In the case that the association of objects/traces between the red and blue sides of DBSP is incorrect, you can use the script dbsp_splice to manually stitch together two coadded spectra.
You must provide the path to the raw files, so that the raw headers can be packaged into the final data product, as well as the filename to save the spliced spectrum to, and the red and blue coadds you wish to splice together.
Note that although the red_file and blue_file arguments are optional, you must supply at least one of them.
Typically the splicing between arms is quite good, and you only need to adjust it when the optimal extraction FWHM vary greatly between the red and blue sides. This can happen when an extended object is near another object on the slit. You can check the extraction FWHM visually by looking at the Extraction QA page.
In this GUI, the blue and red spectra are plotted as well as the spliced spectrum. After using the usual Matplotlib tools to zoom and pan to the region of interest, you can adjust the red multiplier to multiply the red flux by a constant, and change if flux will be interpolated across detector gaps, as well as save your changes to the same file you opened. After you close the GUI, you will be presented with the next spectrum to adjust the splicing of.
The default splicing that DBSP_DRP does uses a red multiplier of 1, since the flux calibration provided by PypeIt is excellent.
