Redundant copies of London Stage PDFs

[wintere]

The London Stage PDFs

directory	size	files	active
/PDFS/source/	835 MB	11	no
/images/pdfs/	1.92 GB	10,455	yes

Under the current repository structure, most pages of the HathiTrust scans of The London stage, 1660-1800... are stored twice in the website repository. The whole PDFs are archived in the /PDFS/source folder, while the individual pages generated by the JavaScript pipeline are stored in /images/pdfs. At serving time, images are retrieved from static copies stored in the /images/pdfs/ folder of the repository.

It is important to archive how and why the individual page images were generated. However, especially given the size of the source files (85% of the recommended maximum size for an active development repository) is not an ideal place to do so.

Proposals:

1. SHORT-TERM Move the PDFs/source/ folder and the Javascript pipeline to another repository, like the database-code or data. Remove the folder from the repository cache. Expected reduction in repository size: >=835 MB.
2. LONG-TERM Move the individual PDF pages in /images/ to an agreed upon third-party hosting service and out of this repository, too. Remove the folder from the repository cache. Expected reduction in repository size: >=1.92 GB.