Use std::optional for peak fitting (which can fail) and other minor refactorings

[lobis]

New we skip the peak if it's not being found, instead of returning energy -1

[mariajmz]

Some examples of TRestDetectorSignalEvents analyzed with a macro using TRestDetectorSignal::GetPeakGauss():

void DrawEvent(TRestDetectorSignalEvent* event, TCanvas* canvas) {
    Int_t numberOfSignals = event->GetNumberOfSignals();
    std::cout << "Number of signals in event: " << numberOfSignals << std::endl;
    canvas->Clear();
    TPad* pad = event->DrawEvent();
    if (pad) {
        pad->SetCanvas(canvas);
        canvas->cd();
        pad->Draw();
        pad->cd();
    } else {
        std::cout << "Failed to draw event, pad is null." << std::endl;
    }
    
    for (int i = 0; i < numberOfSignals; i++) {
        TRestDetectorSignal* signal = event->GetSignal(i);
        std::optional<std::pair<Double_t, Double_t>> maxGaussOpt = signal->GetPeakGauss();
        if (maxGaussOpt) {
            auto maxGauss = maxGaussOpt.value();
            std::cout << "Max peak for signal " << i << " : " << maxGauss.first << " " << maxGauss.second << std::endl;
            TMarker* marker = new TMarker(maxGauss.first, maxGauss.second, 20);
            marker->SetMarkerColor(2);
            marker->Draw("same");
        }
    }

    canvas->Update();
}

20 ns of sampling time:

40 ns of sampling time:

The previous event, when I fit it with the method GetMaxGauss of mariajmz_fits branch, I get this:

Another event:

Also, I tried using gaussFit in TRestDetectorSignalToHitsProcess in a rml file with the same events and I get the usual errors when the fit fails:

Plotting the spectra (hitsAna_energy) with this branch and mariajmz_fits, they look similar:

This branch:

mariajmz_fits branch: