Is it possible to search the indexes inside a `.ck` folder without the existence of actual project files?

[Linloir]

I'm using CK with CI pipelines which automatically generate .ck indexes for every commit made to the repository.
The .ck indexes are stored elsewhere that is not in the original repository.

Currently I'm trying to embed the semantic search feature into my agent's workflow, and I want to only download the pre-packed .ck index without downloading the repository, but I can't see an option in the CLI interface.

I'm wondering whether the underlying implementation allows this feature, that is does ck --index stores original chunk texts or just stores things like line_nos which requires the original files to fetch the original chunk text?

[runonthespot]

Currently it doesn't chunk the text - it maintains span references, and a hash for each chunk, for storing the information - the reasoning being that a chunk of text is just a redundant copy (it also allows us to create redundant overlapping chunks + semantic chunks at significantly less extra cost. If you could explain the use case in a bit more detail, we could think about how we could facilitate you better?