[Q] how does the content: query filter search/index content?

[0n0mat0p03ia]

I recently used content:"foo" to query a folder containing epub files. To my surprise, Cardinal returned a list of epubs that contained "foo" (although it did not return all the epubs that contained "foo"). Behind the scenes, how is Cardinal searching/indexing file contents?

Edit: it's worth noting that Cardinal spiked every CPU core to 100% when querying ext:epub content:"foo" on a folder containing ~350 epubs. (System responsiveness wasn't really affected. I wouldn't have known if I didn't happen to be looking my stats app at the same time).

[ldm0]

Behind the scenes, how is Cardinal searching/indexing file contents?

Cardinal doesn't indexing file contents actually. Currently Cardinal just scans for file content in parallel during searching.

Cardinal spiked every CPU core to 100% when querying ext:epub content:"foo"

Yeah, that's the expected behaviour. I think we could also use spotlight cache here actually(wavering), which is less accurate but faster and less CPU consuming.

[0n0mat0p03ia]

I think we could also use spotlight cache here actually(wavering), which is less accurate but faster and less CPU consuming.

Spotlight doesn't index epubs (my use case. It might be possible using a third-party metadata importer). But otherwise it might be a better option. Or perhaps allow the user to choose Spotlight vs. Cardinal content search via a setting or run-time flag.