SdfLoader::SdfLoader (Scanner &scanner) :
TL_CP_GET(data),
TL_CP_GET(properties),
TL_CP_GET(_offsets),
TL_CP_GET(_preread)
{
data.clear();
properties.clear();
// detect if input is gzipped
byte id[2];
int pos = scanner.tell();
scanner.readCharsFix(2, (char *)id);
scanner.seek(pos, SEEK_SET);
if (id[0] == 0x1f && id[1] == 0x8b)
{
_scanner = new GZipScanner(scanner);
_own_scanner = true;
}
else
{
_scanner = &scanner;
_own_scanner = false;
}
_current_number = 0;
_max_offset = 0;
_offsets.clear();
_preread.clear();
}
bool Scanner::isSingleLine (Scanner &scanner)
{
int pos = scanner.tell();
scanner.skipLine();
bool res = scanner.isEOF();
scanner.seek(pos, SEEK_SET);
return res;
}
SdfLoader::SdfLoader (Scanner &scanner) :
CP_INIT,
TL_CP_GET(data),
TL_CP_GET(properties),
TL_CP_GET(_offsets),
TL_CP_GET(_preread)
{
data.clear();
properties.clear();
// detect if input is gzipped
byte id[2];
int pos = scanner.tell();
scanner.readCharsFix(2, (char *)id);
scanner.seek(pos, SEEK_SET);
if (id[0] == 0x1f && id[1] == 0x8b)
{
_scanner = new GZipScanner(scanner);
_own_scanner = true;
}
else
{
_scanner = &scanner;
_own_scanner = false;
}
_current_number = 0;
_max_offset = 0;
_offsets.clear();
_preread.clear();
}
SdfLoader::~SdfLoader()
{
if (_own_scanner)
delete _scanner;
}
int SdfLoader::tell ()
{
return _scanner->tell();
}
int SdfLoader::currentNumber ()
{
return _current_number;
}
int SdfLoader::count ()
{
int offset = _scanner->tell();
int cn = _current_number;
if (offset != _max_offset)
{
_scanner->seek(_max_offset, SEEK_SET);
_preread.clear();
_current_number = _offsets.size();
}
while (!isEOF())
readNext();
int res = _current_number;
if (res != cn)
{
_scanner->seek(offset, SEEK_SET);
_preread.clear();
_current_number = cn;
}
return res;
}
bool MoleculeAutoLoader::tryMDLCT (Scanner &scanner, Array<char> &outbuf)
{
int pos = scanner.tell();
bool endmark = false;
QS_DEF(Array<char>, curline);
outbuf.clear();
while (!scanner.isEOF())
{
int len = scanner.readByte();
if (len > 90) // Molfiles and Rxnfiles actually have 80 characters limit
{
scanner.seek(pos, SEEK_SET);
// Garbage after endmark means end of data.
// (See the note below about endmarks)
if (endmark)
return true;
return false;
}
curline.clear();
while (len-- > 0)
{
if (scanner.isEOF())
{
scanner.seek(pos, SEEK_SET);
return false;
}
int c = scanner.readChar();
curline.push(c);
}
curline.push(0);
if (endmark)
{
// We can not properly read the data to the end as there
// is often garbage after the actual MDLCT data.
// Instead, we are doing this lousy check:
// "M END" or "$END MOL" can be followed only
// by "$END CTAB" (in Markush queries), or
// by "$MOL" (in Rxnfiles). Otherwise, this
// is actually the end of data.
if (strcmp(curline.ptr(), "$END CTAB") != 0 &&
strcmp(curline.ptr(), "$MOL") != 0)
{
scanner.seek(pos, SEEK_SET);
return true;
}
}
if (strcmp(curline.ptr(), "M END") == 0)
endmark = true;
else if (strcmp(curline.ptr(), "$END MOL") == 0)
endmark = true;
else
endmark = false;
outbuf.appendString(curline.ptr(), false);
outbuf.push('\n');
}
scanner.seek(pos, SEEK_SET);
// It happened once that a valid Molfile had successfully
// made its way through the above while() cycle, and thus
// falsely recognized as MDLCT. To fight this case, we include
// here a check that the last line was actually an endmark
return endmark;
}
