACKET_SIZES, &s, " ", true);
+    strip_number_before_yenc(&s)
+}
+
+/// `.replace(/\b\d+(\s+yEnc\s*)$/i, '$1')`: a size without a unit right before the closing yEnc.
+fn strip_number_before_yenc(s: &str) -> String {
+    let tail_start = s.trim_end_matches(js::is_space).len();
+    let body = &s[..tail_start];
+    if body.len() < 4 || !body.is_char_boundary(body.len() - 4) || !body[body.len() - 4..].eq_ignore_ascii_case("yenc") {
+        return s.to_string();
+    }
+    let before = &body[..body.len() - 4];
+    let digits_end = before.trim_end_matches(js::is_space).len();
+    if digits_end == before.len() {
+        return s.to_string(); // \s+ needs a space
+    }
+    let b = before.as_bytes();
+    let digits_start = b[..digits_end].iter().rposition(|c| !c.is_ascii_digit()).map_or(0, |i| i + 1);
+    if digits_start == digits_end {
+        return s.to_string();
+    }
+    // \b: the digit run starts the text or follows a non-word character.
+    if digits_start > 0 && (b[digits_start - 1].is_ascii_alphanumeric() || b[digits_start - 1] == b'_') {
+        return s.to_string();
+    }
+    format!("{}{}", &s[..digits_start], &s[digits_end..])
 }
 
 /// The end of subjectRemainder: what's left of brackets, double spaces, trim (after the separators became single spaces).
@@ -517,7 +542,9 @@ pub fn random_poster(from: &str) -> bool {
         Some(address) => address,
         None => js::trim(from).split(' ').next().unwrap_or(""),
     };
-    looks_random(address.split_once('@').map_or(address, |(local, _)| local))
+    // Separators don't make a token less random.
+    let local: String = address.split_once('@').map_or(address, |(local, _)| local).chars().filter(|c| !matches!(c, '-' | '_' | '.')).collect();
+    looks_random(&local)
 }
 
 pub fn release_identity(poster: &str, subject: &str, name: &str) -> Release {
@@ -527,7 +554,15 @@ pub fn release_identity(poster: &str, subject: &str, name: &str) -> Release {
     let file_total = counter_total(subject).and_then(|t| t.parse::<i64>().ok());
     let stem = file_stem(name);
     // A name that isn't quoted (unquoted_name) is taken out as the quotes would have been.
-    let remainder = if subject.contains(&format!("\"{name}\"")) { subject_remainder(subject) } else { subject_remainder(&subject.replacen(name, " ", 1).replace('"', " ")) };
+    // The subject here has its spaces collapsed; the name may not.
+    let remainder = if subject.contains(&format!("\"{name}\"")) {
+        subject_remainder(subject)
+    } else {
+        let collapsed = js::collapse_spaces(name);
+        let collapsed = js::trim(&collapsed);
+        let without = if subject.contains(name) { subject.replacen(name, " ", 1) } else { subject.replacen(collapsed, " ", 1) };
+        subject_remainder(&without.replace('"', " "))
+    };
     // A document or comic is a release of its own.
     if test(&DOCUMENT, name) {
         let shown = replace(&DOCUMENT, &stem, "", false);
@@ -559,6 +594,13 @@ pub fn release_identity(poster: &str, subject: &str, name: &str) -> Release {
         let name = if prefix.is_empty() { stem.clone() } else { prefix.clone() };
         return Release { key: hash(&["stem", poster, &name]), name, source: "stem", file_total };
     }
+    // An audiobook's or album's tracks without a name around them: by their name without its numbers.
+    if test(&AUDIO, name) {
+        let tracks = image_prefix(&stem);
+        if !tracks.is_empty() {
+            return Release { key: hash(&["stem", poster, &tracks]), name: tracks, source: "stem", file_total };
+        }
+    }
     Release { key: hash(&["stem", poster, &stem]), name: stem, source: "stem", file_total }
 }
 
@@ -1056,6 +1098,30 @@ mod image_set_tests {
         assert!(!random_poster("w.merklein@freenet.de (muffin)"));
     }
 
+    #[test]
+    fn more_splits_from_the_reindex() {
+        use super::random_poster;
+        let makz = "Makz@BlueRadiance.uni (makz)";
+        // The name as unquoted_name gives it (spaces kept), the subject collapsed and without its counter.
+        let a = release_identity(makz, r"G:\posting\thegoodnudes\lara_bath 15of19 larabath_2.5.jpg", r"G:\posting\thegoodnudes\lara_bath  15of19 larabath_2.5.jpg");
+        let b = release_identity(makz, r"G:\posting\thegoodnudes\lara_bath 16of19 larabath_2.6.jpg", r"G:\posting\thegoodnudes\lara_bath  16of19 larabath_2.6.jpg");
+        assert_eq!(a.key, b.key);
+        let mara = "Mara <guest@cox.net>";
+        let a = release_identity(mara, r#"RP - File 268 of 295 - yEnc "Elizabeth Peters - (Peabody 09) Seeing A Large Cat - D11.16-22.mp3""#, "Elizabeth Peters - (Peabody 09) Seeing A Large Cat - D11.16-22.mp3");
+        let b = release_identity(mara, r#"RP - File 269 of 295 - yEnc "Elizabeth Peters - (Peabody 09) Seeing A Large Cat - D11.17-22.mp3""#, "Elizabeth Peters - (Peabody 09) Seeing A Large Cat - D11.17-22.mp3");
+        assert_eq!(a.key, b.key);
+        let mb = "misterb@sure.com (MisterB)";
+        let a = release_identity(mb, r#"Herb Ernst@256 - NMR -[07/17] "Herb Ernst - Dreamflight II - 05 - The Light of Forever.mp3" - 8123 yEnc"#, "Herb Ernst - Dreamflight II - 05 - The Light of Forever.mp3");
+        let b = release_identity(mb, r#"Herb Ernst@256 - NMR -[08/17] "Herb Ernst - Dreamflight II - 06 - Other.mp3" - 9001 yEnc"#, "Herb Ernst - Dreamflight II - 06 - Other.mp3");
+        assert_eq!(a.key, b.key);
+        for s in [r#"x - "a.mp3" - 8123 yEnc"#, "a8123 yEnc", "-8123 yEnc ", "yEnc"] {
+            assert_eq!(super::subject_remainder(s), super::subject_remainder_re(s), "{s}");
+        }
+        assert!(random_poster("SOFZRw1-OTEd@EEFUbh.srG"));
+        assert!(random_poster("vMgGYlZB9ZcTNk8WxYFGV_OOW@x6YR.abc"));
+        assert!(!random_poster("w.merklein@freenet.de (muffin)"));
+    }
+
     #[test]
     fn sizes_in_brackets_and_one_word_sets() {
         for s in [r#"Aletta Ocean - "100628-035.jpg" yEnc [347K]"#, r#"Corie - "2877-041.jpg" yEnc [290K]"#, "x - !!!blob2010-102katie009.jpg 336573 byt"] {

lib/release/identity.ts

@@ -77,6 +77,8 @@ export function subjectRemainder(subject: string): string {
     // File sizes ("76.88 MB", "335555 Bytes", "336573 byt" cut off, "[347K]") differ per file.
     .replace(/\b\d+(?:[.,]\d+)?\s*(?:[kmgt]i?b|bytes?|byt)\b/gi, ' ')
     .replace(/\[\s*\d+(?:[.,]\d+)?\s*[kmgt]\s*\]/gi, ' ')
+    // A size without a unit right before the closing yEnc ("…mp3" - 8123 yEnc").
+    .replace(/\b\d+(\s+yEnc\s*)$/i, '$1')
     // File sizes follow "yEnc" at the end; other numbers (years, post ids) stay.
     .replace(/\byEnc\b(\s+\d+)?\s*$/i, ' ')
     .replace(/\byEnc\b/gi, ' ')
@@ -195,7 +197,8 @@ export function randomPoster(from: string): boolean {
   const close = open === -1 ? -1 : from.indexOf('>', open + 1)
   const address = close !== -1 ? from.slice(open + 1, close) : from.trim().split(' ')[0]
   const at = address.indexOf('@')
-  return looksRandom(at === -1 ? address : address.slice(0, at))
+  // Separators don't make a token less random ("SOFZRw1-OTEd", "vMgGYlZB9ZcTNk8WxYFGV_OOW").
+  return looksRandom((at === -1 ? address : address.slice(0, at)).replace(/[-_.]/g, ''))
 }
 
 export function releaseIdentity(input: ReleaseInput): ReleaseIdentity | null {
@@ -216,7 +219,13 @@ export function releaseIdentity(input: ReleaseInput): ReleaseIdentity | null {
   // Text outside the quotes that names the release, e.g. "Some.Release [01/45] - "x.rar"". A name that isn't
   // quoted (unquotedName: "DVDFab 8.2.0.5.rar (006/332)", "[keira_s04_0037_l.jpg" yEnc] (2/2)") is taken out
   // as the quotes would have been: otherwise every file of a set would name a release of its own.
-  const remainder = subjectRemainder(input.subject.includes(`"${input.name}"`) ? input.subject : input.subject.replace(input.name, ' ').replace(/"/g, ' '))
+  // The subject here has its spaces collapsed; the name may not ("lara_bath  15of19 larabath_2.5.jpg").
+  const collapsed = input.name.replace(/\s+/g, ' ').trim()
+  const remainder = subjectRemainder(
+    input.subject.includes(`"${input.name}"`)
+      ? input.subject
+      : (input.subject.includes(input.name) ? input.subject.replace(input.name, ' ') : input.subject.replace(collapsed, ' ')).replace(/"/g, ' ')
+  )
   // A document or comic is a release of its own (collections post hundreds under one subject).
   if (DOCUMENT.test(input.name)) return { key: hash('stem', poster, stem), name: stem.replace(DOCUMENT, '') || stem, fileTotal, source: 'stem' }
   const image = IMAGE.test(input.name)
@@ -252,6 +261,12 @@ export function releaseIdentity(input: ReleaseInput): ReleaseIdentity | null {
   }
   // An image without words around it: its set by its name (keyed as a stem, so "x.jpg" joins "x.part01.rar").
   if (image) return { key: hash('stem', poster, prefix || stem), name: prefix || stem, fileTotal, source: 'stem' }
+  // An audiobook's or album's tracks without a name around them: by their name without its numbers, as an
+  // image set ("… Seeing A Large Cat - D11.16-22.mp3", "Candy Dulfer - Candy Store - 11 - everytime.mp3").
+  if (AUDIO.test(input.name)) {
+    const tracks = stem.replace(/\d+\D*$/, '').replace(/\d+/g, ' ').replace(/[\s._-]+/g, ' ').trim()
+    if (tracks) return { key: hash('stem', poster, tracks), name: tracks, fileTotal, source: 'stem' }
+  }
   // The [i/M] total is informational, not part of the key: posters often
   // upload the par2 index without a counter, and it belongs to the same post.
   return { key: hash('stem', poster, stem), name: stem, fileTotal, source: 'stem' }

tests/unit/release/identity.spec.ts

@@ -194,6 +194,26 @@ test.group('release / par2 names', () => {
     assert.equal(id(`${album} - 07 - Rascals - Sattva.mp3 (1/43)`).key, id(`${album} - 08 - Incredible String Band - Koeeoaddi There.mp3 (1/47)`).key)
   })
 
+  test('more splits from the re-index: spaces in unquoted names, audio tracks, sizes before yEnc, posters with separators', async ({ assert }) => {
+    const { parseSubject } = await import('#lib/headers/index')
+    const { randomPoster } = await import('#lib/release/identity')
+    // As the edge worker: the subject without its counter and with collapsed spaces, the name as parsed.
+    const id = (poster: string, subject: string) => {
+      const name = parseSubject(subject)!.name
+      const base = subject.replace(/\(\d+\/\d+\)\s*$/, '').replace(/\s+/g, ' ').trim()
+      return releaseIdentity({ poster, subject: base, name, weak: false })!
+    }
+    const makz = 'Makz@BlueRadiance.uni (makz)'
+    assert.equal(id(makz, 'G:\\posting\\thegoodnudes\\lara_bath  15of19 larabath_2.5.jpg (1/1)').key, id(makz, 'G:\\posting\\thegoodnudes\\lara_bath  16of19 larabath_2.6.jpg (1/1)').key)
+    const mara = 'Mara <guest@cox.net>'
+    assert.equal(id(mara, 'RP - File 268 of 295 - yEnc "Elizabeth Peters - (Peabody 09) Seeing A Large Cat - D11.16-22.mp3" (1/3)').key, id(mara, 'RP - File 269 of 295 - yEnc "Elizabeth Peters - (Peabody 09) Seeing A Large Cat - D11.17-22.mp3" (1/3)').key)
+    const mb = 'misterb@sure.com (MisterB)'
+    assert.equal(id(mb, 'Herb Ernst@256 - NMR -[07/17] "Herb Ernst - Dreamflight II - 05 - The Light of Forever.mp3" - 8123 yEnc (1/9)').key, id(mb, 'Herb Ernst@256 - NMR -[08/17] "Herb Ernst - Dreamflight II - 06 - Other.mp3" - 9001 yEnc (1/9)').key)
+    assert.isTrue(randomPoster('SOFZRw1-OTEd@EEFUbh.srG'))
+    assert.isTrue(randomPoster('vMgGYlZB9ZcTNk8WxYFGV_OOW@x6YR.abc'))
+    assert.isFalse(randomPoster('w.merklein@freenet.de (muffin)'))
+  })
+
   test('image sets: sizes in brackets and one-word set names', ({ assert }) => {
     const id = (subject: string, name: string) => releaseIdentity({ poster: 'p', subject, name, weak: false })!
     assert.equal(id('Aletta Ocean - "100628-035.jpg" yEnc [347K]', '100628-035.jpg').key, id('Aletta Ocean - "100628-036.jpg" yEnc [298K]', '100628-036.jpg').key)