I haven't lost interest, and development is still in progress.

There have been few impediments to the developments these past few years though, mostly kids, and more recently a global pandemic. ;)

Apart from the warnings you mentionned, any specific error in the logs ?

Nothing related to memory ?

I can't completely exclude it (as I understand the value of the feature) but it's not in the backlog for now.

Ubooquity parses the OPF file contained in the Epub file (see Epub specifications for details).
It stores the following fields:

• dc:title
• dc:creator
• dc:identifier
• dc:publisher
• dc:description
• dc:subject
• dc:language
• dc:date

It also stores Calibre specific "rating" and "series" fields.

When searching books, Ubooquity looks at:

• title
• authors (taken from "dc:creator")
• file name
• tags (taken from "dc:subject")
• series name

I was expecting Ubooquity to be taking the metadata from the .opf file is that not the case?

Hello Morgan,

Ubooquity uses metadata that is stored inside the epub file.
It does not read external opf files.

Indeed.

The library currently used to read the PDF sometimes consumes a huge amount of memory for some files. The only workaround for now is to increase the memory allocated to Ubooquity.

Just in case: you didn't define an exclusion pattern in the advanced options ?

Caused by: java.net.BindException: Address already in use

You have to uUse different ports and paths when running several Ubooquity instances at the same time.

Try launching it with the --headless command line option.

--headless
Use it if you don’t want the Ubooquity interface to be displayed, or to avoid compatibility issues on devices with no graphical environment (e.g. NAS).

'Raw files' allows you to expose a simple listing of clickable files link to be able to download them.

No metadata, no thumbnails, nothing, just links.