seams 0.1.14

// WHY: Centralized abbreviation handling for sentence boundary detection
// Extracted from working test strategies to prevent false sentence splits

use std::collections::HashSet;

/// Title abbreviations that cause false sentence boundaries when followed by proper nouns
/// These are the first part of 2-segment identifiers like "Dr. Smith", "Mr. Johnson"
/// Single capital letters are commonly used as abbreviations for proper names
pub const TITLE_ABBREVIATIONS: &[&str] = &[
    "Dr.", "Mr.", "Mrs.", "Ms.", "Prof.", "Sr.", "Jr.", "St.",
    // Single capital letters - commonly used as abbreviations for proper names
    "A.", "B.", "C.", "D.", "E.", "F.", "G.", "H.", "I.", "J.", "K.", "L.", "M.",
    "N.", "O.", "P.", "Q.", "R.", "S.", "T.", "U.", "V.", "W.", "X.", "Y.", "Z.",
    // Clergy & religious
    "Rev.", "Fr.", "Rab.", "Msgr.",
    // Political / civic
    "Gov.", "Sen.", "Rep.", "Amb.", "Hon.", "Ald.",
    // Military ranks
    "Adm.", "Capt.", "Cmdr.", "Col.", "Gen.", "Lt.", "Maj.", "Sgt.", "Cpl.", "Pvt.", "Ens.",
    // Academic / professional (beyond Prof.)
    "Pres.", "Supt.",
    // Courtesy plurals & French forms sometimes used in English
    "Messrs.", "Mme.", "Mlle.", "M."
];

/// All abbreviations that should not cause sentence splits  
#[cfg(test)]
pub const ABBREVIATIONS: &[&str] = &[
    "Dr.", "Mr.", "Mrs.", "Ms.", "Prof.", "Sr.", "Jr.",
    "U.S.A.", "U.K.", "N.Y.C.", "L.A.", "D.C.",
    "ft.", "in.", "lbs.", "oz.", "mi.", "km.",
    "a.m.", "p.m.", "etc.", "vs.", "ea.", "deg.", "et al."
];

/// Efficient abbreviation lookup using HashSet for O(1) performance
pub struct AbbreviationChecker {
    #[cfg(test)]
    abbreviations: HashSet<&'static str>,
    title_abbreviations: HashSet<&'static str>,
}

impl AbbreviationChecker {
    /// Create new abbreviation checker with default abbreviation sets
    pub fn new() -> Self {
        Self {
            #[cfg(test)]
            abbreviations: ABBREVIATIONS.iter().copied().collect(),
            title_abbreviations: TITLE_ABBREVIATIONS.iter().copied().collect(),
        }
    }

    /// Check if a word is a known abbreviation
    #[cfg(test)]
    pub fn is_abbreviation(&self, word: &str) -> bool {
        self.abbreviations.contains(word)
    }

    /// Check if a word is a title abbreviation (causes false positives with proper nouns)
    pub fn is_title_abbreviation(&self, word: &str) -> bool {
        self.title_abbreviations.contains(word)
    }

    /// Check if text ends with an abbreviation that should not split sentences
    /// WHY: Examines the last word in context to determine if punctuation is part of abbreviation
    #[cfg(test)]
    pub fn ends_with_abbreviation(&self, text: &str) -> bool {
        if let Some(last_word) = text.split_whitespace().last() {
            // Remove quotes and other punctuation to get clean word
            let clean_word = last_word.trim_matches(|c: char| {
                matches!(c, '"' | '\'' | '\u{201C}' | '\u{201D}' | '\u{2018}' | '\u{2019}')
            });
            self.is_abbreviation(clean_word)
        } else {
            false
        }
    }

    /// Check if text ends with a title abbreviation that could cause false positive
    /// WHY: Title abbreviations like "Dr." often precede proper nouns and should not split sentences
    pub fn ends_with_title_abbreviation(&self, text: &str) -> bool {
        if let Some(last_word) = text.split_whitespace().last() {
            // Remove quotes and other punctuation to get clean word
            let clean_word = last_word.trim_matches(|c: char| {
                matches!(c, '"' | '\'' | '\u{201C}' | '\u{201D}' | '\u{2018}' | '\u{2019}')
            });
            self.is_title_abbreviation(clean_word)
        } else {
            false
        }
    }
}

impl Default for AbbreviationChecker {
    fn default() -> Self {
        Self::new()
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use std::sync::OnceLock;
    
    // WHY: Single shared checker instance reduces test overhead
    static SHARED_CHECKER: OnceLock<AbbreviationChecker> = OnceLock::new();
    
    fn get_checker() -> &'static AbbreviationChecker {
        SHARED_CHECKER.get_or_init(AbbreviationChecker::new)
    }

    #[test]
    fn test_abbreviation_detection_comprehensive() {
        let checker = get_checker();
        
        // Test basic abbreviation detection
        let abbreviations = ["Dr.", "U.S.A.", "p.m.", "Prof.", "mi."];
        for abbr in &abbreviations {
            assert!(checker.is_abbreviation(abbr), "Should detect {abbr} as abbreviation");
        }
        assert!(!checker.is_abbreviation("Hello"));
        
        // Test title vs non-title abbreviation classification
        let title_abbreviations = ["Dr.", "Prof.", "Mr.", "Mrs."];
        for abbr in &title_abbreviations {
            assert!(checker.is_title_abbreviation(abbr), "Should detect {abbr} as title abbreviation");
        }
        assert!(!checker.is_title_abbreviation("U.S.A."));
        assert!(!checker.is_title_abbreviation("p.m."));
        
        // Test text ending detection
        let ending_tests = [
            ("Meeting at 5 p.m.", true),
            ("He lives in the U.S.A.", true),
            ("This is a sentence", false),
            ("Call Dr.", true),
            ("See Prof.", true),
        ];
        for (text, should_end_with_abbr) in &ending_tests {
            assert_eq!(checker.ends_with_abbreviation(text), *should_end_with_abbr, 
                "ends_with_abbreviation failed for: {text}");
        }
        
        // Test punctuation handling with quotes (only quotes are stripped, not other punctuation)
        let punctuation_tests = [
            ("He said 'Dr.'", true),
            ("She mentioned \"Prof.\"", true),
            ("Meeting at 5 p.m.?", false), // Question mark is not stripped, so "p.m.?" != "p.m."
        ];
        for (text, should_detect) in &punctuation_tests {
            assert_eq!(checker.ends_with_abbreviation(text), *should_detect,
                "Punctuation handling failed for: {text}");
        }
        
        // Test single capital letter abbreviations (PG 4300 fix)
        let single_capital_tests = [
            ("Listener, S.", true),
            ("latitude, N.", true), 
            ("longitude, W.", true),
            ("Direction E.", true),
            ("Point A.", true),
            ("Section Z.", true),
        ];
        for (text, should_detect) in &single_capital_tests {
            assert_eq!(checker.ends_with_title_abbreviation(text), *should_detect,
                "Single capital abbreviation failed for: {text}");
        }
    }
}