How does Tabby parse/process the git repository codes into context of code search?

[ANYMS-A]

Hi, I recently tried to figure out that: how Tabby process a git repo (to AST) and save the processed into the Database or VectorDB for context searching at code completion stage？ May I ask which part of the code contains the implementation of this feature?

I saw some CodeSearchSchema declaration in the crates/tabby-common/src/index.rs and had a rough view about the Schema, but I'd like to know more detail about the whole pipeline of how a git repo of code is processed and stored.

Thanks!