Tighten DMHY template label heuristics

tools/rust_dmhy_template_apply/src/main.rs

@@ -201,10 +201,18 @@ static CJK_SEASON_TOKEN_RE: Lazy<Regex> =
 static CJK_SEASON_EMBEDDED_RE: Lazy<Regex> = Lazy::new(|| {
     Regex::new(r"^(.+?)(第[一二三四五六七八九十\d]+[季期部])(.{0,12})$").unwrap()
 });
+static CJK_EPISODE_EMBEDDED_RE: Lazy<Regex> =
+    Lazy::new(|| Regex::new(r"^(.+?)(第?\d{1,4}[话話回集])(.{0,32})$").unwrap());
+static CJK_TITLE_TRAILING_EPISODE_RE: Lazy<Regex> =
+    Lazy::new(|| Regex::new(r"^(.+[\p{Han}\p{Hiragana}\p{Katakana}])(\d{2,3})$").unwrap());
 static ASCII_SEASON_SUFFIX_RE: Lazy<Regex> =
     Lazy::new(|| Regex::new(r"(?i)^(.+?)[\s_.-]+(S\d{1,2})$").unwrap());
 static ORDINAL_SEASON_TOKEN_RE: Lazy<Regex> =
     Lazy::new(|| Regex::new(r"(?i)^\d{1,2}(?:st|nd|rd|th)$").unwrap());
+static WORD_ORDINAL_SEASON_TOKEN_RE: Lazy<Regex> = Lazy::new(|| {
+    Regex::new(r"(?i)^(?:First|Second|Third|Fourth|Fifth|Sixth|Seventh|Eighth|Ninth|Tenth)$")
+        .unwrap()
+});
 static SEASON_WORD_RE: Lazy<Regex> =
     Lazy::new(|| Regex::new(r"(?i)^(?:Season|Saison)$").unwrap());
 static CJK_TITLE_LANG_PREFIX_RE: Lazy<Regex> =
@@ -1764,7 +1772,7 @@ fn has_encoding_noise(value: &str) -> bool {
     let markers = [
         "譁", "蜈", "螟", "蟄", "謇", "邱", "荳", "縺", "繧", "莨", "鬆", "髯", "瀛",
         "楀", "箷", "绲", "刔", "鏃", "湪", "鏍", "犲", "儚", "鐗", "吀", "铦", "躲",
-        "伄", "椋", "伓", "姘", "帽",
+        "伄", "椋", "伓", "姘", "帽", "娆", "洖", "浜", "堝", "澶", "湴", "鐒",
     ];
     let marker_hits = markers
         .iter()
@@ -1783,6 +1791,11 @@ fn has_encoding_noise(value: &str) -> bool {
 fn has_non_anime_noise(value: &str) -> bool {
     let normalized = value.replace('\\', "/").trim().to_ascii_lowercase();
     normalized == "mtv" || normalized.starts_with("mtv/") || normalized.contains("/mtv/")
+        || value.contains("[旅游")
+        || value.contains("[旅游番")
+        || normalized.contains("tokyo deep")
+        || value.contains("日本不思议铁路之旅")
+        || value.contains("ニッポンぶらり鉄道旅")
 }
 
 fn normalized_path_segment(value: &str) -> String {
@@ -2235,6 +2248,7 @@ const KNOWN_TITLE_PHRASES: &[&[&str]] = &[
     &["Phantasy", "Star", "Online", "2", "Episode", "Oracle"],
     &["Durarara", "2", "Ketsu"],
     &["Ghiblies", "Episode", "2"],
+    &["Eien", "no", "831"],
     &["Lupin The Thrid Jigen Daisuke no Bohyou"],
     &["Lupin The Third Jigen Daisuke no Bohyou"],
 ];
@@ -2705,6 +2719,25 @@ fn adjust_contextual_roles(tokens: &[String], groups: &[Group], roles: &[String]
             output[index] = "O".to_string();
             continue;
         }
+        if roles[index] == "TITLE"
+            && matches!(text.as_str(), "TVアニメ" | "テレビアニメ")
+            && output
+                .iter()
+                .enumerate()
+                .any(|(other, role)| other != index && role == "TITLE")
+        {
+            output[index] = "O".to_string();
+            continue;
+        }
+        if output[index] == "TITLE" && text.eq_ignore_ascii_case("Creditless") {
+            let later_special = output[index + 1..]
+                .iter()
+                .any(|role| role == "SPECIAL");
+            if later_special {
+                output[index] = "SPECIAL".to_string();
+                continue;
+            }
+        }
         if roles[index] == "TITLE" && matches!(text.as_str(), "第" | "話" | "话" | "回" | "集")
         {
             output[index] = "O".to_string();
@@ -3039,6 +3072,11 @@ fn normalize_title_token(token: &str) -> (Vec<String>, Vec<String>) {
             labels.push("B-SEASON".to_string());
             continue;
         }
+        if EPISODE_CJK_RE.is_match(&piece) {
+            output_pieces.push(piece);
+            labels.push("B-EPISODE".to_string());
+            continue;
+        }
         if let Some(caps) = ASCII_SEASON_SUFFIX_RE.captures(&piece) {
             let before = caps.get(1).map(|m| m.as_str()).unwrap_or_default();
             let season = caps.get(2).map(|m| m.as_str()).unwrap_or_default();
@@ -3066,6 +3104,33 @@ fn normalize_title_token(token: &str) -> (Vec<String>, Vec<String>) {
             }
             continue;
         }
+        if let Some(caps) = CJK_EPISODE_EMBEDDED_RE.captures(&piece) {
+            let before = caps.get(1).map(|m| m.as_str()).unwrap_or_default();
+            let episode = caps.get(2).map(|m| m.as_str()).unwrap_or_default();
+            let after = caps.get(3).map(|m| m.as_str()).unwrap_or_default();
+            if !before.is_empty() {
+                output_pieces.push(before.to_string());
+                labels.push("B-TITLE".to_string());
+            }
+            output_pieces.push(episode.to_string());
+            labels.push("B-EPISODE".to_string());
+            if !after.is_empty() {
+                output_pieces.push(after.to_string());
+                labels.push("O".to_string());
+            }
+            continue;
+        }
+        if let Some(caps) = CJK_TITLE_TRAILING_EPISODE_RE.captures(&piece) {
+            let before = caps.get(1).map(|m| m.as_str()).unwrap_or_default();
+            let episode = caps.get(2).map(|m| m.as_str()).unwrap_or_default();
+            if !before.is_empty() {
+                output_pieces.push(before.to_string());
+                labels.push("B-TITLE".to_string());
+            }
+            output_pieces.push(episode.to_string());
+            labels.push("B-EPISODE".to_string());
+            continue;
+        }
         output_pieces.push(piece);
         labels.push("B-TITLE".to_string());
     }
@@ -3227,18 +3292,58 @@ fn smooth_title_spans(tokens: &[String], labels: &[String]) -> Vec<String> {
     let joiners = [
         " ", ".", "-", "_", "·", "・", "×", "／", "/", "'", "’", ":", "：", "!", "！", "?",
         "？", ";", "；", ",", "，", "、", "。", "～", "~", "－", "+", "＋", "(", ")",
-        "（", "）", "[", "]", "【", "】", "<", ">", "＜", "＞", "｢", "｣", "「", "」",
-        "☆", "♪", "`", "@",
+        "（", "）", "[", "]", "【", "】", "<", ">", "＜", "＞", "｢", "｣", "「", "」", "《", "》",
+        "☆", "♪", "`", "@", "‐", "‑", "–", "—", "−", "$", "＄", "∽", "꞉", "♥",
     ];
     let title_terminal_punctuation = ["!", "！", "?", "？"];
     let entity_joiners = [
         " ", ".", "-", "_", "·", "・", "×", "／", "/", "'", "’", ":", "：", "!", "！", "?",
         "？", ";", "；", ",", "，", "、", "。", "～", "~", "－", "+", "＋", "(", ")",
-        "（", "）", "[", "]", "【", "】", "<", ">", "＜", "＞", "｢", "｣", "「", "」",
-        "☆", "♪", "`", "@", "&", "＆",
+        "（", "）", "[", "]", "【", "】", "<", ">", "＜", "＞", "｢", "｣", "「", "」", "《", "》",
+        "☆", "♪", "`", "@", "&", "＆", "‐", "‑", "–", "—", "−", "$", "＄", "∽", "꞉", "♥",
     ];
     let mut output = labels.to_vec();
     for (index, (token, label)) in tokens.iter().zip(labels.iter()).enumerate() {
+        if label == "B-TITLE"
+            && token == "TV"
+            && index + 1 < tokens.len()
+            && tokens[index + 1] == "アニメ"
+            && output[index + 2..].iter().any(|label| label == "B-TITLE")
+        {
+            output[index] = "O".to_string();
+            output[index + 1] = "O".to_string();
+            continue;
+        }
+        if label == "B-TITLE"
+            && token == "アニメ"
+            && output[index + 1..].iter().any(|label| label == "B-TITLE")
+        {
+            output[index] = "O".to_string();
+            continue;
+        }
+        if label == "B-TITLE" && token.eq_ignore_ascii_case("part") {
+            let next_number = (index + 1..tokens.len()).find(|&cursor| {
+                !joiners.contains(&tokens[cursor].as_str())
+                    && !tokens[cursor].chars().all(char::is_whitespace)
+            });
+            let nearby_lupin = tokens[..index]
+                .iter()
+                .rev()
+                .take(8)
+                .any(|item| item.eq_ignore_ascii_case("lupin"))
+                || tokens[index + 1..]
+                    .iter()
+                    .take(12)
+                    .any(|item| item.eq_ignore_ascii_case("lupin"));
+            if nearby_lupin
+                && next_number.is_some_and(|cursor| {
+                    tokens[cursor].chars().all(|ch| ch.is_ascii_digit()) && tokens[cursor].len() <= 2
+                })
+            {
+                output[index] = "B-SEASON".to_string();
+                continue;
+            }
+        }
         if label == "B-TITLE"
             && token.chars().all(|ch| ch.is_ascii_digit())
             && token.len() == 3
@@ -3283,6 +3388,11 @@ fn smooth_title_spans(tokens: &[String], labels: &[String]) -> Vec<String> {
                         .chars()
                         .any(|ch| ch.is_alphanumeric() || ('\u{4e00}'..='\u{9fff}').contains(&ch))
             });
+            let later_episode = (index + 1..tokens.len()).any(|cursor| labels[cursor] == "B-EPISODE");
+            if previous_title_word.is_none() && later_episode {
+                output[index] = "B-SEASON".to_string();
+                continue;
+            }
             let previous_word = previous_title_word.map(|cursor| tokens[cursor].to_ascii_lowercase());
             if previous_title_word.is_some()
                 && !matches!(previous_word.as_deref(), Some("lupin"))
@@ -3291,7 +3401,10 @@ fn smooth_title_spans(tokens: &[String], labels: &[String]) -> Vec<String> {
                 continue;
             }
         }
-        if label == "B-TITLE" && ORDINAL_SEASON_TOKEN_RE.is_match(token) {
+        if label == "B-TITLE"
+            && (ORDINAL_SEASON_TOKEN_RE.is_match(token)
+                || WORD_ORDINAL_SEASON_TOKEN_RE.is_match(token))
+        {
             let next_word = (index + 1..tokens.len()).find(|&cursor| {
                 !joiners.contains(&tokens[cursor].as_str())
                     && tokens[cursor].chars().any(|ch| ch.is_alphabetic())
@@ -3311,6 +3424,136 @@ fn smooth_title_spans(tokens: &[String], labels: &[String]) -> Vec<String> {
                 continue;
             }
         }
+        if label == "O"
+            && (EPISODE_CJK_RE.is_match(token)
+                || EPISODE_VALUE_RE.is_match(token)
+                || EPISODE_RANGE_RE.is_match(token))
+        {
+            output[index] = "B-EPISODE".to_string();
+            continue;
+        }
+        if label == "O" && token.chars().all(|ch| ch.is_ascii_digit()) {
+            let previous_non_space = (0..index)
+                .rev()
+                .find(|&cursor| !tokens[cursor].chars().all(char::is_whitespace));
+            let next_non_space = (index + 1..tokens.len())
+                .find(|&cursor| !tokens[cursor].chars().all(char::is_whitespace));
+            if previous_non_space.is_some_and(|cursor| tokens[cursor] == "#") {
+                output[index] = "B-EPISODE".to_string();
+                if next_non_space.is_some_and(|cursor| matches!(tokens[cursor].as_str(), "-" | "~"))
+                {
+                    if let Some(separator) = next_non_space {
+                        output[separator] = "B-EPISODE".to_string();
+                        if let Some(right) = (separator + 1..tokens.len())
+                            .find(|&cursor| !tokens[cursor].chars().all(char::is_whitespace))
+                        {
+                            if tokens[right].chars().all(|ch| ch.is_ascii_digit()) {
+                                output[right] = "B-EPISODE".to_string();
+                            }
+                        }
+                    }
+                }
+                continue;
+            }
+            if previous_non_space.is_some_and(|cursor| tokens[cursor] == "第")
+                && next_non_space
+                    .is_some_and(|cursor| {
+                        matches!(tokens[cursor].as_str(), "话" | "話" | "回" | "集")
+                            || tokens[cursor].starts_with('话')
+                            || tokens[cursor].starts_with('話')
+                            || tokens[cursor].starts_with('回')
+                            || tokens[cursor].starts_with('集')
+                    })
+            {
+                if let Some(cursor) = previous_non_space {
+                    output[cursor] = "B-EPISODE".to_string();
+                }
+                output[index] = "B-EPISODE".to_string();
+                if let Some(cursor) = next_non_space {
+                    if matches!(tokens[cursor].as_str(), "话" | "話" | "回" | "集") {
+                        output[cursor] = "B-EPISODE".to_string();
+                    }
+                }
+                continue;
+            }
+        }
+        if matches!(label.as_str(), "B-TITLE" | "O")
+            && token.chars().all(|ch| ch.is_ascii_digit())
+            && token.len() <= 3
+        {
+            let previous_word = (0..index)
+                .rev()
+                .find(|&cursor| {
+                    !joiners.contains(&tokens[cursor].as_str())
+                        && tokens[cursor].chars().any(|ch| ch.is_alphabetic())
+                })
+                .map(|cursor| tokens[cursor].to_ascii_lowercase());
+            let next_structural = (index + 1..tokens.len())
+                .find(|&cursor| !joiners.contains(&tokens[cursor].as_str()))
+                .map(|cursor| tokens[cursor].as_str());
+            let next_non_space = (index + 1..tokens.len())
+                .find(|&cursor| !tokens[cursor].chars().all(char::is_whitespace))
+                .map(|cursor| tokens[cursor].as_str());
+            let later_technical_block = output[index + 1..]
+                .iter()
+                .any(|label| matches!(label.as_str(), "B-SOURCE" | "B-RESOLUTION"));
+            let nearby_lupin_part = previous_word.as_deref() == Some("part")
+                && (tokens[..index]
+                    .iter()
+                    .rev()
+                    .take(8)
+                    .any(|item| item.eq_ignore_ascii_case("lupin"))
+                    || tokens[index + 1..]
+                        .iter()
+                        .take(12)
+                        .any(|item| item.eq_ignore_ascii_case("lupin")));
+            if nearby_lupin_part {
+                output[index] = "B-SEASON".to_string();
+                continue;
+            }
+            let followed_by_title_word = (index + 1..tokens.len())
+                .find(|&cursor| {
+                    !joiners.contains(&tokens[cursor].as_str())
+                        && !matches!(tokens[cursor].as_str(), "-" | "－" | "," | "，" | ":" | "：")
+                })
+                .is_some_and(|cursor| {
+                    !matches!(tokens[cursor].as_str(), "[" | "【" | "(" | "（" | "]" | "】")
+                        && output
+                            .get(cursor)
+                            .is_some_and(|label| label == "B-TITLE")
+                        && tokens[cursor].chars().any(|ch| ch.is_alphabetic())
+                });
+            if followed_by_title_word && matches!(previous_word.as_deref(), Some("movie" | "part"))
+            {
+                output[index] = "B-TITLE".to_string();
+                continue;
+            }
+            if (later_technical_block
+                || next_non_space.is_some_and(|token| matches!(token, "[" | "【" | "(" | "（"))
+                || next_structural.is_some_and(|token| matches!(token, "[" | "【" | "(" | "（")))
+                && matches!(previous_word.as_deref(), Some("movie" | "part"))
+            {
+                output[index] = "B-SPECIAL".to_string();
+                continue;
+            }
+            let eien_title_number = token == "831"
+                && previous_word.as_deref() == Some("no")
+                && (0..index).any(|cursor| {
+                    output[cursor] == "B-TITLE" && tokens[cursor].eq_ignore_ascii_case("Eien")
+                });
+            if eien_title_number {
+                for joiner_index in (0..index).rev() {
+                    if tokens[joiner_index].eq_ignore_ascii_case("no") {
+                        break;
+                    }
+                    if joiners.contains(&tokens[joiner_index].as_str()) {
+                        output[joiner_index] = "B-TITLE".to_string();
+                    }
+                }
+                output[index] = "B-TITLE".to_string();
+                continue;
+            }
+        }
         if label == "O"
             && token.chars().all(|ch| ch.is_ascii_digit())
             && token.len() <= 3
@@ -3320,8 +3563,8 @@ fn smooth_title_spans(tokens: &[String], labels: &[String]) -> Vec<String> {
                 .find(|&cursor| !tokens[cursor].chars().all(char::is_whitespace));
             let next_non_space = (index + 1..tokens.len())
                 .find(|&cursor| !tokens[cursor].chars().all(char::is_whitespace));
-            if previous_non_space.is_some_and(|cursor| tokens[cursor] == "[")
-                && next_non_space.is_some_and(|cursor| tokens[cursor] == "]")
+            if previous_non_space.is_some_and(|cursor| matches!(tokens[cursor].as_str(), "[" | "【"))
+                && next_non_space.is_some_and(|cursor| matches!(tokens[cursor].as_str(), "]" | "】"))
                 && output[..index].iter().any(|label| label == "B-TITLE")
                 && output[index + 1..]
                     .iter()
@@ -3330,11 +3573,45 @@ fn smooth_title_spans(tokens: &[String], labels: &[String]) -> Vec<String> {
                 output[index] = "B-EPISODE".to_string();
                 continue;
             }
+            if previous_non_space.is_some_and(|cursor| matches!(tokens[cursor].as_str(), "-" | "－"))
+                && output[..index].iter().any(|label| label == "B-TITLE")
+                && output[index + 1..]
+                    .iter()
+                    .any(|label| matches!(label.as_str(), "B-SOURCE" | "B-RESOLUTION"))
+            {
+                output[index] = "B-EPISODE".to_string();
+                continue;
+            }
+            if next_non_space.is_none()
+                && previous_non_space.is_some_and(|cursor| {
+                    output[cursor] == "B-TITLE"
+                        && tokens[cursor].chars().any(|ch| {
+                            ('\u{4e00}'..='\u{9fff}').contains(&ch)
+                                || ('\u{3040}'..='\u{30ff}').contains(&ch)
+                        })
+                })
+            {
+                output[index] = "B-EPISODE".to_string();
+                continue;
+            }
         }
         if label == "B-EPISODE" && token.chars().all(|ch| ch.is_ascii_digit()) {
             let previous_non_space = (0..index)
                 .rev()
                 .find(|&cursor| !tokens[cursor].chars().all(char::is_whitespace));
+            let next_non_space = (index + 1..tokens.len())
+                .find(|&cursor| !tokens[cursor].chars().all(char::is_whitespace));
+            if previous_non_space.is_some_and(|cursor| tokens[cursor] == "第")
+                && next_non_space
+                    .is_some_and(|cursor| matches!(tokens[cursor].as_str(), "话" | "話" | "回" | "集"))
+            {
+                if let Some(cursor) = previous_non_space {
+                    output[cursor] = "B-EPISODE".to_string();
+                }
+                if let Some(cursor) = next_non_space {
+                    output[cursor] = "B-EPISODE".to_string();
+                }
+            }
             if previous_non_space
                 .is_some_and(|cursor| matches!(tokens[cursor].as_str(), "×" | "x" | "X"))
             {
@@ -3366,7 +3643,13 @@ fn smooth_title_spans(tokens: &[String], labels: &[String]) -> Vec<String> {
                 }
                 break;
             }
-            if matches!(previous_word.as_deref(), Some("lesson" | "part" | "no")) {
+            let previous_non_space = (0..index)
+                .rev()
+                .find(|&cursor| !tokens[cursor].chars().all(char::is_whitespace))
+                .map(|cursor| tokens[cursor].as_str());
+            if matches!(previous_word.as_deref(), Some("lesson" | "part"))
+                || (previous_word.as_deref() == Some("no") && previous_non_space == Some("."))
+            {
                 output[index] = "O".to_string();
                 continue;
             }
@@ -3851,6 +4134,10 @@ mod tests {
         assert!(has_encoding_noise(
             "[4K_SDR][DBD-Raws&HKG瀛楀箷绲刔[鏃ュ湪鏍″湌][01][2160P]"
         ));
+        assert!(has_encoding_noise("ATRI -My Dear Moments-/娆″洖浜堝憡 EP01 Log01"));
+        assert!(has_non_anime_noise(
+            "13-[旅游番][花丸字幕组][日本不思议铁路之旅][15.03.19-16.02.03][720&1080][中日双语]/铁道旅 15.03.19 720"
+        ));
 
         let tintin = "Adventures of Tintin (1991) Season 1-3 S01-S03 (1080p BluRay x265 HEVC 10bit EAC3 2.0 Garshasp)/Season 1/Adventures of Tintin (1991) - S01E06 - Cigars of the Pharaoh (Part 1) (1080p BluRay x265 Garshasp)";
         let (trimmed, was_trimmed) = training_filename_for(tintin);
@@ -4091,6 +4378,49 @@ mod tests {
         assert!(bleach_movie.contains(&("3".to_string(), "B-TITLE".to_string())));
         assert!(!bleach_movie.contains(&("3".to_string(), "B-EPISODE".to_string())));
 
+        let conan_movie =
+            labels_for("[DBD-Raws][Detective Conan Movie 27 The Million-Dollar Pentagram][PV][01][1080P]");
+        assert!(conan_movie.contains(&("27".to_string(), "B-TITLE".to_string())));
+        assert!(conan_movie.contains(&("PV".to_string(), "B-SPECIAL".to_string())));
+
+        let madoka_movie =
+            labels_for("[DBD-Raws][Puella Magi Madoka Magica the Movie 01 Beginnings][NCED][1080P]");
+        assert!(madoka_movie.contains(&("01".to_string(), "B-TITLE".to_string())));
+        assert!(madoka_movie.contains(&("Beginnings".to_string(), "B-TITLE".to_string())));
+
+        let fate_first_order =
+            labels_for("[DBD-Raws][Fate Grand Order ‐First Order‐][PV][01][1080P]");
+        assert!(fate_first_order.contains(&("Fate".to_string(), "B-TITLE".to_string())));
+        assert!(fate_first_order.contains(&("‐".to_string(), "B-TITLE".to_string())));
+        assert!(fate_first_order.contains(&("First".to_string(), "B-TITLE".to_string())));
+
+        let trillion_game = labels_for("[ANi] 一兆＄遊戲 - 03 [1080P][Baha][WEB-DL][AAC AVC][CHT]");
+        assert!(trillion_game.contains(&("一兆".to_string(), "B-TITLE".to_string())));
+        assert!(trillion_game.contains(&("＄".to_string(), "B-TITLE".to_string())));
+        assert!(trillion_game.contains(&("遊戲".to_string(), "B-TITLE".to_string())));
+
+        let lapis = labels_for("[Nekomoe kissaten&LoliHouse] Lapis Re꞉LiGHTs - PV01 [BDRip 1080p]");
+        assert!(lapis.contains(&("Re".to_string(), "B-TITLE".to_string())));
+        assert!(lapis.contains(&("꞉".to_string(), "B-TITLE".to_string())));
+        assert!(lapis.contains(&("LiGHTs".to_string(), "B-TITLE".to_string())));
+
+        let rezero = labels_for("TVアニメ『Re：ゼロから始める異世界生活』第10話「鬼がかったやり方」予告");
+        assert!(!rezero.contains(&("TV".to_string(), "B-TITLE".to_string())));
+        assert!(!rezero.contains(&("アニメ".to_string(), "B-TITLE".to_string())));
+        assert!(rezero.contains(&("Re".to_string(), "B-TITLE".to_string())));
+        assert!(rezero.contains(&("第".to_string(), "B-EPISODE".to_string())));
+        assert!(rezero.contains(&("話".to_string(), "B-EPISODE".to_string())));
+
+        let shark = labels_for("アニメ『おでかけ子ザメ』第10話「かじゅえん」");
+        assert!(!shark.contains(&("アニメ".to_string(), "B-TITLE".to_string())));
+        assert!(shark.contains(&("おでかけ子ザメ".to_string(), "B-TITLE".to_string())));
+
+        let creditless = labels_for(
+            "[ANK-Raws] デート・ア・ライブⅡ Creditless ED (Bdrip 1920x1080 HEVC FLAC)",
+        );
+        assert!(creditless.contains(&("Creditless".to_string(), "B-SPECIAL".to_string())));
+        assert!(creditless.contains(&("ED".to_string(), "B-SPECIAL".to_string())));
+
         let no_number = labels_for("[甜甜圈字幕组] 小讨厌 081「爷爷的礼物 No.1」");
         assert!(no_number.contains(&("081".to_string(), "B-EPISODE".to_string())));
         assert!(!no_number.contains(&("1".to_string(), "B-EPISODE".to_string())));
@@ -4110,6 +4440,16 @@ mod tests {
         assert!(hayate.contains(&("Season".to_string(), "B-SEASON".to_string())));
         assert!(hayate.contains(&("23".to_string(), "B-EPISODE".to_string())));
 
+        let yama = labels_for("[A.I.R.nesSub][Yama_no_Susume_Second_Season][08][720p]");
+        assert!(yama.contains(&("Yama".to_string(), "B-TITLE".to_string())));
+        assert!(yama.contains(&("Second".to_string(), "B-SEASON".to_string())));
+        assert!(yama.contains(&("Season".to_string(), "B-SEASON".to_string())));
+
+        let one_room = labels_for("[DMG][One Room Second Season][00][1080P][BIG5]");
+        assert!(one_room.contains(&("One".to_string(), "B-TITLE".to_string())));
+        assert!(one_room.contains(&("Second".to_string(), "B-SEASON".to_string())));
+        assert!(one_room.contains(&("Season".to_string(), "B-SEASON".to_string())));
+
         let jade = labels_for("[GM-Team][国漫][诛仙 第2季][Jade Dynasty Ⅱ][2024][12][AVC][GB][1080P]");
         assert!(jade.contains(&("Jade".to_string(), "B-TITLE".to_string())));
         assert!(jade.contains(&("Dynasty".to_string(), "B-TITLE".to_string())));
@@ -4122,6 +4462,11 @@ mod tests {
         assert!(yu_no.contains(&("NO".to_string(), "B-TITLE".to_string())));
         assert!(yu_no.contains(&("23".to_string(), "B-EPISODE".to_string())));
 
+        let yu_no_dash =
+            labels_for("[LowPower-Raws] この世の果てで恋を唄う少女YU-NO - 01 (BD 1080P x264 FLAC)");
+        assert!(yu_no_dash.contains(&("NO".to_string(), "B-TITLE".to_string())));
+        assert!(yu_no_dash.contains(&("01".to_string(), "B-EPISODE".to_string())));
+
         let fox = labels_for(
             "[GM-Team][国漫][狐妖小红娘 尾生篇][Fox Spirit Matchmaker Ⅷ][2019][05][AVC][GB][1080P]",
         );
@@ -4143,6 +4488,68 @@ mod tests {
         assert!(date_live_special.contains(&("Ⅱ".to_string(), "B-SEASON".to_string())));
         assert!(date_live_special.contains(&("CM01".to_string(), "B-SPECIAL".to_string())));
 
+        let lupin_part =
+            labels_for("[SnowDream][Part 5_Lupin Sansei Part 5][01][BIG5][720P]");
+        assert!(lupin_part.contains(&("Lupin".to_string(), "B-TITLE".to_string())));
+        assert!(lupin_part.contains(&("Sansei".to_string(), "B-TITLE".to_string())));
+        assert!(!lupin_part.contains(&("Part".to_string(), "B-TITLE".to_string())));
+        assert!(lupin_part.contains(&("5".to_string(), "B-SEASON".to_string())));
+        assert!(!lupin_part.contains(&("5".to_string(), "B-SPECIAL".to_string())));
+
+        let roman_leaf = dmhy_record("Ⅰ 001 魯邦燃起了鬥志", "tpl_test", &suggested_roles("TEXT SEP EPISODE SEP TEXT")).unwrap();
+        assert!(roman_leaf
+            .tokens
+            .iter()
+            .zip(roman_leaf.labels.iter())
+            .any(|(token, label)| token == "Ⅰ" && label == "B-SEASON"));
+        assert!(audit_warnings(&roman_leaf).contains(&"no_title".to_string()));
+
+        let hallow = labels_for("[c.c动漫 ccwzz.cc][驱魔少年HALLOW][第09话][GB][720p]");
+        assert!(hallow.contains(&("驱魔少年HALLOW".to_string(), "B-TITLE".to_string())));
+        assert!(hallow.contains(&("第09话".to_string(), "B-EPISODE".to_string())));
+
+        let fairy = labels_for("[魔導少年 最終章][EP35][繁体][1080P]");
+        assert!(fairy.contains(&("魔導少年".to_string(), "B-TITLE".to_string())));
+        assert!(fairy.contains(&("EP35".to_string(), "B-EPISODE".to_string())));
+
+        let mebius = labels_for("【CXRAW】【ウルトラマンメビウス】【22】【日々の未来】【DVDrip】【x264 Hi10P AAC】【MP4】");
+        assert!(mebius.contains(&("ウルトラマンメビウス".to_string(), "B-TITLE".to_string())));
+        assert!(mebius.contains(&("22".to_string(), "B-EPISODE".to_string())));
+
+        let battle = labels_for("斗破苍穹三年之约第01话");
+        assert!(battle.contains(&("斗破苍穹三年之约".to_string(), "B-TITLE".to_string())));
+        assert!(battle.contains(&("第".to_string(), "B-EPISODE".to_string())));
+        assert!(battle.contains(&("01".to_string(), "B-EPISODE".to_string())));
+        assert!(battle.contains(&("话".to_string(), "B-EPISODE".to_string())));
+
+        let hakumei = labels_for("妖精森林的小不点01");
+        assert!(hakumei.contains(&("妖精森林的小不点".to_string(), "B-TITLE".to_string())));
+        assert!(hakumei.contains(&("01".to_string(), "B-EPISODE".to_string())));
+
+        let decimal_episode_title = labels_for("无限系统树：第1话可能性的起点");
+        assert!(decimal_episode_title.contains(&("无限系统树".to_string(), "B-TITLE".to_string())));
+        assert!(decimal_episode_title.contains(&("第".to_string(), "B-EPISODE".to_string())));
+        assert!(decimal_episode_title.contains(&("1".to_string(), "B-EPISODE".to_string())));
+
+        let hash_range = labels_for("花田少年史#1-3");
+        assert!(hash_range.contains(&("花田少年史".to_string(), "B-TITLE".to_string())));
+        assert!(hash_range.contains(&("1".to_string(), "B-EPISODE".to_string())));
+        assert!(hash_range.contains(&("-".to_string(), "B-EPISODE".to_string())));
+        assert!(hash_range.contains(&("3".to_string(), "B-EPISODE".to_string())));
+
+        let movie_number = labels_for("[Kamigami] Haikyuu!! Movie - 01 [BD 1080p x265 Ma10p AAC]");
+        assert!(movie_number.contains(&("Haikyuu".to_string(), "B-TITLE".to_string())));
+        assert!(movie_number.contains(&("01".to_string(), "B-SPECIAL".to_string())));
+        assert!(!movie_number.contains(&("01".to_string(), "B-EPISODE".to_string())));
+
+        let ajin_movie = labels_for("[Moozzi2] Ajin The Movie - 01 (BD 1920x1080 x.264 FLACx2)");
+        assert!(ajin_movie.contains(&("Ajin".to_string(), "B-TITLE".to_string())));
+        assert!(ajin_movie.contains(&("01".to_string(), "B-SPECIAL".to_string())));
+
+        let eien = labels_for("[Nekomoe kissaten&LoliHouse] Eien no 831 [WebRip 1080p HEVC-10bit AAC ASSx2]");
+        assert!(eien.contains(&("Eien".to_string(), "B-TITLE".to_string())));
+        assert!(eien.contains(&("831".to_string(), "B-TITLE".to_string())));
+
         let ep_only = dmhy_record("Ep.25", "tpl_test", &suggested_roles("TEXT SEP EPISODE")).unwrap();
         assert!(audit_warnings(&ep_only).contains(&"no_title".to_string()));
     }