Method rules_fallback

cake-core/src/models/luxtts/tokenizer.rs:165–236

Simple rule-based English grapheme→IPA fallback for OOV words.

(&self, word: &str)

    /// Simple rule-based English grapheme→IPA fallback for OOV words.
    fn rules_fallback(&self, word: &str) -> String {
        let mut ipa = String::new();
        let chars: Vec<char> = word.chars().collect();
        let mut i = 0;

        while i < chars.len() {
            let remaining = &word[i..];
            // Try digraphs first
            let (phoneme, advance) = if remaining.len() >= 2 {
                match &remaining[..2] {
                    "th" => ("\u{03B8}", 2), // θ
                    "sh" => ("\u{0283}", 2), // ʃ
                    "ch" => ("t\u{0283}", 2), // tʃ
                    "ng" => ("\u{014B}", 2), // ŋ
                    "ph" => ("f", 2),
                    "wh" => ("w", 2),
                    "ck" => ("k", 2),
                    "ee" => ("i\u{02D0}", 2), // iː
                    "oo" => ("u\u{02D0}", 2), // uː
                    "ou" => ("a\u{028A}", 2), // aʊ
                    "ow" => ("a\u{028A}", 2), // aʊ
                    "ai" => ("e\u{026A}", 2), // eɪ
                    "ay" => ("e\u{026A}", 2), // eɪ
                    "ea" => ("i\u{02D0}", 2), // iː
                    _ => ("", 0),
                }
            } else {
                ("", 0)
            };

            if advance > 0 {
                ipa.push_str(phoneme);
                i += advance;
                continue;
            }

            // Single character mapping
            let phoneme = match chars[i] {
                'a' => "\u{00E6}", // æ
                'b' => "b",
                'c' => "k",
                'd' => "d",
                'e' => "\u{025B}", // ɛ
                'f' => "f",
                'g' => "g",
                'h' => "h",
                'i' => "\u{026A}", // ɪ
                'j' => "d\u{0292}", // dʒ
                'k' => "k",
                'l' => "l",
                'm' => "m",
                'n' => "n",
                'o' => "\u{0251}", // ɑ
                'p' => "p",
                'q' => "k",
                'r' => "\u{0279}", // ɹ
                's' => "s",
                't' => "t",