Add more plots regarding sky line profiles and scaling in skySubtractionQa

To obtain more sky subtraction QA information, specifically regarding sky line profile modeling and scaling, I propose adding options to the so-called “Jared” QA plots (or the “median chis by sky brightness percentile” plots).
1: eliminate “sky continuum” pixels. These pixels occupy a significant fraction of the pixels analyzed in the current plots and are likely affected by other QA items (e.g., scattering correction) as well. By eliminating these pixels, we want to focus more on sky line modeling and scaling.
2: in addition, revise the calculation of the “percentile”. Currently, it is calculated with respect to the peak of the entire spectrum, including faint and bright sky lines as well as the sky continuum. We also want to know the distribution of chis at different levels of the “normalized” sky lines (or LSF). This avoids comparing chis near the peaks of faint sky lines with the LSF tail of bright sky lines.