SmilesParser Class Reference

Parse a SMILES string into a molecule. More...

#include <avogadro/io/smilesparser.h>

Public Member Functions
bool	parse (const std::string &smiles, Core::Molecule &molecule)
const std::string &	error () const
size_t	errorPosition () const
const std::vector< std::string > &	warnings () const
const std::vector< size_t > &	atomMaps () const

Detailed Description

This is the inverse of SmilesWriter, and follows the OpenSMILES grammar (http://opensmiles.org/opensmiles.html): the organic subset, bracket atoms, branches, ring closures and dot-disconnected components.

Aromatic input – a lowercase atom, or an aromatic bond, anywhere in the string – is accepted: Core::kekulize() turns the aromatic ring or ring system back into alternating single and double bond orders, which is how Avogadro stores structures internally. If no assignment satisfies every aromatic atom's valence – an unkekulizable input – parsing fails with an error positioned at the first aromatic token in the string, and the molecule is left empty, the same as any other parse failure.

Chirality markers and the directional bond markers '/' and '\' are parsed – so the rest of the string still makes sense – but the stereochemistry they describe is discarded: a freshly parsed molecule has no coordinates, and Avogadro represents stereochemistry through coordinates rather than through a stored descriptor. warnings() reports this once per parse, not once per occurrence.

Avogadro's convention is that every hydrogen is a real atom in the graph, so every hydrogen this class infers – from the organic subset valence model on a bare atom, or from a bracket's H count – is materialized as one. A bracket atom's count is stated in the string, but a bare aromatic atom's implied count depends on the bond orders kekulization produces, so hydrogens for those atoms are added only after kekulization succeeds. As a result, the molecule this class produces orders its atoms as: first, every atom written explicitly in the SMILES, in the order it appears in the string (branches included, depth first); then every hydrogen this class added. The order of the added hydrogens themselves is unspecified – they are interchangeable – and does not simply follow their parent atoms' order, since it is now split across two passes.

Member Function Documentation

parse()

bool parse	(	const std::string &	smiles,
		Core::Molecule &	molecule )

Parse smiles into molecule, replacing any atoms and bonds it already had. The molecule is left with none of either on failure.

Returns

True on success.

error()

const std::string & error

(

)

const

Returns

A description of the last failure, or an empty string.

errorPosition()

size_t errorPosition

(

)

const

Returns

Character offset into the input at which the last failure was detected.

warnings()

const std::vector< std::string > & warnings

(

)

const

Returns

Non-fatal notes about information the parse discarded.

atomMaps()

const std::vector< size_t > & atomMaps

(

)

const

Daylight atom map classes, one entry per atom of the molecule produced, zero where the atom carried none.

These are exposed but never applied: a map class is data from the input, and using it to place an atom would let a malformed or hostile string reorder or alias atoms. Callers wanting to correlate with a mapped write must subtract one, since SmilesWriter emits index + 1.

---

The documentation for this class was generated from the following file:

• smilesparser.h