A proof of concept metadata search shouldn't be too hard to whip up, on the basis of some simple stuff along these lines:
Code:
SELECT * FROM bookinfo WHERE filename REGEXP 'README.*'
The main problem is gathering the metadata and keeping it up to date. One time I was unable to use my H2O for *at least* 10 minutes after I unplugged its USB. That may be okay if you purposefully initiate a full scan, but otherwise…
Also see
https://github.com/koreader/koreader/issues/6922
Edit: GLOB is supported out of the box, forget REGEXP.