Templates: separate file, wavelength spacing

[kgozman6159]

Hello! I know it's been ~10 years since this issue was last brought up (ie #133), but I was wondering if there is now a way to read in templates from another file? I have something like 100 templates (I am trying to fit individual stellar velocities as opposed to galaxy redshifts, so I have my own set of specific synthetic spectra I'd like to use), and it is pain when all the spectral fluxes are just hardcoded into the templates.js file -- I wrote a python script to write the javascript dictionary part of the templates.js file for my given set of templates which I then copy/paste into the .js file, but this makes the file huge and laggy (though it works -- I am able to use my own set of templates). I tried reformatting the code to read in the template arrays from another file, but I don't know enough about javascript (this is my first time trying to read/modify js code) to make it work...

Another small question: if I understand correctly, the template arrays just take in the starting and ending wavelengths of the templates and whether the spacing is log-linear or not, but there's nowhere that specifies what the actual spacing is between wavelengths -- does it just assume that the templates are rebinned with one angstrom per bin/pixel? And is end_wavelength inclusive or exclusive?

Thank you!

[Samreay]

Hi Katya! Sorry for the delay responding to this, it's been a long couple of weeks! Unfortunately its probably even harder to do this now than it would have been when Marz was first written, because the Angular version I wrote it all in well and truly beyond its support window (making it harder to jump back in as all the guides are for far more modern versions). Technically I can't see why Marz couldn't be updated so a template file could be dropped into the same place as the FITS files go, the issue simply comes down to writing this code. I wouldn't know where to start any more (reading code is hard enough after six months have passed, let alone ten years in which I haven't touched JS or this project).

If you've already got the project forked and have augmented the templates with your own (to create, as you said, a massive and laggy JS file), that is probably as good as we can get without substantial refactoring.

the actual spacing is between wavelengths

It's assumed to be linear spacing (well linear either in... linear space or log-space, depending on that flag you noticed).

does it just assume that the templates are rebinned with one angstrom per bin/pixel

Nah it uses the length of the spec and then creates a linear scale (with inclusive start and end) between the start and end lambda. You can see this in processTemplates if curious.