SMARTS (SMiles ARbitrary Target Specification) substructure searching. More...

#include <openbabel/parsmart.h>

Public Types
enum	MatchType { All, Single, AllUnique }
Public Member Functions
	OBSmartsPattern ()
virtual	~OBSmartsPattern ()
	OBSmartsPattern (const OBSmartsPattern &cp)
OBSmartsPattern &	operator= (const OBSmartsPattern &cp)
void	WriteMapList (std::ostream &)
Initialization Methods

bool	Init (const char *pattern)
bool	Init (const std::string &pattern)
Pattern Properties

const std::string &	GetSMARTS () const
std::string &	GetSMARTS ()
bool	Empty () const
bool	IsValid () const
unsigned int	NumAtoms () const
unsigned int	NumBonds () const
void	GetBond (int &src, int &dst, int &ord, int idx)
int	GetAtomicNum (int idx)
int	GetCharge (int idx)
int	GetVectorBinding (int idx) const
Matching methods (SMARTS on a specific OBMol)

bool	Match (OBMol &mol, bool single=false)
bool	Match (OBMol &mol, std::vector< std::vector< int > > &mlist, MatchType mtype=All) const
bool	HasMatch (OBMol &mol) const
bool	RestrictedMatch (OBMol &mol, std::vector< std::pair< int, int > > &pairs, bool single=false)
bool	RestrictedMatch (OBMol &mol, OBBitVec &bv, bool single=false)
unsigned int	NumMatches () const
std::vector< std::vector< int > > &	GetMapList ()
std::vector< std::vector< int > >::iterator	BeginMList ()
std::vector< std::vector< int > >::iterator	EndMList ()
std::vector< std::vector< int > > &	GetUMapList ()
Protected Member Functions
Pattern *	ParseSMARTSPattern (void)
Pattern *	ParseSMARTSPart (Pattern *, int)
Pattern *	SMARTSError (Pattern *pat)
Pattern *	ParseSMARTSError (Pattern pat, BondExpr expr)
AtomExpr *	ParseSimpleAtomPrimitive (void)
AtomExpr *	ParseComplexAtomPrimitive (void)
AtomExpr *	ParseAtomExpr (int level)
BondExpr *	ParseBondPrimitive (void)
BondExpr *	ParseBondExpr (int level)
Pattern *	ParseSMARTSString (char *ptr)
Pattern *	ParseSMARTSRecord (char *ptr)
int	GetVectorBinding ()
Pattern *	SMARTSParser (Pattern pat, ParseState stat, int prev, int part)
Protected Attributes
OBSmartsPrivate *	_d
std::vector< bool >	_growbond
std::vector< std::vector< int > >	_mlist
Pattern *	_pat
std::string	_str
char *	_buffer
char *	LexPtr
char *	MainPtr

Detailed Description

SMARTS (SMiles ARbitrary Target Specification) substructure searching.

Substructure search is an incredibly useful tool in the context of a small molecule programming library. Having an efficient substructure search engine reduces the amount of hard code needed for molecule perception, as well as increases the flexibility of certain operations. For instance, atom typing can be easily performed based on hard coded rules of element type and bond orders (or hybridization). Alternatively, atom typing can also be done by matching a set of substructure rules read at run time. In the latter case customization based on application (such as changing the pH) becomes a facile operation. Fortunately for Open Babel and its users, Roger Sayle donated a SMARTS parser which became the basis for SMARTS matching in Open Babel.

For more information on the SMARTS support in Open Babel, see the wiki page: http://openbabel.org/wiki/SMARTS

The SMARTS matcher, or OBSmartsPattern, is a separate object which can match patterns in the OBMol class. The following code demonstrates how to use the OBSmartsPattern class:

    OBMol mol;
    ...
    OBSmartsPattern sp;
    sp.Init("CC");
    sp.Match(mol);
    vector<vector<int> > maplist;
    maplist = sp.GetMapList();
    //or maplist = sp.GetUMapList();
    //print out the results
    vector<vector<int> >::iterator i;
    vector<int>::iterator j;
    for (i = maplist.begin();i != maplist.end();++i)
    {
    for (j = i->begin();j != i->end();++j)
    cout << j << ' `;
    cout << endl;
    }

The preceding code reads in a molecule, initializes a SMARTS pattern of two single-bonded carbons, and locates all instances of the pattern in the molecule. Note that calling the Match() function does not return the results of the substructure match. The results from a match are stored in the OBSmartsPattern, and a call to GetMapList() or GetUMapList() must be made to extract the results. The function GetMapList() returns all matches of a particular pattern while GetUMapList() returns only the unique matches. For instance, the pattern [OD1]~C~[OD1] describes a carboxylate group. This pattern will match both atom number permutations of the carboxylate, and if GetMapList() is called, both matches will be returned. If GetUMapList() is called only unique matches of the pattern will be returned. A unique match is defined as one which does not cover the identical atoms that a previous match has covered.

Member Enumeration Documentation

enum MatchType

Enumerator:

All
Single
AllUnique

Constructor & Destructor Documentation

OBSmartsPattern ( ) [inline]

~OBSmartsPattern ( ) [virtual]

OBSmartsPattern ( const OBSmartsPattern & cp ) [inline]

Member Function Documentation

Pattern * ParseSMARTSPattern ( void ) [protected]

Referenced by OBSmartsPattern::ParseComplexAtomPrimitive(), and OBSmartsPattern::ParseSMARTSString().

Pattern * ParseSMARTSPart	(	Pattern *	result,
		int	part
	)			`[protected]`

Referenced by OBSmartsPattern::ParseSMARTSPattern().

Pattern * SMARTSError ( Pattern * pat ) [protected]

Referenced by OBSmartsPattern::ParseSMARTSError(), OBSmartsPattern::ParseSMARTSPart(), OBSmartsPattern::ParseSMARTSPattern(), and OBSmartsPattern::ParseSMARTSString().

Pattern * ParseSMARTSError	(	Pattern *	pat,
		BondExpr *	expr
	)			`[protected]`

Referenced by OBSmartsPattern::SMARTSParser().

AtomExpr * ParseSimpleAtomPrimitive ( void ) [protected]

Referenced by OBSmartsPattern::SMARTSParser().

AtomExpr * ParseComplexAtomPrimitive ( void ) [protected]

Referenced by OBSmartsPattern::ParseAtomExpr().

AtomExpr * ParseAtomExpr ( int level ) [protected]

Referenced by OBSmartsPattern::SMARTSParser().

BondExpr * ParseBondPrimitive ( void ) [protected]

Referenced by OBSmartsPattern::ParseBondExpr().

BondExpr * ParseBondExpr ( int level ) [protected]

Referenced by OBSmartsPattern::SMARTSParser().

Pattern * ParseSMARTSString ( char * ptr ) [protected]

Referenced by OBSmartsPattern::ParseSMARTSRecord().

Pattern * ParseSMARTSRecord ( char * ptr ) [protected]

Referenced by OBSmartsPattern::Init().

int GetVectorBinding ( ) [protected]

Referenced by OBChemTsfm::Init(), and OBSmartsPattern::SMARTSParser().

Pattern * SMARTSParser	(	Pattern *	pat,
		ParseState *	stat,
		int	prev,
		int	part
	)			`[protected]`

Referenced by OBSmartsPattern::ParseSMARTSPart().

OBSmartsPattern& operator= ( const OBSmartsPattern & cp ) [inline]

bool Init ( const char * pattern )

Parse the pattern SMARTS string.

Returns:: Whether the pattern is a valid SMARTS expression

Referenced by patty::assign_rules(), OBBondTyper::AssignFunctionalGroupBonds(), OpenBabel::CorrectBadResonanceForm(), OBMol::DoTransformations(), OBChemTsfm::Init(), OBBuilder::LoadFragments(), OBAtom::MatchesSMARTS(), OBMol::NewPerceiveKekuleBonds(), OBRotorList::OBRotorList(), OBAromaticTyper::ParseLine(), OBRingTyper::ParseLine(), OBAtomTyper::ParseLine(), OBPhModel::ParseLine(), OBBondTyper::ParseLine(), and patty::read_rules().

bool Init ( const std::string & pattern )

Parse the pattern SMARTS string.

Returns:: Whether the pattern is a valid SMARTS expression

const std::string& GetSMARTS ( ) const [inline]

Returns:: the SMARTS string which is currently used

std::string& GetSMARTS ( ) [inline]

Returns:: the SMARTS string which is currently used

bool Empty ( ) const [inline]

Returns:: If the SMARTS pattern is an empty expression (e.g., invalid)

bool IsValid ( ) const [inline]

Returns:: If the SMARTS pattern is a valid expression

unsigned int NumAtoms ( ) const [inline]

Returns:: the number of atoms in the SMARTS pattern

Referenced by OBChemTsfm::Init(), OBChemTsfm::IsAcid(), OBChemTsfm::IsBase(), and OBPhModel::ParseLine().

unsigned int NumBonds ( ) const [inline]

Returns:: the number of bonds in the SMARTS pattern

Referenced by OBChemTsfm::Init().

void GetBond	(	int &	src,
		int &	dst,
		int &	ord,
		int	idx
	)

Access the bond idx in the internal pattern

Parameters:

	src	The index of the beginning atom
	dst	The index of the end atom
	ord	The bond order of this bond
	idx	The index of the bond in the SMARTS pattern

Referenced by OBChemTsfm::Init().

int GetAtomicNum ( int idx )

Returns:: the atomic number of the atom idx in the internal pattern

Referenced by OBChemTsfm::Init().

int GetCharge ( int idx )

Returns:: the formal charge of the atom idx in the internal pattern

Referenced by OBChemTsfm::Init(), OBChemTsfm::IsAcid(), and OBChemTsfm::IsBase().

int GetVectorBinding ( int idx ) const [inline]

Returns:: the vector binding of the atom idx in the internal pattern

bool Match	(	OBMol &	mol,
		bool	single = `false`
	)

Perform SMARTS matching for the pattern specified using Init().

Parameters:

	mol	The molecule to use for matching
	single	Whether only a single match is required (faster). Default is false.

Returns:: Whether matches occurred

Referenced by OBChemTsfm::Apply(), OBBondTyper::AssignFunctionalGroupBonds(), OpenBabel::CorrectBadResonanceForm(), OBMol::DoTransformations(), OBSmartsPattern::HasMatch(), OBAtom::MatchesSMARTS(), and OBMol::NewPerceiveKekuleBonds().

bool Match	(	OBMol &	mol,
		std::vector< std::vector< int > > &	mlist,
		MatchType	mtype = `All`
	)			const

Perform SMARTS matching for the pattern specified using Init(). This version is (more) thread safe.

Parameters:

	mol	The molecule to use for matching
	mlist	The resulting match list
	mtype	The match type to use. Default is All.

Returns:: Whether matches occurred

bool HasMatch ( OBMol & mol ) const

Thread safe check for any SMARTS match

Parameters:

mol

The molecule to use for matching

Returns:: Whether there exists any match

bool RestrictedMatch	(	OBMol &	mol,
		std::vector< std::pair< int, int > > &	pairs,
		bool	single = `false`
	)

Referenced by OBRotorRules::GetRotorIncrements().

bool RestrictedMatch	(	OBMol &	mol,
		OBBitVec &	bv,
		bool	single = `false`
	)

unsigned int NumMatches ( ) const [inline]

Returns:: the number of non-unique SMARTS matches To get the number of unique SMARTS matches, query GetUMapList()->size()

std::vector<std::vector<int> >& GetMapList ( ) [inline]

Returns:: the entire list of non-unique matches for this pattern

See also:: GetUMapList()

Referenced by OBRotorRules::GetRotorIncrements().

std::vector<std::vector<int> >::iterator BeginMList ( ) [inline]

Returns:: An iterator over the (non-unique) match list, starting at the beginning

std::vector<std::vector<int> >::iterator EndMList ( ) [inline]

Returns:: An iterator over the non-unique match list, set to the end

std::vector< std::vector< int > > & GetUMapList ( )

Returns:: the entire list of unique matches for this pattern A unique match is defined as one which does not cover the identical atoms that a previous match has covered.

For instance, the pattern [OD1]~C~[OD1] describes a carboxylate group. This pattern will match both atom number permutations of the carboxylate, and if GetMapList() is called, both matches will be returned. If GetUMapList() is called only unique matches of the pattern will be returned.

Referenced by OBChemTsfm::Apply(), OBBondTyper::AssignFunctionalGroupBonds(), OpenBabel::CorrectBadResonanceForm(), and OBAtom::MatchesSMARTS().

void WriteMapList ( std::ostream & ofs )

Debugging -- write a list of matches to the output stream.