Reduce website file count

### What problem does your feature request solve?

Our webserver serves files from a shared lustre file system, which has restrictions on how many files we can store. Large CSET runs will hit quota limits if they include a lot of files.

### Describe the solution you'd like

As part of the finalise webpage task tar up all of the plots and create an index json file recording their positions in the tarball. Use javascript to fetch the image directly out of the tarball using the [range](https://developer.mozilla.org/en-US/docs/Web/HTTP/Reference/Headers/Range) header.

We've done the same setup at NCI for obsmon and RES [(internal link)](https://git.nci.org.au/bom/ngm/tarwebplots).

### Describe alternatives you've considered