On 2012-09-15 05:40, Greg Landrum wrote:
Hi Jan,
On Sat, Sep 15, 2012 at 12:47 AM, Jan Holst Jensen
<j...@biochemfusion.com> wrote:
I have now managed to produce a nice little self-contained DLL that uses
RDKit to do substructure mapping. It receives a query and a target molecule
as two MDL molfile strings. When the inputs are valid it behaves well and
returns the atom mapping to me. Yes :-).
Excellent!
However, when I give it an invalid MDL molfile as input, e.g. a text file
with regular prose, MolDataStreamToMol() throws a highly unexpected error
message.
This piece of my test code:
ifstream query_file ("C:\\SubstructData\\results.txt");
auto_ptr<RWMol> query (NULL);
unsigned int line = 0;
try {
query.reset( MolDataStreamToMol(query_file, line, false, false,
false) );
} catch (std::exception& e) {
cout << "STD-ERROR: " << e.what() << endl;
return 1;
} catch (...) {
cout << "UNK-ERROR: Unknown exception." << endl;
return 1;
}
produces this output:
[00:13:24] CTAB version string invalid at line 4
STD-ERROR: Unknown exception
In MolFileParser.cpp line 1881 I can locate the reported error message:
std::ostringstream errout;
errout<<"CTAB version string invalid at line "<<line;
if(strictParsing){
if(res) delete res;
throw FileParseException(errout.str());
} else {
BOOST_LOG(rdWarningLog) << errout.str() << std::endl;
}
Since I run with strictParsing==false the BOOST_LOG(rdWarningLog) call must
be the one that emits the first line of output, and I assume that the boost
warning log does not throw an exception. I haven't been able to figure out
where the exception gets thrown, but that must be later in the code.
If I switch to strict parsing (last parameter in MolDataStreamToMol() set to
true) I don't see the boost log message anymore, only the single line
"STD-ERROR: Unknown exception". So strict parsing actually yields less error
message context.
I hope the exception error message can be changed to something more
meaningful ?
I would guess that the exception being thrown is a FileParseException.
These don't do anything to override the what() method, so you get the
"unknown exception" text. I will fix that.
Here's the code from the python wrapper for constructing a molecule
from a mol block
($RDBASE/Code/GraphMol/Wrap/rdmolfiles.cpp:MolFromMolFile()). It
catches the exceptions that are likely to arise and returns a NULL if
something fails. This might be useful:
RWMol *newM=0;
try {
newM = MolFileToMol(molFilename, sanitize,removeHs,strictParsing);
} catch (RDKit::BadFileException &e) {
PyErr_SetString(PyExc_IOError,e.message());
throw python::error_already_set();
} catch (RDKit::FileParseException &e) {
BOOST_LOG(rdWarningLog) << e.message() <<std::endl;
} catch (...) {
}
return static_cast<ROMol *>(newM);
-greg
Hi Greg,
Ah, so the actual message is in .message() instead of .what() - that
explains it. I changed my test code to:
try {
query.reset( MolDataStreamToMol(query_file, line, false,
false, true) );
* } catch (RDKit::BadFileException &e) {**
** cout << "RDK-ERROR: " << e.message() << endl;**
** return 1;**
** } catch (RDKit::FileParseException &e) {**
** cout << "RDK-ERROR: " << e.message() << endl;**
** return 1;*
} catch (std::exception& e) {
cout << "STD-ERROR: " << e.what() << endl;
return 1;
} catch (...) {
cout << "UNK-ERROR: Unknown exception." << endl;
return 1;
}
and now I get this output:
RDK-ERROR: CTAB version string invalid at line 4
Problem solved for now - thanks :-). Having the actual message available
in .what() as well in future will be nice though.
Cheers
-- Jan
------------------------------------------------------------------------------
How fast is your code?
3 out of 4 devs don\\\'t know how their code performs in production.
Find out how slow your code is with AppDynamics Lite.
http://ad.doubleclick.net/clk;262219672;13503038;z?
http://info.appdynamics.com/FreeJavaPerformanceDownload.html
_______________________________________________
Rdkit-discuss mailing list
Rdkit-discuss@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
