bool Scanner::isSingleLine (Scanner &scanner)
{
int pos = scanner.tell();
scanner.skipLine();
bool res = scanner.isEOF();
scanner.seek(pos, SEEK_SET);
return res;
}
/* Read full (with hash chains, for future modifying) *
* lzw dictionary function */
void LzwDict::load( Scanner &_scanner )
{
int i, j, HashCode, n, k;
_modified = false;
_alphabetSize = _scanner.readBinaryInt();
_nextCode = _scanner.readBinaryInt();
n = _scanner.readBinaryInt();
_bitcodeSize = _scanner.readBinaryInt();
_maxCode = (1 << _bitcodeSize) - 1;
_storage.clear();
_storage.resize(n);
for (i = 0; i < n; i++)
{
k = _scanner.readBinaryDword();
_storage[i].Prefix = k;
_storage[i].AppendChar = _scanner.readByte();
}
_freePtr = _scanner.readBinaryInt();
_hashKeys.clear_resize(SIZE);
_nextPointers.clear_resize(SIZE);
for (i = 0; i < SIZE; i++)
{
_nextPointers[i] = -1;
_hashKeys[i] = -1;
}
while (!_scanner.isEOF())
{
HashCode = _scanner.readBinaryInt();
j = _scanner.readBinaryInt();
_hashKeys[HashCode] = j;
i = j;
j = _scanner.readBinaryInt();
while (j != -1)
{
_nextPointers[i] = j;
i = j;
j = _scanner.readBinaryInt();
}
}
}
void GrossFormula::fromString (Scanner &scanner, Array<int> &gross)
{
gross.clear_resize(ELEM_MAX);
gross.zerofill();
scanner.skipSpace();
while (!scanner.isEOF())
{
int elem = Element::read(scanner);
scanner.skipSpace();
int counter = 1;
if (isdigit(scanner.lookNext()))
{
counter = scanner.readUnsigned();
scanner.skipSpace();
}
gross[elem] += counter;
}
}
void _importSMILES (OracleEnv &env, const char *table, const char *smiles_col,
const char *id_col, const char *file_name)
{
FileScanner fscanner(file_name);
AutoPtr<GZipScanner> gzscanner;
Scanner *scanner;
int nwritten = 0;
QS_DEF(Array<char>, id);
QS_DEF(Array<char>, str);
env.dbgPrintfTS("importing into table %s\n", table);
// detect if input is gzipped
byte magic[2];
int pos = fscanner.tell();
fscanner.readCharsFix(2, (char *)magic);
fscanner.seek(pos, SEEK_SET);
if (magic[0] == 0x1f && magic[1] == 0x8b)
{
gzscanner.reset(new GZipScanner(fscanner));
scanner = gzscanner.get();
}
else
scanner = &fscanner;
while (!scanner->isEOF())
{
id.clear();
scanner->readLine(str, false);
BufferScanner strscan(str);
strscan.skipSpace();
while (!strscan.isEOF() && !isspace(strscan.readChar()))
;
strscan.skipSpace();
if (strscan.lookNext() == '|')
{
strscan.readChar();
while (!strscan.isEOF() && strscan.readChar() != '|')
;
strscan.skipSpace();
}
if (!strscan.isEOF() && id_col != 0)
strscan.readLine(id, true);
OracleStatement statement(env);
statement.append("INSERT INTO %s(%s", table, smiles_col);
if (id_col != 0)
statement.append(", %s", id_col);
statement.append(") VALUES(:smiles");
if (id_col != 0)
{
if (id.size() > 1)
statement.append(", :id");
else
statement.append(", NULL");
}
statement.append(")");
statement.prepare();
str.push(0);
statement.bindStringByName(":smiles", str.ptr(), str.size());
if (id.size() > 1)
statement.bindStringByName(":id", id.ptr(), id.size());
statement.execute();
nwritten++;
if (nwritten % 1000 == 0)
{
env.dbgPrintfTS("imported %d items, commiting\n", nwritten);
OracleStatement::executeSingle(env, "COMMIT");
}
}
if (nwritten % 1000 != 0)
{
env.dbgPrintfTS("imported %d items, commiting\n", nwritten);
OracleStatement::executeSingle(env, "COMMIT");
}
}
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;
}
