chematic_iupac/

lib.rs

//! `chematic-iupac` — local IUPAC name generation, no network required.
//!
//! Supports:
//! - Linear alkanes and cycloalkanes
//! - Alkenes (`-ene`) and alkynes (`-yne`) with one unsaturation
//! - Simple derivatives: alcohols (`-ol`), amines (`-amine`), aldehydes (`-al`),
//!   ketones (`-one` with position locant), carboxylic acids (`-oic acid`)
//! - Esters (`alkyl alkanoate`) — linear, primary esters
//! - Primary/secondary amides (`-anamide`)
//! - Halogen substituents: fluoro-, chloro-, bromo-, iodo-
//! - Common aromatic heterocycles: benzene, pyridine, furan, thiophene,
//!   pyrrole, imidazole, pyrimidine
//!
//! Complex polycyclic systems, stereo descriptors, and structures outside
//! the above scope return [`IupacError::NotSupported`].

#![forbid(unsafe_code)]

use chematic_core::{AtomIdx, BondOrder, Molecule, implicit_hcount};
use chematic_perception::find_sssr;
use std::collections::{HashSet, VecDeque};

// ---------------------------------------------------------------------------
// Public error type
// ---------------------------------------------------------------------------

/// Error returned by [`name`].
#[derive(Debug, Clone, PartialEq, Eq)]
pub enum IupacError {
    /// The molecule contains no atoms.
    Empty,
    /// The molecule is outside the supported naming scope.
    NotSupported,
}

impl core::fmt::Display for IupacError {
    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
        match self {
            Self::Empty => write!(f, "empty molecule"),
            Self::NotSupported => write!(f, "IUPAC name not supported for this structure"),
        }
    }
}

impl std::error::Error for IupacError {}

// ---------------------------------------------------------------------------
// Public entry point
// ---------------------------------------------------------------------------

/// Generate a local IUPAC name for `mol`.
///
/// Returns `Err(IupacError::NotSupported)` for structures outside the current
/// scope (polycyclic systems, multi-functional groups, stereocenters, etc.).
pub fn name(mol: &Molecule) -> Result<String, IupacError> {
    if mol.atom_count() == 0 {
        return Err(IupacError::Empty);
    }
    Namer::new(mol).name()
}

// ---------------------------------------------------------------------------
// Internal namer
// ---------------------------------------------------------------------------

struct Namer<'a> {
    mol: &'a Molecule,
}

impl<'a> Namer<'a> {
    fn new(mol: &'a Molecule) -> Self {
        Self { mol }
    }

    fn name(&self) -> Result<String, IupacError> {
        let mol = self.mol;

        if count_components(mol) != 1 {
            return Err(IupacError::NotSupported);
        }

        let rings = find_sssr(mol);
        let ring_atoms: HashSet<AtomIdx> = rings
            .rings()
            .iter()
            .flat_map(|r| r.iter().copied())
            .collect();

        let carbons:  Vec<AtomIdx> = atoms_of(mol, 6);
        let o_atoms:  Vec<AtomIdx> = atoms_of(mol, 8);
        let n_atoms:  Vec<AtomIdx> = atoms_of(mol, 7);
        let s_atoms:  Vec<AtomIdx> = atoms_of(mol, 16);
        let halogens: Vec<AtomIdx> = mol
            .atoms()
            .filter(|(_, a)| matches!(a.element.atomic_number(), 9 | 17 | 35 | 53))
            .map(|(i, _)| i)
            .collect();

        // Reject elements outside C, H, N, O, S, halogens.
        let het_elements: HashSet<u8> = mol
            .atoms()
            .filter(|(_, a)| { let an = a.element.atomic_number(); an != 6 && an != 1 })
            .map(|(_, a)| a.element.atomic_number())
            .collect();
        if het_elements.iter().any(|&an| !matches!(an, 7 | 8 | 9 | 16 | 17 | 35 | 53)) {
            return Err(IupacError::NotSupported);
        }

        let cyclic = !ring_atoms.is_empty();

        if cyclic {
            let any_aromatic = ring_atoms.iter().any(|&i| mol.atom(i).aromatic);
            if any_aromatic {
                return self.name_aromatic_ring(&ring_atoms);
            }
            // Non-aromatic ring: allow OH substituent (cycloalkanol), block others.
            let only_oxygen = het_elements.len() == 1 && het_elements.contains(&8);
            if !het_elements.is_empty() && !only_oxygen {
                return Err(IupacError::NotSupported);
            }
            if only_oxygen {
                return self.name_cycloalkanol(&ring_atoms, &carbons, &o_atoms);
            }
            return self.name_cycloalkane(&ring_atoms, &carbons);
        }

        // Acyclic dispatch on heteroatom composition.
        match (o_atoms.len(), n_atoms.len(), s_atoms.len(), halogens.len()) {
            (0, 0, 0, 0) => self.name_acyclic_hydrocarbon(&carbons),
            (1, 0, 0, 0) => self.name_one_oxygen(&carbons, o_atoms[0]),
            (2, 0, 0, 0) => self.name_two_oxygens(&carbons, &o_atoms),
            (1, 1, 0, 0) => self.name_amide(&carbons, o_atoms[0], n_atoms[0]),
            (0, 1, 0, 0) => {
                // Nitrile (C≡N) takes priority over amine.
                if self.is_nitrile(n_atoms[0]) {
                    self.name_nitrile(&carbons, n_atoms[0])
                } else {
                    self.name_amine(&carbons, n_atoms[0])
                }
            }
            (0, 0, 1, 0) => self.name_thiol(&carbons, s_atoms[0]),
            (0, 0, 0, _) if !halogens.is_empty() => {
                if het_elements.len() != 1 {
                    return Err(IupacError::NotSupported);
                }
                let prefix = match het_elements.iter().next().copied().unwrap() {
                    9  => "fluoro",
                    17 => "chloro",
                    35 => "bromo",
                    53 => "iodo",
                    _  => return Err(IupacError::NotSupported),
                };
                self.name_haloalkane(&carbons, &halogens, prefix)
            }
            _ => Err(IupacError::NotSupported),
        }
    }

    // -----------------------------------------------------------------------
    // Aromatic ring naming
    // -----------------------------------------------------------------------

    fn name_aromatic_ring(&self, ring_atoms: &HashSet<AtomIdx>) -> Result<String, IupacError> {
        let mol = self.mol;
        // All ring atoms must be aromatic.
        if !ring_atoms.iter().all(|&i| mol.atom(i).aromatic) {
            return Err(IupacError::NotSupported);
        }

        let n_n = ring_atoms.iter().filter(|&&i| mol.atom(i).element.atomic_number() == 7).count();
        let n_o = ring_atoms.iter().filter(|&&i| mol.atom(i).element.atomic_number() == 8).count();
        let n_s = ring_atoms.iter().filter(|&&i| mol.atom(i).element.atomic_number() == 16).count();
        let sz  = ring_atoms.len();

        // Case 1: Pure aromatic ring (no substituents).
        if ring_atoms.len() == mol.atom_count() {
            return match (sz, n_n, n_o, n_s) {
                (6, 0, 0, 0) => Ok("benzene".into()),
                (6, 1, 0, 0) => Ok("pyridine".into()),
                (6, 2, 0, 0) => Ok("pyrimidine".into()),
                (5, 0, 1, 0) => Ok("furan".into()),
                (5, 0, 0, 1) => Ok("thiophene".into()),
                (5, 1, 0, 0) => Ok("pyrrole".into()),
                (5, 2, 0, 0) => Ok("imidazole".into()),
                _            => Err(IupacError::NotSupported),
            };
        }

        // Case 2: Monosubstituted benzene (phenol, toluene, aniline, etc.)
        // Only support pure benzene ring (6 C, no N/O/S in ring).
        if sz == 6 && n_n == 0 && n_o == 0 && n_s == 0 {
            let sub_atoms: Vec<AtomIdx> = mol.atoms()
                .filter(|(i, _)| !ring_atoms.contains(i))
                .map(|(i, _)| i)
                .collect();
            return self.name_monosubstituted_benzene(ring_atoms, &sub_atoms);
        }

        Err(IupacError::NotSupported)
    }

    // -----------------------------------------------------------------------
    // Monosubstituted benzene naming
    // -----------------------------------------------------------------------

    fn name_monosubstituted_benzene(
        &self,
        ring_atoms: &HashSet<AtomIdx>,
        sub_atoms: &[AtomIdx],
    ) -> Result<String, IupacError> {
        let mol = self.mol;
        // Count how many ring C have substituents.
        let attach_count = ring_atoms.iter().filter(|&&r| {
            mol.neighbors(r).any(|(nb, _)| !ring_atoms.contains(&nb))
        }).count();
        if attach_count == 2 {
            return self.name_disubstituted_benzene(ring_atoms, sub_atoms);
        }
        if attach_count == 3 {
            return self.name_trisubstituted_benzene(ring_atoms);
        }
        if attach_count != 1 {
            return Err(IupacError::NotSupported);
        }

        // Classify substituent by element counts + bond types.
        let mut n_c = 0usize; let mut n_n = 0usize;
        let mut n_o = 0usize; let mut n_hal = 0usize;
        let mut halogen_an = 0u8;
        for &a in sub_atoms {
            match mol.atom(a).element.atomic_number() {
                6  => n_c += 1,
                7  => n_n += 1,
                8  => n_o += 1,
                1  => {},
                an @ (9 | 17 | 35 | 53) => { n_hal += 1; halogen_an = an; }
                _  => return Err(IupacError::NotSupported),
            }
        }

        let sub_set: HashSet<AtomIdx> = sub_atoms.iter().copied().collect();
        let has_triple = mol.bonds().any(|(_, b)| {
            b.order == BondOrder::Triple
                && (sub_set.contains(&b.atom1) || sub_set.contains(&b.atom2))
        });
        let has_double = mol.bonds().any(|(_, b)| {
            b.order == BondOrder::Double
                && (sub_set.contains(&b.atom1) || sub_set.contains(&b.atom2))
        });

        match (n_c, n_n, n_o, n_hal, has_double, has_triple) {
            // Phenol: c1ccccc1O
            (0, 0, 1, 0, false, false) => Ok("phenol".into()),
            // Aniline: c1ccccc1N
            (0, 1, 0, 0, false, false) => Ok("aniline".into()),
            // Halo-benzenes
            (0, 0, 0, 1, false, false) => {
                let prefix = match halogen_an {
                    9 => "fluoro", 17 => "chloro", 35 => "bromo", 53 => "iodo", _ => return Err(IupacError::NotSupported),
                };
                Ok(format!("{prefix}benzene"))
            }
            // Toluene: c1ccccc1C (one CH3)
            (1, 0, 0, 0, false, false) => Ok("toluene".into()),
            // Benzaldehyde: c1ccccc1C=O (n_c=1, n_o=1, has_double)
            (1, 0, 1, 0, true, false) => Ok("benzaldehyde".into()),
            // Benzoic acid: c1ccccc1C(=O)O (n_c=1, n_o=2, has_double)
            (1, 0, 2, 0, true, false) => Ok("benzoic acid".into()),
            // Benzonitrile: c1ccccc1C#N (n_c=1, n_n=1, has_triple)
            (1, 1, 0, 0, false, true) => Ok("benzonitrile".into()),
            _ => Err(IupacError::NotSupported),
        }
    }

    // -----------------------------------------------------------------------
    // Cycloalkane naming
    // -----------------------------------------------------------------------

    fn name_cycloalkane(
        &self,
        ring_atoms: &HashSet<AtomIdx>,
        carbons: &[AtomIdx],
    ) -> Result<String, IupacError> {
        let mol = self.mol;
        if carbons.iter().any(|&c| mol.atom(c).aromatic) {
            return Err(IupacError::NotSupported);
        }
        // All carbons in ring: unsubstituted cycloalkane.
        if ring_atoms.len() == carbons.len() {
            return Ok(format!("cyclo{}", alkane_suffix(ring_atoms.len())));
        }
        let outside: Vec<AtomIdx> = carbons.iter()
            .filter(|&&c| !ring_atoms.contains(&c))
            .copied()
            .collect();

        let is_terminal_methyl = |sub: AtomIdx| -> bool {
            mol.neighbors(sub)
                .filter(|(nb, _)| mol.atom(*nb).element.atomic_number() == 6 && !ring_atoms.contains(nb))
                .count() == 0
        };

        if outside.len() == 1 && is_terminal_methyl(outside[0]) {
            return Ok(format!("methylcyclo{}", alkane_suffix(ring_atoms.len())));
        }

        if outside.len() == 2 && is_terminal_methyl(outside[0]) && is_terminal_methyl(outside[1]) {
            let att_a = mol.neighbors(outside[0])
                .find(|(nb, _)| ring_atoms.contains(nb))
                .map(|(nb, _)| nb)
                .ok_or(IupacError::NotSupported)?;
            let att_b = mol.neighbors(outside[1])
                .find(|(nb, _)| ring_atoms.contains(nb))
                .map(|(nb, _)| nb)
                .ok_or(IupacError::NotSupported)?;
            // BFS shortest path within ring.
            let raw_dist = {
                let mut dist = 0usize;
                let mut queue = VecDeque::new();
                let mut visited: HashSet<AtomIdx> = HashSet::new();
                queue.push_back((att_a, 0usize));
                visited.insert(att_a);
                'bfs: while let Some((cur, d)) = queue.pop_front() {
                    if cur == att_b { dist = d; break 'bfs; }
                    for (nb, _) in mol.neighbors(cur) {
                        if ring_atoms.contains(&nb) && visited.insert(nb) {
                            queue.push_back((nb, d + 1));
                        }
                    }
                }
                dist
            };
            let ring_dist = raw_dist.min(ring_atoms.len() - raw_dist);
            return Ok(format!("1,{}-dimethylcyclo{}", ring_dist + 1, alkane_suffix(ring_atoms.len())));
        }

        Err(IupacError::NotSupported)
    }

    // -----------------------------------------------------------------------
    // Cycloalkanol naming (cyclopentanol, cyclohexanol, ...)
    // -----------------------------------------------------------------------

    fn name_cycloalkanol(
        &self,
        ring_atoms: &HashSet<AtomIdx>,
        carbons: &[AtomIdx],
        o_atoms: &[AtomIdx],
    ) -> Result<String, IupacError> {
        let mol = self.mol;
        // Only one OH substituent.
        if o_atoms.len() != 1 { return Err(IupacError::NotSupported); }
        let o_idx = o_atoms[0];
        // O must be single-bond –OH (not carbonyl).
        if mol.neighbors(o_idx).any(|(_, bi)| mol.bond(bi).order == BondOrder::Double) {
            return Err(IupacError::NotSupported);
        }
        // O must have implicit H.
        if implicit_hcount(mol, o_idx) == 0 {
            return Err(IupacError::NotSupported);
        }
        // No exocyclic carbons (unsubstituted ring + OH only).
        let exo_c = carbons.iter().filter(|&&c| !ring_atoms.contains(&c)).count();
        if exo_c > 0 { return Err(IupacError::NotSupported); }
        Ok(format!("cyclo{}ol", alkane_base(ring_atoms.len())))
    }

    // -----------------------------------------------------------------------
    // Acyclic hydrocarbon naming
    // -----------------------------------------------------------------------

    fn name_acyclic_hydrocarbon(&self, carbons: &[AtomIdx]) -> Result<String, IupacError> {
        let mol = self.mol;
        let n = carbons.len();

        let double_bonds = mol.bonds().filter(|(_, b)| b.order == BondOrder::Double).count();
        let triple_bonds = mol.bonds().filter(|(_, b)| b.order == BondOrder::Triple).count();
        if double_bonds > 1 || triple_bonds > 1 || (double_bonds > 0 && triple_bonds > 0) {
            return Err(IupacError::NotSupported);
        }

        // Check for branching.
        let c_set: HashSet<AtomIdx> = carbons.iter().copied().collect();
        let is_branched = carbons.iter().any(|&c| {
            mol.neighbors(c).filter(|(nb, _)| c_set.contains(nb)).count() > 2
        });

        if is_branched {
            // Only saturated branched alkanes supported for now.
            if double_bonds > 0 || triple_bonds > 0 {
                return Err(IupacError::NotSupported);
            }
            return self.name_branched_alkane(carbons);
        }

        if triple_bonds == 1 {
            if n >= 4 {
                let pos = unsaturation_locant(mol, carbons, BondOrder::Triple);
                Ok(format!("{}-{}-yne", alkane_stem(n), pos))
            } else {
                Ok(alkyne_suffix(n))
            }
        } else if double_bonds == 1 {
            if n >= 4 {
                let pos = unsaturation_locant(mol, carbons, BondOrder::Double);
                Ok(format!("{}-{}-ene", alkane_stem(n), pos))
            } else {
                Ok(alkene_suffix(n))
            }
        } else {
            Ok(alkane_suffix(n))
        }
    }

    // -----------------------------------------------------------------------
    // One-oxygen compound: alcohol / aldehyde / ketone
    // -----------------------------------------------------------------------

    fn name_one_oxygen(&self, carbons: &[AtomIdx], o_idx: AtomIdx) -> Result<String, IupacError> {
        let mol = self.mol;
        let is_double = mol.neighbors(o_idx).any(|(_, bi)| mol.bond(bi).order == BondOrder::Double);

        if !is_double {
            // Ether check: O with 2 C neighbors and no implicit H → R-O-R
            let o_c_nb: Vec<AtomIdx> = mol.neighbors(o_idx)
                .filter(|(nb, _)| mol.atom(*nb).element.atomic_number() == 6)
                .map(|(nb, _)| nb)
                .collect();
            if o_c_nb.len() == 2 && implicit_hcount(mol, o_idx) == 0 {
                return self.name_ether(carbons, o_idx, o_c_nb[0], o_c_nb[1]);
            }

            // Find the OH carbon.
            let oh_c = mol.neighbors(o_idx)
                .filter(|(nb, _)| mol.atom(*nb).element.atomic_number() == 6)
                .map(|(nb, _)| nb)
                .next()
                .ok_or(IupacError::NotSupported)?;

            // Check for branching.
            let c_set: HashSet<AtomIdx> = carbons.iter().copied().collect();
            let is_branched = carbons.iter().any(|&c| {
                mol.neighbors(c).filter(|(nb, _)| c_set.contains(nb)).count() > 2
            });
            if is_branched {
                return self.name_branched_alcohol(carbons, oh_c);
            }

            // Straight-chain: determine OH position using longest chain.
            let chain = find_longest_c_chain(mol, carbons);
            let n = chain.len();
            let pos_fwd = chain.iter().position(|&c| c == oh_c).map(|p| p + 1).unwrap_or(1);
            let pos = pos_fwd.min(n + 1 - pos_fwd);
            if pos == 1 && n <= 2 {
                // Short common names without locant: methanol, ethanol.
                return Ok(format!("{}anol", alkane_stem(n)));
            }
            return Ok(format!("{}-{}-ol", alkane_base(n), pos));
        }

        // Carbonyl: find the C=O carbon.
        let carbonyl_c = mol
            .neighbors(o_idx)
            .filter(|(nb, _)| mol.atom(*nb).element.atomic_number() == 6)
            .map(|(nb, _)| nb)
            .next()
            .ok_or(IupacError::NotSupported)?;

        if implicit_hcount(mol, carbonyl_c) > 0 {
            // Aldehyde: CHO is position 1; find chain from carbonyl_c.
            let c_set: HashSet<AtomIdx> = carbons.iter().copied().collect();
            let chain = chain_from_anchor(mol, &c_set, carbonyl_c);
            let n = chain.len();
            let chain_set: HashSet<AtomIdx> = chain.iter().copied().collect();
            let mut subs: Vec<(usize, usize)> = Vec::new();
            for (pos0, &chain_c) in chain.iter().enumerate() {
                if pos0 == 0 { continue; }
                let position = pos0 + 1;
                for (nb, _) in mol.neighbors(chain_c) {
                    if c_set.contains(&nb) && !chain_set.contains(&nb) {
                        let sub_len = count_c_chain(mol, nb, chain_c);
                        if sub_len > 4 { return Err(IupacError::NotSupported); }
                        subs.push((position, sub_len));
                    }
                }
            }
            let prefix = if subs.is_empty() { String::new() } else { format_substituents(&subs) };
            return Ok(format!("{}{}anal", prefix, alkane_stem(n)));
        }

        // Ketone: internal C=O — find principal chain and position.
        let chain = find_longest_c_chain(mol, carbons);
        let n = chain.len();
        if n < 3 { return Err(IupacError::NotSupported); }
        let chain_set: HashSet<AtomIdx> = chain.iter().copied().collect();
        let all_c_set: HashSet<AtomIdx> = carbons.iter().copied().collect();
        let pos_fwd = chain.iter().position(|&c| c == carbonyl_c)
            .map(|p| p + 1).ok_or(IupacError::NotSupported)?;
        let pos = pos_fwd.min(n + 1 - pos_fwd);
        let reversed = pos_fwd > n + 1 - pos_fwd;
        // Collect alkyl substituents on the chain.
        let mut subs: Vec<(usize, usize)> = Vec::new();
        for (idx, &chain_c) in chain.iter().enumerate() {
            let position = idx + 1;
            for (nb, _) in mol.neighbors(chain_c) {
                if all_c_set.contains(&nb) && !chain_set.contains(&nb) {
                    let sub_len = count_c_chain(mol, nb, chain_c);
                    if sub_len > 4 { return Err(IupacError::NotSupported); }
                    let adj_pos = if reversed { n + 1 - position } else { position };
                    subs.push((adj_pos, sub_len));
                }
            }
        }
        let prefix = if subs.is_empty() { String::new() } else { format_substituents(&subs) };
        Ok(format!("{}{}-{}-one", prefix, alkane_base(n), pos))
    }

    // -----------------------------------------------------------------------
    // Ether naming (R-O-R → "alkoxyalkane")
    // -----------------------------------------------------------------------

    fn name_ether(
        &self,
        carbons: &[AtomIdx],
        o_idx: AtomIdx,
        side_a: AtomIdx,
        side_b: AtomIdx,
    ) -> Result<String, IupacError> {
        let mol = self.mol;
        // Only unbranched ethers.
        let c_set: HashSet<AtomIdx> = carbons.iter().copied().collect();
        if carbons.iter().any(|&c| {
            mol.neighbors(c).filter(|(nb, _)| c_set.contains(nb)).count() > 2
        }) {
            return Err(IupacError::NotSupported);
        }
        let len_a = count_c_chain(mol, side_a, o_idx);
        let len_b = count_c_chain(mol, side_b, o_idx);
        let (alkoxy_len, parent_len) = if len_a <= len_b { (len_a, len_b) } else { (len_b, len_a) };
        let alkoxy = format!("{}oxy", alkane_stem(alkoxy_len));
        let parent = alkane_suffix(parent_len);
        // Add locant "1-" when parent ≥ 3 C and chains differ (O position is ambiguous).
        if parent_len >= 3 && alkoxy_len != parent_len {
            Ok(format!("1-{alkoxy}{parent}"))
        } else {
            Ok(format!("{alkoxy}{parent}"))
        }
    }

    // -----------------------------------------------------------------------
    // Two-oxygen compound: carboxylic acid or ester
    // -----------------------------------------------------------------------

    fn name_two_oxygens(&self, carbons: &[AtomIdx], o_atoms: &[AtomIdx]) -> Result<String, IupacError> {
        let mol = self.mol;
        let o1 = o_atoms[0];
        let o2 = o_atoms[1];

        let o1_dbl = mol.neighbors(o1).any(|(_, bi)| mol.bond(bi).order == BondOrder::Double);
        let o2_dbl = mol.neighbors(o2).any(|(_, bi)| mol.bond(bi).order == BondOrder::Double);

        let (carbonyl_o, ester_o) = match (o1_dbl, o2_dbl) {
            (true, false) => (o1, o2),
            (false, true) => (o2, o1),
            _ => return Err(IupacError::NotSupported),
        };

        // Carbonyl C is bonded to the =O oxygen.
        let carbonyl_c = mol
            .neighbors(carbonyl_o)
            .filter(|(nb, _)| mol.atom(*nb).element.atomic_number() == 6)
            .map(|(nb, _)| nb)
            .next()
            .ok_or(IupacError::NotSupported)?;

        // Carbonyl C must also be bonded to the single-bond O.
        if !mol.neighbors(carbonyl_c).any(|(nb, _)| nb == ester_o) {
            return Err(IupacError::NotSupported);
        }

        // Is the single-bond O also bonded to another C (→ ester) or only H (→ acid)?
        let alcohol_c = mol
            .neighbors(ester_o)
            .filter(|(nb, _)| *nb != carbonyl_c && mol.atom(*nb).element.atomic_number() == 6)
            .map(|(nb, _)| nb)
            .next();

        let c_set: HashSet<AtomIdx> = carbons.iter().copied().collect();
        if let Some(alc_c) = alcohol_c {
            // Ester: find acid chain from carbonyl_c (handles branched acid parts).
            let chain_acid = chain_from_anchor(mol, &c_set, carbonyl_c);
            let acid_n = chain_acid.len();
            let chain_acid_set: HashSet<AtomIdx> = chain_acid.iter().copied().collect();
            let mut subs: Vec<(usize, usize)> = Vec::new();
            for (pos0, &chain_c) in chain_acid.iter().enumerate() {
                if pos0 == 0 { continue; }
                let position = pos0 + 1;
                for (nb, _) in mol.neighbors(chain_c) {
                    if c_set.contains(&nb) && !chain_acid_set.contains(&nb) {
                        let sub_len = count_c_chain(mol, nb, chain_c);
                        if sub_len > 4 { return Err(IupacError::NotSupported); }
                        subs.push((position, sub_len));
                    }
                }
            }
            let alcohol_n = count_c_chain(mol, alc_c, ester_o);
            let acid_part = if subs.is_empty() {
                format!("{}anoate", alkane_stem(acid_n))
            } else {
                format!("{}{}anoate", format_substituents(&subs), alkane_stem(acid_n))
            };
            Ok(format!("{}yl {}", alkane_stem(alcohol_n), acid_part))
        } else {
            // Carboxylic acid — find principal chain from carboxyl C (always position 1).
            let chain = chain_from_anchor(mol, &c_set, carbonyl_c);
            let n = chain.len();
            let chain_set: HashSet<AtomIdx> = chain.iter().copied().collect();
            let mut subs: Vec<(usize, usize)> = Vec::new();
            for (pos0, &chain_c) in chain.iter().enumerate() {
                if pos0 == 0 { continue; }
                let position = pos0 + 1;
                for (nb, _) in mol.neighbors(chain_c) {
                    if c_set.contains(&nb) && !chain_set.contains(&nb) {
                        let sub_len = count_c_chain(mol, nb, chain_c);
                        if sub_len > 4 { return Err(IupacError::NotSupported); }
                        subs.push((position, sub_len));
                    }
                }
            }
            if subs.is_empty() {
                Ok(format!("{}anoic acid", alkane_stem(n)))
            } else {
                Ok(format!("{}{}anoic acid", format_substituents(&subs), alkane_stem(n)))
            }
        }
    }

    // -----------------------------------------------------------------------
    // Amide: C(=O)–N
    // -----------------------------------------------------------------------

    fn name_amide(
        &self,
        _carbons: &[AtomIdx],
        o_idx: AtomIdx,
        n_idx: AtomIdx,
    ) -> Result<String, IupacError> {
        let mol = self.mol;

        // O must be a carbonyl (C=O).
        if !mol.neighbors(o_idx).any(|(_, bi)| mol.bond(bi).order == BondOrder::Double) {
            return Err(IupacError::NotSupported);
        }

        let carbonyl_c = mol
            .neighbors(o_idx)
            .filter(|(nb, _)| mol.atom(*nb).element.atomic_number() == 6)
            .map(|(nb, _)| nb)
            .next()
            .ok_or(IupacError::NotSupported)?;

        // Carbonyl C must be bonded to N.
        if !mol.neighbors(carbonyl_c).any(|(nb, _)| nb == n_idx) {
            return Err(IupacError::NotSupported);
        }

        // Only primary/secondary amides (N has ≥ 1 H).
        if implicit_hcount(mol, n_idx) == 0 {
            return Err(IupacError::NotSupported);
        }

        // Amide chain from carbonyl_c (handles branched structures).
        let c_set: HashSet<AtomIdx> = mol.atoms()
            .filter(|(_, a)| a.element.atomic_number() == 6)
            .map(|(i, _)| i)
            .collect();
        let chain = chain_from_anchor(mol, &c_set, carbonyl_c);
        let n = chain.len();
        let chain_set: HashSet<AtomIdx> = chain.iter().copied().collect();
        let mut subs: Vec<(usize, usize)> = Vec::new();
        for (pos0, &chain_c) in chain.iter().enumerate() {
            if pos0 == 0 { continue; }
            let position = pos0 + 1;
            for (nb, _) in mol.neighbors(chain_c) {
                if c_set.contains(&nb) && !chain_set.contains(&nb) {
                    let sub_len = count_c_chain(mol, nb, chain_c);
                    if sub_len > 4 { return Err(IupacError::NotSupported); }
                    subs.push((position, sub_len));
                }
            }
        }
        let prefix = if subs.is_empty() { String::new() } else { format_substituents(&subs) };
        Ok(format!("{}{}anamide", prefix, alkane_stem(n)))
    }

    // -----------------------------------------------------------------------
    // Amine naming
    // -----------------------------------------------------------------------

    fn name_amine(&self, carbons: &[AtomIdx], n_idx: AtomIdx) -> Result<String, IupacError> {
        let mol = self.mol;
        let n_h = implicit_hcount(mol, n_idx);
        let c_sides: Vec<AtomIdx> = mol.neighbors(n_idx)
            .filter(|(nb, _)| mol.atom(*nb).element.atomic_number() == 6)
            .map(|(nb, _)| nb)
            .collect();
        let mut chain_lens: Vec<usize> = c_sides.iter()
            .map(|&nb| count_c_chain(mol, nb, n_idx))
            .collect();
        chain_lens.sort_unstable_by(|a, b| b.cmp(a)); // descending
        match n_h {
            2 => {
                // Find N-bearing C's position on the principal chain.
                let chain = find_longest_c_chain(mol, carbons);
                let n_chain = chain.len();
                let chain_set: HashSet<AtomIdx> = chain.iter().copied().collect();
                let amine_c = mol.neighbors(n_idx)
                    .filter(|(nb, _)| mol.atom(*nb).element.atomic_number() == 6
                                   && chain_set.contains(nb))
                    .map(|(nb, _)| nb)
                    .next()
                    .ok_or(IupacError::NotSupported)?;
                let pos_fwd = chain.iter().position(|&c| c == amine_c)
                    .map(|p| p + 1).unwrap_or(1);
                let pos = pos_fwd.min(n_chain + 1 - pos_fwd);
                Ok(format!("{}an-{}-amine", alkane_stem(n_chain), pos))
            }
            1 => {
                if chain_lens.len() != 2 { return Err(IupacError::NotSupported); }
                let parent_len = chain_lens[0];
                let sub_len    = chain_lens[1];
                Ok(format!("N-{}yl{}anamine", alkane_stem(sub_len), alkane_stem(parent_len)))
            }
            0 => {
                if chain_lens.len() != 3 { return Err(IupacError::NotSupported); }
                let parent_len = chain_lens[0];
                let sub1 = chain_lens[1];
                let sub2 = chain_lens[2];
                if sub1 == sub2 {
                    Ok(format!("N,N-di{}yl{}anamine", alkane_stem(sub1), alkane_stem(parent_len)))
                } else {
                    let (lo, hi) = (sub1.min(sub2), sub1.max(sub2));
                    Ok(format!("N-{}yl-N-{}yl{}anamine", alkane_stem(lo), alkane_stem(hi), alkane_stem(parent_len)))
                }
            }
            _ => Err(IupacError::NotSupported),
        }
    }

    // -----------------------------------------------------------------------
    // Haloalkane naming
    // -----------------------------------------------------------------------

    fn name_haloalkane(
        &self,
        carbons: &[AtomIdx],
        halogen_atoms: &[AtomIdx],
        prefix: &str,
    ) -> Result<String, IupacError> {
        let mol = self.mol;
        let chain = find_longest_c_chain(mol, carbons);
        let n = chain.len();
        let chain_set: HashSet<AtomIdx> = chain.iter().copied().collect();

        // Find the locant of each halogen on the chain.
        let mut locants: Vec<usize> = Vec::new();
        for &hal in halogen_atoms {
            let hal_c = mol.neighbors(hal)
                .filter(|(nb, _)| chain_set.contains(nb))
                .map(|(nb, _)| nb)
                .next()
                .ok_or(IupacError::NotSupported)?;
            let pos = chain.iter().position(|&c| c == hal_c).map(|p| p + 1)
                .ok_or(IupacError::NotSupported)?;
            locants.push(pos);
        }

        // Apply lowest-locant rule (compare forward vs reversed numbering).
        let locants_rev: Vec<usize> = locants.iter().map(|&p| n + 1 - p).collect();
        let best = if locants.iter().min() <= locants_rev.iter().min() {
            locants
        } else {
            locants_rev
        };

        let count = halogen_atoms.len();
        let mult = match count {
            1 => prefix.to_string(),
            2 => format!("di{prefix}"),
            3 => format!("tri{prefix}"),
            _ => return Err(IupacError::NotSupported),
        };

        let mut sorted_locs = best;
        sorted_locs.sort_unstable();
        let locant_str = sorted_locs.iter().map(|l| l.to_string()).collect::<Vec<_>>().join(",");

        // Omit locant for short unambiguous cases (n≤2, single halogen at terminal).
        if n <= 2 && count == 1 {
            Ok(format!("{mult}{}", alkane_suffix(n)))
        } else {
            Ok(format!("{locant_str}-{mult}{}", alkane_suffix(n)))
        }
    }

    // -----------------------------------------------------------------------
    // Thiol naming (R-SH → "...anethiol")
    // -----------------------------------------------------------------------

    fn name_thiol(&self, carbons: &[AtomIdx], s_idx: AtomIdx) -> Result<String, IupacError> {
        let mol = self.mol;
        if implicit_hcount(mol, s_idx) == 0 {
            return Err(IupacError::NotSupported);
        }
        let chain = find_longest_c_chain(mol, carbons);
        let n = chain.len();
        let chain_set: HashSet<AtomIdx> = chain.iter().copied().collect();
        let thiol_c = mol.neighbors(s_idx)
            .filter(|(nb, _)| chain_set.contains(nb))
            .map(|(nb, _)| nb)
            .next()
            .ok_or(IupacError::NotSupported)?;
        let pos_fwd = chain.iter().position(|&c| c == thiol_c).map(|p| p + 1).unwrap_or(1);
        let pos = pos_fwd.min(n + 1 - pos_fwd);
        // Terminal SH (pos=1): no locant; internal: add locant.
        if pos == 1 {
            Ok(format!("{}anethiol", alkane_stem(n)))
        } else {
            Ok(format!("{}ane-{}-thiol", alkane_stem(n), pos))
        }
    }

    // -----------------------------------------------------------------------
    // Branched alcohol naming (e.g., "propan-2-ol")
    // -----------------------------------------------------------------------

    fn name_branched_alcohol(
        &self,
        carbons: &[AtomIdx],
        oh_c: AtomIdx,
    ) -> Result<String, IupacError> {
        // Find principal chain.
        let chain = find_longest_c_chain(self.mol, carbons);
        let n = chain.len();
        if n < 2 { return Err(IupacError::NotSupported); }

        let chain_set: HashSet<AtomIdx> = chain.iter().copied().collect();
        let all_c_set: HashSet<AtomIdx> = carbons.iter().copied().collect();

        // The OH carbon must be on the principal chain.
        let pos_on_chain = if chain_set.contains(&oh_c) {
            chain.iter().position(|&c| c == oh_c).map(|p| p + 1)
        } else {
            None
        };

        let pos_fwd = pos_on_chain.ok_or(IupacError::NotSupported)?;
        let pos = pos_fwd.min(n + 1 - pos_fwd);

        // Also collect any alkyl substituents on the chain.
        let mut subs: Vec<(usize, usize)> = Vec::new();
        for (pos0, &chain_c) in chain.iter().enumerate() {
            let position = pos0 + 1;
            for (nb, _) in self.mol.neighbors(chain_c) {
                if all_c_set.contains(&nb) && !chain_set.contains(&nb) {
                    let sub_len = count_c_chain(self.mol, nb, chain_c);
                    if sub_len > 4 { return Err(IupacError::NotSupported); }
                    subs.push((position, sub_len));
                }
            }
        }

        // Re-number subs with the same locant direction as OH position.
        if pos_fwd > n + 1 - pos_fwd {
            // Reverse direction was chosen for OH; re-number subs accordingly.
            subs = subs.iter().map(|&(p, l)| (n + 1 - p, l)).collect();
        }

        let prefix = if subs.is_empty() {
            String::new()
        } else {
            subs.sort_unstable();
            let subs_rev: Vec<(usize, usize)> = subs.iter()
                .map(|&(p, l)| (n + 1 - p, l))
                .collect();
            let first_fwd = subs.iter().map(|&(p, _)| p).min().unwrap_or(usize::MAX);
            let first_rev = subs_rev.iter().map(|&(p, _)| p).min().unwrap_or(usize::MAX);
            let best = if first_fwd <= first_rev { subs.clone() } else { subs_rev };
            format!("{}-", format_substituents(&best))
        };

        Ok(format!("{}{}-{}-ol", prefix, alkane_base(n), pos))
    }

    // -----------------------------------------------------------------------
    // Disubstituted benzene naming (e.g., "4-chlorophenol")
    // -----------------------------------------------------------------------

    fn name_disubstituted_benzene(
        &self,
        ring_atoms: &HashSet<AtomIdx>,
        _sub_atoms: &[AtomIdx],
    ) -> Result<String, IupacError> {
        let mol = self.mol;

        // Identify the two ring C attachment points and their substituent sets.
        let attach_points: Vec<AtomIdx> = ring_atoms.iter()
            .filter(|&&r| mol.neighbors(r).any(|(nb, _)| !ring_atoms.contains(&nb)))
            .copied()
            .collect();
        if attach_points.len() != 2 {
            return Err(IupacError::NotSupported);
        }

        // Compute ring distance (shortest path within ring) between the two attachment points.
        let ring_dist = {
            let ring_vec: Vec<AtomIdx> = ring_atoms.iter().copied().collect();
            let mut dist = usize::MAX;
            // BFS within the ring
            let mut queue = VecDeque::new();
            let mut visited: HashSet<AtomIdx> = HashSet::new();
            queue.push_back((attach_points[0], 0usize));
            visited.insert(attach_points[0]);
            while let Some((cur, d)) = queue.pop_front() {
                if cur == attach_points[1] { dist = d; break; }
                for (nb, _) in mol.neighbors(cur) {
                    if ring_atoms.contains(&nb) && visited.insert(nb) {
                        queue.push_back((nb, d + 1));
                    }
                }
            }
            // Take minimum of this and the longer path
            dist.min(ring_vec.len() - dist)
        };

        // Classify each substituent group.
        let classify_sub = |attach: AtomIdx| -> Option<(&str, bool)> {
            // Returns (substituent_name, is_principal)
            // is_principal: true if this substituent determines the compound root name
            // Collect sub atoms for this attachment (unused for now, just for documentation)
            // Simple: just look at atoms directly bonded to attach that are not in ring
            let direct: Vec<AtomIdx> = mol.neighbors(attach)
                .filter(|(nb, _)| !ring_atoms.contains(nb))
                .map(|(nb, _)| nb)
                .collect();
            if direct.is_empty() { return None; }
            let first = direct[0];
            let an = mol.atom(first).element.atomic_number();
            match an {
                8 if !mol.neighbors(first).any(|(_, bi)| mol.bond(bi).order == BondOrder::Double) => {
                    Some(("hydroxy", true)) // -OH → phenol as principal
                }
                7 if implicit_hcount(mol, first) > 0 => Some(("amino", true)), // -NH2 → aniline
                6 => Some(("methyl", false)), // -CH3 → toluene substituent
                17 => Some(("chloro", false)),
                35 => Some(("bromo", false)),
                9  => Some(("fluoro", false)),
                53 => Some(("iodo", false)),
                _ => None,
            }
        };

        let sub_a = classify_sub(attach_points[0]);
        let sub_b = classify_sub(attach_points[1]);

        let (sub_a, sub_b) = match (sub_a, sub_b) {
            (Some(a), Some(b)) => (a, b),
            _ => return Err(IupacError::NotSupported),
        };

        // Determine locant prefix (1,2= ortho, 1,3= meta, 1,4= para for 6-ring).
        let pos2 = ring_dist + 1; // position of the second substituent from first

        // Build name: principal group determines root, non-principal is prefix.
        let (prefix_sub, root_name) = if sub_a.1 {
            // sub_a is principal (phenol/aniline): prefix comes from sub_b
            let root = match sub_a.0 {
                "hydroxy" => "phenol",
                "amino" => "aniline",
                _ => return Err(IupacError::NotSupported),
            };
            (sub_b.0, root)
        } else if sub_b.1 {
            let root = match sub_b.0 {
                "hydroxy" => "phenol",
                "amino" => "aniline",
                _ => return Err(IupacError::NotSupported),
            };
            (sub_a.0, root)
        } else {
            // Neither is principal — both are substituents on benzene.
            // Alphabetically first substituent gets locant 1.
            let (s1, s2) = if sub_a.0 <= sub_b.0 {
                (sub_a.0, sub_b.0)
            } else {
                (sub_b.0, sub_a.0)
            };
            return if s1 == s2 {
                Ok(format!("1,{}-di{}benzene", pos2, s1))
            } else {
                Ok(format!("1-{}-{}-{}benzene", s1, pos2, s2))
            };
        };

        Ok(format!("{}-{}{}", pos2, prefix_sub, root_name))
    }

    // -----------------------------------------------------------------------
    // Trisubstituted benzene naming
    // -----------------------------------------------------------------------

    fn name_trisubstituted_benzene(
        &self,
        ring_atoms: &HashSet<AtomIdx>,
    ) -> Result<String, IupacError> {
        let mol = self.mol;
        let attach_points: Vec<AtomIdx> = ring_atoms.iter()
            .filter(|&&r| mol.neighbors(r).any(|(nb, _)| !ring_atoms.contains(&nb)))
            .copied()
            .collect();
        if attach_points.len() != 3 {
            return Err(IupacError::NotSupported);
        }

        let locant_map = best_benzene_locants(mol, ring_atoms, &attach_points);

        // Classify each substituent.
        let mut sub_list: Vec<(usize, String)> = Vec::new();
        for &(locant, attach) in &locant_map {
            let sub = self.classify_benzene_sub_simple(attach, ring_atoms)
                .ok_or(IupacError::NotSupported)?;
            sub_list.push((locant, sub));
        }

        // Sort alphabetically by substituent name, then numerically by locant.
        sub_list.sort_by(|a, b| a.1.cmp(&b.1).then(a.0.cmp(&b.0)));

        // Group identical substituents for di/tri multiplier.
        let mut groups: Vec<(String, Vec<usize>)> = Vec::new();
        for (locant, name) in sub_list {
            if let Some(last) = groups.last_mut() {
                if last.0 == name {
                    last.1.push(locant);
                    continue;
                }
            }
            groups.push((name, vec![locant]));
        }

        let mut parts: Vec<String> = Vec::new();
        for (name, mut locs) in groups {
            locs.sort_unstable();
            let locant_str = locs.iter().map(|l| l.to_string()).collect::<Vec<_>>().join(",");
            let mult = match locs.len() {
                1 => String::new(),
                2 => "di".to_string(),
                3 => "tri".to_string(),
                _ => return Err(IupacError::NotSupported),
            };
            parts.push(format!("{}-{}{}", locant_str, mult, name));
        }

        Ok(format!("{}benzene", parts.join("-")))
    }

    /// Classify a single benzene substituent by element type (for trisubstituted naming).
    fn classify_benzene_sub_simple(
        &self,
        attach: AtomIdx,
        ring_atoms: &HashSet<AtomIdx>,
    ) -> Option<String> {
        let mol = self.mol;
        let direct: Vec<AtomIdx> = mol.neighbors(attach)
            .filter(|(nb, _)| !ring_atoms.contains(nb))
            .map(|(nb, _)| nb)
            .collect();
        if direct.is_empty() { return None; }
        let first = direct[0];
        match mol.atom(first).element.atomic_number() {
            6  => Some("methyl".to_string()),
            7  => Some("amino".to_string()),
            8  => Some("hydroxy".to_string()),
            9  => Some("fluoro".to_string()),
            17 => Some("chloro".to_string()),
            35 => Some("bromo".to_string()),
            53 => Some("iodo".to_string()),
            _  => None,
        }
    }

    // -----------------------------------------------------------------------
    // Nitrile naming (R-C≡N → "...nitrile")
    // -----------------------------------------------------------------------

    fn is_nitrile(&self, n_idx: AtomIdx) -> bool {
        self.mol.neighbors(n_idx)
            .any(|(_, bi)| self.mol.bond(bi).order == BondOrder::Triple)
    }

    fn name_nitrile(&self, carbons: &[AtomIdx], n_idx: AtomIdx) -> Result<String, IupacError> {
        let mol = self.mol;
        // Find the C≡N carbon.
        let nitrile_c = mol.neighbors(n_idx)
            .filter(|(_, bi)| mol.bond(*bi).order == BondOrder::Triple)
            .map(|(nb, _)| nb)
            .next()
            .ok_or(IupacError::NotSupported)?;
        // Count the total C chain (nitrile C + alkyl chain).
        // count_c_chain gives all C reachable from nitrile_c without crossing N.
        let n_carbons = count_c_chain(mol, nitrile_c, n_idx);
        // n_carbons already includes the nitrile carbon itself.
        if n_carbons == 0 { return Err(IupacError::NotSupported); }
        // Verify no branching on the C chain
        let c_set: std::collections::HashSet<AtomIdx> = carbons.iter().copied().collect();
        for &c in carbons {
            if mol.neighbors(c)
                .filter(|(nb, _)| c_set.contains(nb))
                .count() > 2
            {
                return Err(IupacError::NotSupported); // branched nitrile not supported
            }
        }
        Ok(format!("{}enitrile", alkane_base(n_carbons)))
    }

    // -----------------------------------------------------------------------
    // Branched alkane naming (e.g., "2-methylpropane", "2,2-dimethylpropane")
    // -----------------------------------------------------------------------

    fn name_branched_alkane(&self, carbons: &[AtomIdx]) -> Result<String, IupacError> {
        let mol = self.mol;

        // Find the principal chain (longest C–C path).
        let chain = find_longest_c_chain(mol, carbons);
        let n = chain.len();
        if n < 2 {
            return Err(IupacError::NotSupported);
        }

        let chain_set: std::collections::HashSet<AtomIdx> = chain.iter().copied().collect();
        let all_c_set: std::collections::HashSet<AtomIdx> = carbons.iter().copied().collect();

        // Collect substituents: (chain_position_1based, alkyl_length).
        let mut subs: Vec<(usize, usize)> = Vec::new();
        for (pos0, &chain_c) in chain.iter().enumerate() {
            let position = pos0 + 1;
            for (nb, _) in mol.neighbors(chain_c) {
                if all_c_set.contains(&nb) && !chain_set.contains(&nb) {
                    // Substituent rooted at `nb`, blocked by chain_c.
                    let sub_len = count_c_chain(mol, nb, chain_c);
                    if sub_len > 4 {
                        return Err(IupacError::NotSupported);
                    }
                    subs.push((position, sub_len));
                }
            }
        }

        if subs.is_empty() {
            return Err(IupacError::NotSupported);
        }

        // Apply IUPAC lowest-locant rule: try forward and reverse numbering.
        let subs_rev: Vec<(usize, usize)> = subs.iter()
            .map(|&(pos, len)| (n + 1 - pos, len))
            .collect();

        let first_fwd = subs.iter().map(|&(p, _)| p).min().unwrap_or(usize::MAX);
        let first_rev = subs_rev.iter().map(|&(p, _)| p).min().unwrap_or(usize::MAX);
        let best_subs = if first_fwd <= first_rev { subs } else { subs_rev };

        Ok(format!(
            "{}{}",
            format_substituents(&best_subs),
            alkane_suffix(n)
        ))
    }
}

// ---------------------------------------------------------------------------
// Graph helpers
// ---------------------------------------------------------------------------

fn atoms_of(mol: &Molecule, atomic_num: u8) -> Vec<AtomIdx> {
    mol.atoms()
        .filter(|(_, a)| a.element.atomic_number() == atomic_num)
        .map(|(i, _)| i)
        .collect()
}

/// BFS count of C atoms reachable from `start` without crossing `blocked`.
fn count_c_chain(mol: &Molecule, start: AtomIdx, blocked: AtomIdx) -> usize {
    let mut visited = HashSet::new();
    let mut queue = VecDeque::new();
    visited.insert(start);
    queue.push_back(start);
    while let Some(cur) = queue.pop_front() {
        for (nb, _) in mol.neighbors(cur) {
            if nb == blocked { continue; }
            if mol.atom(nb).element.atomic_number() == 6 && visited.insert(nb) {
                queue.push_back(nb);
            }
        }
    }
    visited.len()
}

/// Find the longest carbon chain in a C-subgraph using two-pass BFS.
///
/// Returns the sequence of AtomIdx forming the longest simple path.
/// For branched alkanes this gives the principal chain (IUPAC rule: longest chain).
fn find_longest_c_chain(mol: &Molecule, carbons: &[AtomIdx]) -> Vec<AtomIdx> {
    if carbons.is_empty() { return Vec::new(); }

    let c_set: std::collections::HashSet<AtomIdx> = carbons.iter().copied().collect();

    // BFS to find the farthest atom from a given start, returning (farthest, parents).
    let bfs_far = |start: AtomIdx| -> (AtomIdx, std::collections::HashMap<AtomIdx, AtomIdx>) {
        let mut parent: std::collections::HashMap<AtomIdx, AtomIdx> = std::collections::HashMap::new();
        let mut visited: std::collections::HashSet<AtomIdx> = std::collections::HashSet::new();
        let mut queue = VecDeque::new();
        let mut farthest = start;
        visited.insert(start);
        queue.push_back(start);
        while let Some(cur) = queue.pop_front() {
            farthest = cur;
            for (nb, _) in mol.neighbors(cur) {
                if c_set.contains(&nb) && visited.insert(nb) {
                    parent.insert(nb, cur);
                    queue.push_back(nb);
                }
            }
        }
        (farthest, parent)
    };

    let reconstruct = |end: AtomIdx, start: AtomIdx,
                        parents: &std::collections::HashMap<AtomIdx, AtomIdx>| -> Vec<AtomIdx> {
        let mut path = vec![end];
        let mut cur = end;
        while cur != start {
            cur = parents[&cur];
            path.push(cur);
        }
        path.reverse();
        path
    };

    // Pass 1: BFS from first carbon to find one endpoint of the longest chain.
    let (end1, _) = bfs_far(carbons[0]);
    // Pass 2: BFS from end1 to find the other endpoint.
    let (end2, parents) = bfs_far(end1);

    reconstruct(end2, end1, &parents)
}

/// Format substituents as an IUPAC prefix string ("2-methyl", "2,2-dimethyl", etc.).
fn format_substituents(subs: &[(usize, usize)]) -> String {
    // Group by alkyl name; sort alphabetically.
    let mut groups: std::collections::BTreeMap<&str, Vec<usize>> =
        std::collections::BTreeMap::new();
    for &(pos, len) in subs {
        let alkyl = match len {
            1 => "methyl",
            2 => "ethyl",
            3 => "propyl",
            4 => "butyl",
            _ => continue,
        };
        groups.entry(alkyl).or_default().push(pos);
    }

    let mut parts: Vec<String> = Vec::new();
    for (alkyl, mut positions) in groups {
        positions.sort_unstable();
        let locants = positions.iter().map(|p| p.to_string()).collect::<Vec<_>>().join(",");
        let mult = match positions.len() {
            1 => String::new(),
            2 => "di".to_string(),
            3 => "tri".to_string(),
            _ => "?".to_string(),
        };
        parts.push(format!("{}-{}{}", locants, mult, alkyl));
    }
    parts.join("-")
}

/// BFS chain anchored at `anchor` (always at index 0 = IUPAC position 1 in result).
fn chain_from_anchor(
    mol: &Molecule,
    c_set: &HashSet<AtomIdx>,
    anchor: AtomIdx,
) -> Vec<AtomIdx> {
    let mut parent: std::collections::HashMap<AtomIdx, AtomIdx> =
        std::collections::HashMap::new();
    let mut visited: HashSet<AtomIdx> = HashSet::new();
    let mut queue = VecDeque::new();
    let mut farthest = anchor;
    visited.insert(anchor);
    queue.push_back(anchor);
    while let Some(cur) = queue.pop_front() {
        farthest = cur;
        for (nb, _) in mol.neighbors(cur) {
            if c_set.contains(&nb) && visited.insert(nb) {
                parent.insert(nb, cur);
                queue.push_back(nb);
            }
        }
    }
    let mut path = vec![farthest];
    let mut cur = farthest;
    while cur != anchor { cur = parent[&cur]; path.push(cur); }
    path.reverse();
    path
}

/// Return the IUPAC locant (1-based, lowest) of a double or triple bond on the chain.
fn unsaturation_locant(mol: &Molecule, carbons: &[AtomIdx], order: BondOrder) -> usize {
    let chain = find_longest_c_chain(mol, carbons);
    let n = chain.len();
    for (_, b) in mol.bonds() {
        if b.order == order {
            if let (Some(p1), Some(p2)) = (
                chain.iter().position(|&c| c == b.atom1),
                chain.iter().position(|&c| c == b.atom2),
            ) {
                let fwd = p1.min(p2) + 1;  // 1-based lower position in forward direction
                let rev = n - p1.max(p2);  // 1-based lower position in reversed direction
                return fwd.min(rev);
            }
        }
    }
    1
}

/// Return ring atoms in cyclic traversal order.
fn ring_order_traversal(mol: &Molecule, ring_atoms: &HashSet<AtomIdx>) -> Vec<AtomIdx> {
    if ring_atoms.is_empty() { return Vec::new(); }
    let start = *ring_atoms.iter().next().unwrap();
    let mut order = vec![start];
    let first_nb = mol.neighbors(start).find(|(nb, _)| ring_atoms.contains(nb)).map(|(nb, _)| nb);
    let mut cur = match first_nb { Some(nb) => nb, None => return order };
    let mut prev = start;
    while cur != start {
        order.push(cur);
        let next = mol.neighbors(cur)
            .find(|(nb, _)| ring_atoms.contains(nb) && *nb != prev)
            .map(|(nb, _)| nb);
        prev = cur;
        match next { Some(nb) => cur = nb, None => break }
    }
    order
}

/// Find the minimum IUPAC locant assignment for `attach_points` on a ring.
/// Returns sorted `(locant, attach_atom)` pairs.
fn best_benzene_locants(
    mol: &Molecule,
    ring_atoms: &HashSet<AtomIdx>,
    attach_points: &[AtomIdx],
) -> Vec<(usize, AtomIdx)> {
    let ring_order = ring_order_traversal(mol, ring_atoms);
    let ring_n = ring_order.len();
    if ring_n == 0 { return Vec::new(); }
    let n = attach_points.len();
    let pos_of: Vec<usize> = attach_points.iter()
        .map(|a| ring_order.iter().position(|r| r == a).unwrap_or(0))
        .collect();
    let mut best_locs: Option<Vec<usize>> = None;
    let mut best_assignment: Vec<(usize, AtomIdx)> = Vec::new();
    for start in 0..n {
        for &reverse in &[false, true] {
            let mut assignment: Vec<(usize, AtomIdx)> = Vec::new();
            for k in 0..n {
                let idx = (start + k) % n;
                let pos = if !reverse {
                    (pos_of[idx] + ring_n - pos_of[start]) % ring_n
                } else {
                    (pos_of[start] + ring_n - pos_of[idx]) % ring_n
                };
                assignment.push((pos + 1, attach_points[idx]));
            }
            assignment.sort_by_key(|&(l, _)| l);
            let locs: Vec<usize> = assignment.iter().map(|&(l, _)| l).collect();
            let is_better = best_locs.as_ref().map_or(true, |b| locs < *b);
            if is_better {
                best_locs = Some(locs);
                best_assignment = assignment;
            }
        }
    }
    best_assignment
}

fn count_components(mol: &Molecule) -> usize {
    let n = mol.atom_count();
    if n == 0 { return 0; }
    let mut visited = vec![false; n];
    let mut count = 0;
    for start in 0..n {
        if visited[start] { continue; }
        count += 1;
        let mut queue = VecDeque::new();
        queue.push_back(AtomIdx(start as u32));
        visited[start] = true;
        while let Some(cur) = queue.pop_front() {
            for (nb, _) in mol.neighbors(cur) {
                if !visited[nb.0 as usize] {
                    visited[nb.0 as usize] = true;
                    queue.push_back(nb);
                }
            }
        }
    }
    count
}

// ---------------------------------------------------------------------------
// Naming helpers
// ---------------------------------------------------------------------------

fn alkane_stem(n: usize) -> &'static str {
    match n {
        1 => "meth", 2 => "eth",  3 => "prop", 4 => "but",
        5 => "pent", 6 => "hex",  7 => "hept", 8 => "oct",
        9 => "non",  10 => "dec", _ => "long",
    }
}

/// Stem with "an" appended — base for most suffix compounds.
fn alkane_base(n: usize) -> String {
    format!("{}an", alkane_stem(n))
}

fn alkane_suffix(n: usize) -> String {
    match n {
        1  => "methane".into(),   2  => "ethane".into(),
        3  => "propane".into(),   4  => "butane".into(),
        5  => "pentane".into(),   6  => "hexane".into(),
        7  => "heptane".into(),   8  => "octane".into(),
        9  => "nonane".into(),    10 => "decane".into(),
        11 => "undecane".into(),  12 => "dodecane".into(),
        13 => "tridecane".into(), 14 => "tetradecane".into(),
        15 => "pentadecane".into(), 16 => "hexadecane".into(),
        17 => "heptadecane".into(), 18 => "octadecane".into(),
        19 => "nonadecane".into(), 20 => "icosane".into(),
        _  => format!("{n}alkane"),
    }
}

fn alkene_suffix(n: usize) -> String { alkane_suffix(n).replace("ane", "ene") }
fn alkyne_suffix(n: usize) -> String { alkane_suffix(n).replace("ane", "yne") }

// ---------------------------------------------------------------------------
// Tests
// ---------------------------------------------------------------------------

#[cfg(test)]
mod tests {
    use super::*;
    use chematic_smiles::parse;

    fn mol(s: &str) -> Molecule { parse(s).unwrap() }

    // --- Existing tests (must remain green) ---------------------------------

    #[test]
    fn test_alkanes() {
        assert_eq!(name(&mol("C")).unwrap(),      "methane");
        assert_eq!(name(&mol("CC")).unwrap(),     "ethane");
        assert_eq!(name(&mol("CCC")).unwrap(),    "propane");
        assert_eq!(name(&mol("CCCC")).unwrap(),   "butane");
        assert_eq!(name(&mol("CCCCC")).unwrap(),  "pentane");
        assert_eq!(name(&mol("CCCCCC")).unwrap(), "hexane");
    }

    #[test]
    fn test_alkenes_alkynes() {
        assert_eq!(name(&mol("C=C")).unwrap(),   "ethene");
        assert_eq!(name(&mol("CC=C")).unwrap(),  "propene");
        assert_eq!(name(&mol("C#C")).unwrap(),   "ethyne");
        assert_eq!(name(&mol("CC#C")).unwrap(),  "propyne");
    }

    #[test]
    fn test_cycloalkanes() {
        assert_eq!(name(&mol("C1CC1")).unwrap(),   "cyclopropane");
        assert_eq!(name(&mol("C1CCC1")).unwrap(),  "cyclobutane");
        assert_eq!(name(&mol("C1CCCC1")).unwrap(), "cyclopentane");
        assert_eq!(name(&mol("C1CCCCC1")).unwrap(),"cyclohexane");
    }

    #[test]
    fn test_alcohol() {
        assert_eq!(name(&mol("CO")).unwrap(),   "methanol");
        assert_eq!(name(&mol("CCO")).unwrap(),  "ethanol");
        assert_eq!(name(&mol("CCCO")).unwrap(), "propan-1-ol");
    }

    #[test]
    fn test_amine() {
        assert_eq!(name(&mol("CN")).unwrap(),  "methan-1-amine");
        assert_eq!(name(&mol("CCN")).unwrap(), "ethan-1-amine");
    }

    #[test]
    fn test_haloalkane() {
        assert_eq!(name(&mol("CCCl")).unwrap(), "chloroethane");
        assert_eq!(name(&mol("CCBr")).unwrap(), "bromoethane");
        assert_eq!(name(&mol("CF")).unwrap(),   "fluoromethane");
        assert_eq!(name(&mol("CI")).unwrap(),   "iodomethane");
    }

    #[test]
    fn test_not_supported() {
        assert!(name(&mol("CC.CC")).is_err());  // disconnected
    }

    #[test]
    fn test_empty() {
        use chematic_core::MoleculeBuilder;
        let mol = MoleculeBuilder::new().build();
        assert_eq!(name(&mol), Err(IupacError::Empty));
    }

    // --- New: benzene & aromatic heterocycles --------------------------------

    #[test]
    fn test_benzene() {
        assert_eq!(name(&mol("c1ccccc1")).unwrap(), "benzene");
    }

    #[test]
    fn test_aromatic_heterocycles() {
        assert_eq!(name(&mol("c1ccncc1")).unwrap(),   "pyridine");
        assert_eq!(name(&mol("c1ccoc1")).unwrap(),    "furan");
        assert_eq!(name(&mol("c1ccsc1")).unwrap(),    "thiophene");
        assert_eq!(name(&mol("c1cc[nH]c1")).unwrap(), "pyrrole");
        assert_eq!(name(&mol("c1cnc[nH]1")).unwrap(), "imidazole");
    }

    // --- New: ketones with position locant -----------------------------------

    #[test]
    fn test_ketones() {
        assert_eq!(name(&mol("CC(=O)C")).unwrap(),    "propan-2-one");
        assert_eq!(name(&mol("CC(=O)CC")).unwrap(),   "butan-2-one");
        assert_eq!(name(&mol("CCC(=O)CC")).unwrap(),  "pentan-3-one");
        assert_eq!(name(&mol("CCCC(=O)C")).unwrap(),  "pentan-2-one");
    }

    // --- New: carboxylic acids -----------------------------------------------

    #[test]
    fn test_carboxylic_acids() {
        assert_eq!(name(&mol("CC(=O)O")).unwrap(),  "ethanoic acid");
        assert_eq!(name(&mol("CCC(=O)O")).unwrap(), "propanoic acid");
        assert_eq!(name(&mol("C(=O)O")).unwrap(),   "methanoic acid");
    }

    // --- New: esters ---------------------------------------------------------

    #[test]
    fn test_esters() {
        assert_eq!(name(&mol("CC(=O)OC")).unwrap(),  "methyl ethanoate");
        assert_eq!(name(&mol("C(=O)OC")).unwrap(),   "methyl methanoate");
        assert_eq!(name(&mol("CC(=O)OCC")).unwrap(), "ethyl ethanoate");
    }

    // --- New: amides ---------------------------------------------------------

    #[test]
    fn test_amides() {
        assert_eq!(name(&mol("CC(=O)N")).unwrap(),   "ethanamide");
        assert_eq!(name(&mol("C(=O)N")).unwrap(),    "methanamide");
        assert_eq!(name(&mol("CCC(=O)N")).unwrap(),  "propanamide");
    }

    // ---- New: branched alkanes (v0.1.101) ------------------------------------

    #[test]
    fn test_branched_alkanes() {
        assert_eq!(name(&mol("CC(C)C")).unwrap(),    "2-methylpropane");
        assert_eq!(name(&mol("CC(C)CC")).unwrap(),   "2-methylbutane");
        assert_eq!(name(&mol("CC(C)(C)C")).unwrap(), "2,2-dimethylpropane");
        assert_eq!(name(&mol("CCCC(C)CC")).unwrap(), "3-methylhexane");
    }

    #[test]
    fn test_branched_alkane_lowest_locant() {
        // CCC(C)C = 2-methylbutane (not 3-methylbutane — lower locant wins).
        assert_eq!(name(&mol("CCC(C)C")).unwrap(), "2-methylbutane");
    }

    // ---- New: substituted benzenes (v0.1.101) --------------------------------

    #[test]
    fn test_substituted_benzenes() {
        assert_eq!(name(&mol("c1ccccc1O")).unwrap(),     "phenol");
        assert_eq!(name(&mol("c1ccccc1N")).unwrap(),     "aniline");
        assert_eq!(name(&mol("c1ccccc1Cl")).unwrap(),    "chlorobenzene");
        assert_eq!(name(&mol("c1ccccc1Br")).unwrap(),    "bromobenzene");
    }

    #[test]
    fn test_substituted_benzene_carbonyl() {
        assert_eq!(name(&mol("c1ccccc1C=O")).unwrap(),        "benzaldehyde");
        assert_eq!(name(&mol("c1ccccc1C(=O)O")).unwrap(),     "benzoic acid");
    }

    // ---- New: nitriles (v0.1.101) -------------------------------------------

    #[test]
    fn test_nitriles() {
        assert_eq!(name(&mol("CC#N")).unwrap(),  "ethanenitrile");
        assert_eq!(name(&mol("CCC#N")).unwrap(), "propanenitrile");
    }

    // ---- New Round 2 tests (v0.1.102) ---------------------------------------

    #[test]
    fn test_thiols() {
        assert_eq!(name(&mol("CS")).unwrap(),   "methanethiol");
        assert_eq!(name(&mol("CCS")).unwrap(),  "ethanethiol");
        assert_eq!(name(&mol("CCCS")).unwrap(), "propanethiol");
    }

    #[test]
    fn test_alcohol_locants() {
        assert_eq!(name(&mol("CCCCO")).unwrap(),  "butan-1-ol");
        assert_eq!(name(&mol("CC(O)C")).unwrap(), "propan-2-ol");
        assert_eq!(name(&mol("CCC(O)C")).unwrap(), "butan-2-ol");
    }

    #[test]
    fn test_disubstituted_benzene() {
        // Para-chlorophenol: OH and Cl are 3 bonds apart in the ring (positions 1 and 4).
        assert_eq!(name(&mol("Oc1ccc(Cl)cc1")).unwrap(), "4-chlorophenol");
        // Meta-chlorophenol: OH and Cl are 2 bonds apart (positions 1 and 3).
        assert_eq!(name(&mol("c1ccc(O)cc1Cl")).unwrap(), "3-chlorophenol");
    }

    #[test]
    fn test_methylcycloalkane() {
        assert_eq!(name(&mol("CC1CCCCC1")).unwrap(), "methylcyclohexane");
        assert_eq!(name(&mol("CC1CCCC1")).unwrap(),  "methylcyclopentane");
        assert_eq!(name(&mol("CC1CCC1")).unwrap(),   "methylcyclobutane");
    }

    // ---- New Round 3 tests (v0.1.103) ----------------------------------------

    #[test]
    fn test_ethers() {
        assert_eq!(name(&mol("COC")).unwrap(),    "methoxymethane");
        assert_eq!(name(&mol("COCC")).unwrap(),   "methoxyethane");
        assert_eq!(name(&mol("CCOCC")).unwrap(),  "ethoxyethane");
        assert_eq!(name(&mol("COCCC")).unwrap(),  "1-methoxypropane");
    }

    #[test]
    fn test_trimethylbenzene() {
        assert_eq!(name(&mol("Cc1cccc(C)c1C")).unwrap(),   "1,2,3-trimethylbenzene");
        assert_eq!(name(&mol("Cc1ccc(C)cc1C")).unwrap(),   "1,2,4-trimethylbenzene");
        assert_eq!(name(&mol("Cc1cc(C)cc(C)c1")).unwrap(), "1,3,5-trimethylbenzene");
    }

    #[test]
    fn test_secondary_amine() {
        assert_eq!(name(&mol("CCNCC")).unwrap(),  "N-ethylethanamine");
        assert_eq!(name(&mol("CNCC")).unwrap(),   "N-methylethanamine");
        assert_eq!(name(&mol("CN(C)C")).unwrap(), "N,N-dimethylmethanamine");
    }

    // ---- New Round 9 tests (v0.1.109) ----------------------------------------

    #[test]
    fn test_branched_aldehyde() {
        assert_eq!(name(&mol("CC(C)C=O")).unwrap(),  "2-methylpropanal");
        assert_eq!(name(&mol("CCC(C)C=O")).unwrap(), "2-methylbutanal");
    }

    #[test]
    fn test_branched_amide() {
        assert_eq!(name(&mol("CC(C)C(=O)N")).unwrap(),  "2-methylpropanamide");
        assert_eq!(name(&mol("CCC(C)C(=O)N")).unwrap(), "2-methylbutanamide");
    }

    // ---- New Round 8 tests (v0.1.108) ----------------------------------------

    #[test]
    fn test_branched_ester() {
        assert_eq!(name(&mol("CC(C)C(=O)OC")).unwrap(),  "methyl 2-methylpropanoate");
        assert_eq!(name(&mol("CC(C)C(=O)OCC")).unwrap(), "ethyl 2-methylpropanoate");
    }

    #[test]
    fn test_branched_ketone() {
        assert_eq!(name(&mol("CC(=O)C(C)C")).unwrap(),    "3-methylbutan-2-one");
        assert_eq!(name(&mol("CC(=O)C(C)(C)C")).unwrap(), "3,3-dimethylbutan-2-one");
    }

    // ---- New Round 7 tests (v0.1.107) ----------------------------------------

    #[test]
    fn test_secondary_thiol() {
        assert_eq!(name(&mol("CCC(S)C")).unwrap(),  "butane-2-thiol");
        assert_eq!(name(&mol("CCCC(S)C")).unwrap(), "pentane-2-thiol");
    }

    #[test]
    fn test_branched_carboxylic_acid() {
        assert_eq!(name(&mol("CC(C)C(=O)O")).unwrap(),    "2-methylpropanoic acid");
        assert_eq!(name(&mol("CCC(C)C(=O)O")).unwrap(),   "2-methylbutanoic acid");
        assert_eq!(name(&mol("CC(C)(C)C(=O)O")).unwrap(), "2,2-dimethylpropanoic acid");
    }

    // ---- New Round 6 tests (v0.1.106) ----------------------------------------

    #[test]
    fn test_alkene_locants() {
        assert_eq!(name(&mol("CC=CC")).unwrap(),   "but-2-ene");
        assert_eq!(name(&mol("C=CCC")).unwrap(),   "but-1-ene");
        assert_eq!(name(&mol("CC=CCC")).unwrap(),  "pent-2-ene");
        assert_eq!(name(&mol("C=CCCC")).unwrap(),  "pent-1-ene");
    }

    #[test]
    fn test_alkyne_locants() {
        assert_eq!(name(&mol("CC#CC")).unwrap(),  "but-2-yne");
        assert_eq!(name(&mol("C#CCC")).unwrap(),  "but-1-yne");
    }

    #[test]
    fn test_amine_locants() {
        assert_eq!(name(&mol("CCCN")).unwrap(),     "propan-1-amine");
        assert_eq!(name(&mol("CCC(N)C")).unwrap(),  "butan-2-amine");
        assert_eq!(name(&mol("CC(N)CCC")).unwrap(), "pentan-2-amine");
    }

    // ---- New Round 5 tests (v0.1.105) ----------------------------------------

    #[test]
    fn test_haloalkane_locants() {
        // n=3: terminal → "1-chloropropane"
        assert_eq!(name(&mol("CCCCl")).unwrap(),   "1-chloropropane");
        // n=4: terminal → "1-chlorobutane"
        assert_eq!(name(&mol("CCCCCl")).unwrap(),  "1-chlorobutane");
        // n=4: internal → "2-chlorobutane"
        assert_eq!(name(&mol("CCC(Cl)C")).unwrap(), "2-chlorobutane");
        // n=5: internal → "2-chloropentane"
        assert_eq!(name(&mol("CCCC(Cl)C")).unwrap(), "2-chloropentane");
        // di-halo: ClCCCl = 2C → "1,2-dichloroethane"; ClCCCCl = 3C → "1,3-dichloropropane"
        assert_eq!(name(&mol("ClCCCl")).unwrap(),   "1,2-dichloroethane");
        assert_eq!(name(&mol("ClCCCCl")).unwrap(),  "1,3-dichloropropane");
    }

    #[test]
    fn test_cycloalkanol() {
        assert_eq!(name(&mol("OC1CCC1")).unwrap(),   "cyclobutanol");
        assert_eq!(name(&mol("OC1CCCC1")).unwrap(),  "cyclopentanol");
        assert_eq!(name(&mol("OC1CCCCC1")).unwrap(), "cyclohexanol");
    }

    // ---- New Round 4 tests (v0.1.104) ----------------------------------------

    #[test]
    fn test_disubstituted_benzene_non_principal() {
        // Two halogens (para)
        assert_eq!(name(&mol("Clc1ccc(Br)cc1")).unwrap(), "1-bromo-4-chlorobenzene");
        assert_eq!(name(&mol("Clc1ccc(F)cc1")).unwrap(),  "1-chloro-4-fluorobenzene");
        // Two methyls: ortho (Cc1ccccc1C) and para (Cc1ccc(C)cc1)
        assert_eq!(name(&mol("Cc1ccccc1C")).unwrap(),     "1,2-dimethylbenzene");
        assert_eq!(name(&mol("Cc1ccc(C)cc1")).unwrap(),   "1,4-dimethylbenzene");
        // Methyl + halogen (para)
        assert_eq!(name(&mol("Cc1ccc(Cl)cc1")).unwrap(),  "1-chloro-4-methylbenzene");
    }

    #[test]
    fn test_propyl_substituent() {
        // 11C: longest chain = octane (8C), propyl substituent at C4
        assert_eq!(name(&mol("CCCC(CCC)CCCC")).unwrap(), "4-propyloctane");
    }

    #[test]
    fn test_dimethylcycloalkane() {
        assert_eq!(name(&mol("CC1CCC(C)CC1")).unwrap(), "1,4-dimethylcyclohexane");
        assert_eq!(name(&mol("CC1CCCC1C")).unwrap(),    "1,2-dimethylcyclopentane");
        assert_eq!(name(&mol("CC1CCC(C)C1")).unwrap(),  "1,3-dimethylcyclopentane");
    }
}