cropperjs ignores attempts to modify the data before ready. This change
migrates setup to a one-time 'ready' callback.
Steps:
1. Change the cropping bounds of page 1
2. Change to page 2
3. Change to page 1
4. Change to page 2
5. Download cropped PDF
Observed behavior:
* (3) shows the default (autocrop) bounds
* (5) results in a download without the set bounds
Expected behavior:
* (3) should restore the bounds already set
* (4) should not override the set bounds with the autocrop
Tested:
* Local builds using the docker dev config
