1 |
/********************************************************************** |
2 |
parsmart.h - Daylight SMARTS parser. |
3 |
|
4 |
Copyright (C) 1998-2001 by OpenEye Scientific Software, Inc. |
5 |
Some portions Copyright (C) 2001-2005 by Geoffrey R. Hutchison |
6 |
|
7 |
This file is part of the Open Babel project. |
8 |
For more information, see <http://openbabel.sourceforge.net/> |
9 |
|
10 |
This program is free software; you can redistribute it and/or modify |
11 |
it under the terms of the GNU General Public License as published by |
12 |
the Free Software Foundation version 2 of the License. |
13 |
|
14 |
This program is distributed in the hope that it will be useful, |
15 |
but WITHOUT ANY WARRANTY; without even the implied warranty of |
16 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
17 |
GNU General Public License for more details. |
18 |
***********************************************************************/ |
19 |
|
20 |
#ifndef OB_PARSMART_H |
21 |
#define OB_PARSMART_H |
22 |
|
23 |
#include <string> |
24 |
#include <vector> |
25 |
|
26 |
#include "mol.hpp" |
27 |
|
28 |
/*==========================*/ |
29 |
/* SMARTS Data Structures */ |
30 |
/*==========================*/ |
31 |
|
32 |
#define AE_LEAF 0x01 |
33 |
#define AE_RECUR 0x02 |
34 |
#define AE_NOT 0x03 |
35 |
#define AE_ANDHI 0x04 |
36 |
#define AE_OR 0x05 |
37 |
#define AE_ANDLO 0x06 |
38 |
|
39 |
#define AL_CONST 0x01 |
40 |
#define AL_MASS 0x02 |
41 |
#define AL_AROM 0x03 |
42 |
#define AL_ELEM 0x04 |
43 |
#define AL_HCOUNT 0x05 |
44 |
#define AL_NEGATIVE 0x06 |
45 |
#define AL_POSITIVE 0x07 |
46 |
#define AL_CONNECT 0x08 |
47 |
#define AL_DEGREE 0x09 |
48 |
#define AL_IMPLICIT 0x0a |
49 |
#define AL_RINGS 0x0b |
50 |
#define AL_SIZE 0x0c |
51 |
#define AL_VALENCE 0x0d |
52 |
#define AL_CHIRAL 0x0e |
53 |
#define AL_HYB 0x0f |
54 |
#define AL_CLOCKWISE 1 |
55 |
#define AL_ANTICLOCKWISE 2 |
56 |
|
57 |
namespace OpenBabel |
58 |
{ |
59 |
|
60 |
//! \brief A SMARTS parser internal atomic expression |
61 |
typedef union _AtomExpr { |
62 |
int type; |
63 |
struct |
64 |
{ |
65 |
int type; |
66 |
int prop; |
67 |
int value; |
68 |
} |
69 |
leaf; |
70 |
struct |
71 |
{ |
72 |
int type; |
73 |
void *recur; |
74 |
} |
75 |
recur; |
76 |
struct |
77 |
{ |
78 |
int type; |
79 |
union _AtomExpr *arg; |
80 |
} |
81 |
mon; |
82 |
struct |
83 |
{ |
84 |
int type; |
85 |
union _AtomExpr *lft; |
86 |
union _AtomExpr *rgt; |
87 |
} |
88 |
bin; |
89 |
} AtomExpr; |
90 |
|
91 |
#define BE_LEAF 0x01 |
92 |
#define BE_ANDHI 0x02 |
93 |
#define BE_ANDLO 0x03 |
94 |
#define BE_NOT 0x04 |
95 |
#define BE_OR 0x05 |
96 |
|
97 |
#define BL_CONST 0x01 |
98 |
#define BL_TYPE 0x02 |
99 |
|
100 |
#define BT_SINGLE 0x01 |
101 |
#define BT_DOUBLE 0x02 |
102 |
#define BT_TRIPLE 0x03 |
103 |
#define BT_AROM 0x04 |
104 |
#define BT_UP 0x05 |
105 |
#define BT_DOWN 0x06 |
106 |
#define BT_UPUNSPEC 0x07 |
107 |
#define BT_DOWNUNSPEC 0x08 |
108 |
#define BT_RING 0x09 |
109 |
|
110 |
//! \brief A SMARTS parser internal bond expression |
111 |
typedef union _BondExpr { |
112 |
int type; |
113 |
struct |
114 |
{ |
115 |
int type; |
116 |
int prop; |
117 |
int value; |
118 |
} |
119 |
leaf; |
120 |
struct |
121 |
{ |
122 |
int type; |
123 |
union _BondExpr *arg; |
124 |
} |
125 |
mon; |
126 |
struct |
127 |
{ |
128 |
int type; |
129 |
union _BondExpr *lft; |
130 |
union _BondExpr *rgt; |
131 |
} |
132 |
bin; |
133 |
} BondExpr; |
134 |
|
135 |
//! \brief A SMARTS parser internal bond specification |
136 |
typedef struct |
137 |
{ |
138 |
BondExpr *expr; |
139 |
int src,dst; |
140 |
int visit; |
141 |
bool grow; |
142 |
} |
143 |
BondSpec; |
144 |
|
145 |
//! \brief A SMARTS parser internal bond specification |
146 |
typedef struct |
147 |
{ |
148 |
AtomExpr *expr; |
149 |
int visit; |
150 |
int part; |
151 |
int chiral_flag; |
152 |
int vb; |
153 |
} |
154 |
AtomSpec; |
155 |
|
156 |
//! \brief A SMARTS parser internal pattern |
157 |
typedef struct |
158 |
{ |
159 |
int aalloc,acount; |
160 |
int balloc,bcount; |
161 |
bool ischiral; |
162 |
AtomSpec *atom; |
163 |
BondSpec *bond; |
164 |
int parts; |
165 |
} |
166 |
Pattern; |
167 |
|
168 |
// class introduction in parsmart.cpp |
169 |
//! \brief SMARTS (SMiles ARbitrary Target Specification) substructure searching |
170 |
class OBAPI OBSmartsPattern |
171 |
{ |
172 |
protected: |
173 |
std::vector<bool> _growbond; |
174 |
std::vector<std::vector<int> > _mlist; |
175 |
Pattern *_pat; |
176 |
std::string _str; |
177 |
|
178 |
public: |
179 |
OBSmartsPattern() |
180 |
{ |
181 |
_pat=NULL; |
182 |
} |
183 |
virtual ~OBSmartsPattern(); |
184 |
|
185 |
OBSmartsPattern(const OBSmartsPattern& cp) |
186 |
{ |
187 |
_pat = NULL; |
188 |
*this = cp; |
189 |
} |
190 |
OBSmartsPattern& operator=(const OBSmartsPattern& cp) |
191 |
{ |
192 |
if (_pat) |
193 |
delete [] _pat; |
194 |
_pat = NULL; |
195 |
std::string s = cp._str; |
196 |
Init(s); |
197 |
return (*this); |
198 |
} |
199 |
|
200 |
unsigned int NumMatches() const |
201 |
{ |
202 |
return (unsigned int)_mlist.size(); |
203 |
} |
204 |
unsigned int NumAtoms() const |
205 |
{ |
206 |
return _pat ? _pat->acount : 0; |
207 |
} |
208 |
unsigned int NumBonds() const |
209 |
{ |
210 |
return _pat ? _pat->bcount : 0; |
211 |
} |
212 |
|
213 |
int GetAtomicNum(int); |
214 |
void GetBond(int&,int&,int&,int); |
215 |
int GetCharge(int); |
216 |
const std::string &GetSMARTS() const |
217 |
{ |
218 |
return _str; |
219 |
} |
220 |
std::string &GetSMARTS() |
221 |
{ |
222 |
return _str; |
223 |
} |
224 |
int GetVectorBinding(int idx) const |
225 |
{ |
226 |
return(_pat->atom[idx].vb); |
227 |
} |
228 |
bool Empty() const |
229 |
{ |
230 |
return(_pat == NULL); |
231 |
} |
232 |
bool IsValid() const |
233 |
{ |
234 |
return(_pat != NULL); |
235 |
} |
236 |
bool Init(const char*); |
237 |
bool Init(const std::string&); |
238 |
void WriteMapList(std::ostream&); |
239 |
|
240 |
bool Match(OBMol &mol, bool single=false); |
241 |
bool RestrictedMatch(OBMol &mol, std::vector<std::pair<int,int> > &pairs, bool single=false); |
242 |
bool RestrictedMatch(OBMol &mol, OBBitVec &bv, bool single=false); |
243 |
|
244 |
std::vector<std::vector<int> > &GetMapList() |
245 |
{ |
246 |
return(_mlist); |
247 |
} |
248 |
std::vector<std::vector<int> > &GetUMapList(); |
249 |
std::vector<std::vector<int> >::iterator BeginMList() |
250 |
{ |
251 |
return(_mlist.begin()); |
252 |
} |
253 |
std::vector<std::vector<int> >::iterator EndMList() |
254 |
{ |
255 |
return(_mlist.end()); |
256 |
} |
257 |
}; |
258 |
|
259 |
//! Performs fast, exhaustive matching used to find just a single match in match() using recursion and explicit stack handling. |
260 |
class OBAPI OBSSMatch //used for fast exhaustive matching |
261 |
{ |
262 |
protected: |
263 |
bool *_uatoms; |
264 |
OBMol *_mol; |
265 |
Pattern *_pat; |
266 |
std::vector<int> _map; |
267 |
|
268 |
public: |
269 |
OBSSMatch(OBMol&,Pattern*); |
270 |
~OBSSMatch(); |
271 |
void Match(std::vector<std::vector<int> > &v, int bidx=-1); |
272 |
}; |
273 |
|
274 |
void SmartsLexReplace(std::string &, |
275 |
std::vector<std::pair<std::string,std::string> > &); |
276 |
|
277 |
} // end namespace OpenBabel |
278 |
|
279 |
#endif // OB_PARSMART_H |
280 |
|
281 |
//! \file parsmart.h |
282 |
//! \brief Daylight SMARTS parser. |