bool CGridLabThresholdsLadder::TransferDataFromWindow()
{
bool bRtn = m_pData != NULL;
if(bRtn)
{
CLabLocusThreshold LocusLimits;
int nILS = GetNumberCols() - 1;
int nCol;
CLabRFU &rfu(*m_pData->GetRFUladder());
CLabRFU &rfuILS(*m_pData->GetRFUls());
CLabRFU *pRFU = &rfu;
// get defaults and ILS
for(nCol = 0; bRtn && (nCol <= nILS); nCol += nILS)
{
_GetColumn(nCol,&LocusLimits);
// setup defaults
pRFU->SetFractionMaxPeak(
LocusLimits.GetFractionMaxPeak());
pRFU->SetPullupFractionFilter(
LocusLimits.GetPullupFractionFilter());
pRFU->SetStutterThreshold(
LocusLimits.GetStutter());
pRFU->SetAdenylationThreshold(
LocusLimits.GetAdenylation());
// perform validation here
pRFU = &rfuILS;
}
for(nCol = 1; nCol < nILS; nCol++)
{
if(_GetColumn(nCol,&LocusLimits))
{
rfu.SetLocusThreshold(LocusLimits);
}
else
{
bRtn = false;
nCol = nILS; // loop exit
}
}
}
return bRtn;
}
bool CGridLabThresholdsSample::TransferDataFromWindow()
{
CLabLocusThreshold LocusLimits;
bool bRtn =
(m_pData != NULL) && _GetColumn(0,&LocusLimits);
if(bRtn)
{
CLabRFU &rfu(*m_pData->GetRFUsample());
// setup defaults
m_pData->SetHeterozygousImbalanceLimit(
LocusLimits.GetHeterozygousImbalanceLimit());
m_pData->SetMinBoundForHomozygote(
LocusLimits.GetMinBoundForHomozygote());
rfu.SetFractionMaxPeak(
LocusLimits.GetFractionMaxPeak());
rfu.SetPullupFractionFilter(
LocusLimits.GetPullupFractionFilter());
rfu.SetStutterThreshold(
LocusLimits.GetStutter());
rfu.SetPlusStutterThreshold(
LocusLimits.GetPlusStutter());
rfu.SetAdenylationThreshold(
LocusLimits.GetAdenylation());
// Setup Loci
int n = GetNumberCols();
for(int i = 1; i < n; i++)
{
if(_GetColumn(i,&LocusLimits))
{
rfu.SetLocusThreshold(LocusLimits);
}
else
{
bRtn = false;
i = n; // loop exit
}
}
}
return bRtn;
}
void FillNet(shared_ptr< Layer<float> > data_layer,
shared_ptr< Layer<float> > label_layer, int num_output) {
std::function<float()> rfu = GetRandomUniform<float>(-1.0, 1.0);
std::function<float()> riu = GetRandomUniform(0, num_output);
if (data_layer != NULL) {
std::vector<cv::Mat> images;
std::vector<int> labels;
shared_ptr<caffe::MemoryDataLayer<float>> data_layer_ptr =
boost::dynamic_pointer_cast<caffe::MemoryDataLayer<float>>(data_layer);
// int bn = data_layer_ptr->batch_size();
int bc = data_layer_ptr->channels();
int bh = data_layer_ptr->height();
int bw = data_layer_ptr->width();
cv::Mat image(bh, bw, CV_32FC(bc));
#pragma omp parallel for
for (int h = 0; h < bh; ++h) {
for (int w = 0; w < bw; ++w) {
for (int c = 0; c < bc; ++c) {
*(image.ptr<float>(h, w) + c) = rfu();
}
}
}
images.push_back(image);
labels.push_back(0);
data_layer_ptr->AddMatVector(images, labels);
}
if (label_layer != NULL) {
std::vector<cv::Mat> images;
std::vector<int> labels;
shared_ptr<caffe::MemoryDataLayer<float>> layer_ptr =
boost::dynamic_pointer_cast<caffe::MemoryDataLayer<float>>(label_layer);
// int bn = layer_ptr->batch_size();
int bc = layer_ptr->channels();
int bh = layer_ptr->height();
int bw = layer_ptr->width();
cv::Mat image(bh, bw, CV_32FC(bc));
#pragma omp parallel for
for (int h = 0; h < bh; ++h) {
for (int w = 0; w < bw; ++w) {
for (int c = 0; c < bc; ++c) {
*(image.ptr<float>(h, w) + c) = riu();
}
}
}
images.push_back(image);
labels.push_back(0);
layer_ptr->AddMatVector(images, labels);
}
}
