Impala Documentation

Table::QuidTable* Impala::Core::DataFactory::MakeTrainFeatureQuids ( ) [inline]

There is a problem with the regression tests for Core/Training: The test reuses data from another set, that's why the setid in the quids is wrong.

There is a hack in here to fix it.

Definition at line 171 of file DataFactory.h.

References Impala::Core::VideoSet::FilterRkf(), Impala::Core::Table::TableTem< Col1T, Col2T, Col3T, Col4T, Col5T, Col6T, Col7T, Col8T, Col9T >::Get1(), Impala::Core::Database::RawDataSet::GetSetId(), ILOG_DEBUG, MakeAllFeatureQuids(), mDataSet, mQuidClass, Impala::QUID_CLASS_IMAGE, Impala::QuidSet(), Impala::Core::Table::QuidTable::SetQuidSet(), and Impala::Core::Table::Table::Size().

Referenced by Impala::Core::Training::TrainDataSrcFeature::SelectValid().

00172     {
00173         Table::QuidTable * quids = MakeAllFeatureQuids();
00174         ILOG_DEBUG("MakeTrainFeatureQuids(): got "<< quids->Size() <<" quids");
00175         if(quids->Size() > 0)
00176         {
00177             int oldSetId = QuidSet(quids->Get1(0));
00178             quids->SetQuidSet(mDataSet->GetSetId());
00179             if(mQuidClass != QUID_CLASS_IMAGE)
00180                 VideoSet::FilterRkf(quids, static_cast<VideoSet::VideoSet*>(mDataSet));
00181             quids->SetQuidSet(oldSetId);
00182         }
00183         return quids;
00184     }

Here is the call graph for this function: