n2words 0.1.0

//! Gujarati (India) (`gu-IN`).
//!
//! - Indian numbering system (હજાર, લાખ, કરોડ)
//! - Gujarati script
//! - 3-2-2 grouping pattern (last 3 digits, then groups of 2)
//! - Complete word forms for 0-99
//! - Per-digit decimal reading
//! - Ordinals: irregular forms for 1-6, then the -મું suffix on the cardinal
//!
//! ```
//! use n2words::gu_in::{to_cardinal, to_currency, to_ordinal};
//!
//! assert_eq!(to_cardinal(42)?, "બેતાળીસ");
//! assert_eq!(to_ordinal(10)?, "દસમું");
//! assert_eq!(to_currency(42.5)?, "બેતાળીસ રૂપિયા પચાસ પૈસા");
//! # Ok::<(), n2words::Error>(())
//! ```

use crate::Result;
use crate::currency::vocab::GU as CURRENCY_VOCAB;
use crate::currency::{Currency, assert_minor_unit};
use crate::language::{Kind, Language, form};
use crate::max::Max;
use crate::number::{Number, parse_cardinal, parse_currency, parse_ordinal, split_groups_indian};
use crate::options::options;
use num_bigint::BigUint;
use num_traits::{One, ToPrimitive, Zero};

// ============================================================================
// Vocabulary
// ============================================================================

const ZERO: &str = "શૂન્ય";
const NEGATIVE: &str = "ઋણ";
const DECIMAL_SEP: &str = "દશાંશ";
const HUNDRED: &str = "સો";

// ============================================================================
// Ordinal vocabulary
// ============================================================================

/// Ordinal suffix (adds to cardinal for numbers >= 7).
const ORDINAL_SUFFIX: &str = "મું";

/// Special ordinals for the first few numbers (1-6 have irregular forms).
const ORDINAL_SPECIAL: [&str; 7] = ["", "પહેલું", "બીજું", "ત્રીજું", "ચોથું", "પાંચમું", "છઠ્ઠું"];

#[rustfmt::skip]
const BELOW_HUNDRED: [&str; 100] = [
    "શૂન્ય", "એક", "બે", "ત્રણ", "ચાર", "પાંચ", "છ", "સાત", "આઠ", "નવ",
    "દસ", "અગિયાર", "બાર", "તેર", "ચૌદ", "પંદર", "સોળ", "સત્તર", "અઢાર", "ઓગણીસ",
    "વીસ", "એકવીસ", "બાવીસ", "ત્રેવીસ", "ચોવીસ", "પચીસ", "છવ્વીસ", "સત્તાવીસ", "અઠ્ઠાવીસ", "ઓગણત્રીસ",
    "ત્રીસ", "એકત્રીસ", "બત્રીસ", "તેત્રીસ", "ચોત્રીસ", "પાંત્રીસ", "છત્રીસ", "સાડત્રીસ", "અડત્રીસ", "ઓગણચાલીસ",
    "ચાલીસ", "એકતાલીસ", "બેતાળીસ", "ત્રેતાળીસ", "ચુંમાલીસ", "પિસ્તાલીસ", "છેતાળીસ", "સુડતાળીસ", "અડતાળીસ", "ઓગણપચાસ",
    "પચાસ", "એકાવન", "બાવન", "ત્રેપન", "ચોપન", "પંચાવન", "છપ્પન", "સત્તાવન", "અઠ્ઠાવન", "ઓગણસાઠ",
    "સાઠ", "એકસઠ", "બાસઠ", "ત્રેસઠ", "ચોસઠ", "પાંસઠ", "છાસઠ", "સડસઠ", "અડસઠ", "અગણોસિત્તેર",
    "સિત્તેર", "એકોતેર", "બોતેર", "તોતેર", "ચુમોતેર", "પંચોતેર", "છોતેર", "સિત્યોતેર", "ઇઠ્યોતેર", "ઓગણાએંસી",
    "એંસી", "એક્યાસી", "બ્યાસી", "ત્યાસી", "ચોર્યાસી", "પંચાસી", "છ્યાસી", "સિત્યાસી", "અઠ્યાસી", "નેવ્યાસી",
    "નેવું", "એકાણું", "બાણું", "ત્રાણું", "ચોરાણું", "પંચાણું", "છન્નું", "સત્તાણું", "અઠ્ઠાણું", "નવ્વાણું",
];

/// Scale words: index 0 = units (empty), 1 = thousand, 2 = lakh, 3 = crore, etc.
const SCALE_WORDS: &[&str] = &["", "હજાર", "લાખ", "કરોડ", "અબજ", "ખરબ", "નીલ", "પદ્મ", "શંખ"];

// 3-2-2 Indian grouping: a 3-digit base segment, then 2 digits per scale word
// (SCALE_WORDS[0] = "" is the units slot). Past the table the scale word is
// dropped, which collapses the magnitude — so cap there.

/// The cardinal ceiling, derived from the scale table: `10^19`.
pub const CARDINAL_MAX: Max = Max::indian(SCALE_WORDS.len() as u32);
/// The ordinal ceiling: `10^19`.
pub const ORDINAL_MAX: Max = Max::indian(SCALE_WORDS.len() as u32);
/// The currency ceiling: `10^19`.
pub const CURRENCY_MAX: Max = Max::indian(SCALE_WORDS.len() as u32);

// ============================================================================
// Segment building
// ============================================================================

/// Words for a 0-999 segment.
fn build_segment(n: u32) -> String {
    if n == 0 {
        return String::new();
    }
    if n < 100 {
        return BELOW_HUNDRED[n as usize].to_string();
    }

    let hundreds = (n / 100) as usize;
    let remainder = (n % 100) as usize;

    if remainder == 0 {
        return format!("{} {HUNDRED}", BELOW_HUNDRED[hundreds]);
    }
    format!(
        "{} {HUNDRED} {}",
        BELOW_HUNDRED[hundreds], BELOW_HUNDRED[remainder]
    )
}

// ============================================================================
// Cardinal
// ============================================================================

/// Words for a non-negative integer.
///
/// South Asian 3-2-2 grouping: first 3 digits, then groups of 2.
fn integer_to_words(n: &BigUint) -> String {
    if n.is_zero() {
        return ZERO.to_string();
    }

    // Fast path: numbers < 1000 (direct lookup).
    if let Some(small) = n.to_u64() {
        if small < 1000 {
            return build_segment(small as u32);
        }
    }

    // First segment: last 3 digits; remaining segments: 2 digits each
    // (lakh = 100k, crore = 10M, etc.). Least-significant first.
    let segments = split_groups_indian(n, 1000, 100);

    // Build from the most-significant segment down.
    let mut result = String::new();
    for (i, &segment) in segments.iter().enumerate().rev() {
        if segment == 0 {
            continue;
        }
        if !result.is_empty() {
            result.push(' ');
        }
        if i == 0 {
            // The units segment can be 0-999.
            result.push_str(&build_segment(segment));
        } else {
            // Other segments are 0-99.
            result.push_str(BELOW_HUNDRED[segment as usize]);
        }
        // Scale word for every segment but the units one.
        if i > 0 {
            if let Some(scale) = SCALE_WORDS.get(i).filter(|s| !s.is_empty()) {
                result.push(' ');
                result.push_str(scale);
            }
        }
    }
    result
}

/// Words for the fractional digits, read one digit at a time.
fn decimal_part_to_words(decimal: &str) -> String {
    decimal
        .bytes()
        .map(|b| {
            let d = usize::from(b - b'0');
            if d == 0 { ZERO } else { BELOW_HUNDRED[d] }
        })
        .collect::<Vec<_>>()
        .join(" ")
}

/// Spells a number in Gujarati words.
///
/// ```
/// # use n2words::gu_in::to_cardinal;
/// assert_eq!(to_cardinal(1001)?, "એક હજાર એક");
/// assert_eq!(to_cardinal("100.05")?, "એક સો દશાંશ શૂન્ય પાંચ");
/// # Ok::<(), n2words::Error>(())
/// ```
///
/// # Errors
///
/// [`Error::InvalidNumber`](crate::Error::InvalidNumber) for a malformed
/// value, [`Error::OutOfRange`](crate::Error::OutOfRange) at or past
/// [`CARDINAL_MAX`].
pub fn to_cardinal<'a>(value: impl Into<Number<'a>>) -> Result<String> {
    let number = parse_cardinal(&value.into())?;
    // The fraction is spelled digit by digit, so only the integer part has a ceiling.
    CARDINAL_MAX.check(&number.integer, None)?;

    let mut result = String::new();
    if number.negative {
        result.push_str(NEGATIVE);
        result.push(' ');
    }
    result.push_str(&integer_to_words(&number.integer));
    if let Some(fraction) = &number.fraction {
        result.push(' ');
        result.push_str(DECIMAL_SEP);
        result.push(' ');
        result.push_str(&decimal_part_to_words(fraction));
    }
    Ok(result)
}

// ============================================================================
// Ordinal
// ============================================================================

/// Ordinal words for a positive integer: the first 6 are irregular, then the
/// -મું suffix is added to the cardinal.
fn integer_to_ordinal(n: &BigUint) -> String {
    if let Some(small) = n.to_u64() {
        if (1..=6).contains(&small) {
            return ORDINAL_SPECIAL[small as usize].to_string();
        }
    }
    let mut result = integer_to_words(n);
    result.push_str(ORDINAL_SUFFIX);
    result
}

/// Spells a positive integer as a Gujarati ordinal.
///
/// ```
/// # use n2words::gu_in::to_ordinal;
/// assert_eq!(to_ordinal(1)?, "પહેલું");
/// assert_eq!(to_ordinal(1001)?, "એક હજાર એકમું");
/// # Ok::<(), n2words::Error>(())
/// ```
///
/// # Errors
///
/// [`Error::InvalidOrdinal`](crate::Error::InvalidOrdinal) for zero, a
/// negative or a fraction; [`Error::OutOfRange`](crate::Error::OutOfRange) at
/// or past [`ORDINAL_MAX`].
pub fn to_ordinal<'a>(value: impl Into<Number<'a>>) -> Result<String> {
    let n = parse_ordinal(&value.into())?;
    // Ordinals build on the cardinal speller, so they share its ceiling.
    ORDINAL_MAX.check(&n, None)?;
    Ok(integer_to_ordinal(&n))
}

// ============================================================================
// Currency
// ============================================================================

options! {
    /// Options for [`to_currency_with`].
    pub struct CurrencyOptions {
        /// ISO 4217 currency code to name the amount in
        currency: Currency = Currency::INR, table = CURRENCY_VOCAB;
    }
}

/// Spells an amount in Indian rupees.
///
/// ```
/// # use n2words::gu_in::to_currency;
/// assert_eq!(to_currency(1)?, "એક રૂપિયો");
/// assert_eq!(to_currency(0.01)?, "એક પૈસો");
/// # Ok::<(), n2words::Error>(())
/// ```
///
/// # Errors
///
/// [`Error::InvalidNumber`](crate::Error::InvalidNumber) for a malformed
/// value, [`Error::OutOfRange`](crate::Error::OutOfRange) at or past
/// [`CURRENCY_MAX`].
pub fn to_currency<'a>(value: impl Into<Number<'a>>) -> Result<String> {
    to_currency_with(value, &CurrencyOptions::default())
}

/// Spells an amount in any currency Gujarati has words for.
///
/// # Errors
///
/// As [`to_currency`], plus
/// [`Error::UnsupportedCurrency`](crate::Error::UnsupportedCurrency) for a
/// currency Gujarati has no words for and
/// [`Error::NoMinorUnit`](crate::Error::NoMinorUnit) for a fractional amount
/// in a currency without one.
pub fn to_currency_with<'a>(
    value: impl Into<Number<'a>>,
    options: &CurrencyOptions,
) -> Result<String> {
    let words = CURRENCY_VOCAB.words(options.currency)?;
    let money = parse_currency(&value.into(), options.currency.minor_unit_digits())?;
    CURRENCY_MAX.check(&money.major, None)?;
    assert_minor_unit(money.minor, options.currency)?;

    let mut result = String::new();
    if money.negative {
        result.push_str(NEGATIVE);
        result.push(' ');
    }
    // Rupees part: shown if non-zero, or if there are no paise.
    if !money.major.is_zero() || money.minor == 0 {
        result.push_str(&integer_to_words(&money.major));
        result.push(' ');
        // Singular for 1 rupee, plural otherwise.
        result.push_str(words.major_at(if money.major.is_one() { 0 } else { 1 }));
    }
    // Paise part.
    if money.minor > 0 {
        if !money.major.is_zero() {
            result.push(' ');
        }
        result.push_str(&integer_to_words(&BigUint::from(money.minor)));
        result.push(' ');
        // Singular for 1 paisa, plural otherwise.
        result.push_str(words.minor_at(if money.minor == 1 { 0 } else { 1 }));
    }
    Ok(result)
}

/// Registry entry for `gu-IN`.
pub static LANGUAGE: Language = Language {
    code: "gu-IN",
    name: "Gujarati (India)",
    kind: Kind::Full,
    cardinal: Some(form!(plain to_cardinal, CARDINAL_MAX)),
    ordinal: Some(form!(plain to_ordinal, ORDINAL_MAX)),
    currency: Some(form!(with_options to_currency_with, CurrencyOptions, CURRENCY_MAX)),
};