Replace the initialize_db.sh bash script with a Rust-based implementation

[goshawk-3]

Description

The current database initialization, including schema migrations and importing large key files (e.g., a 1.1 GB SNS PK and a 0.3 GB SKS), is handled via a bash script.

However, much of this logic—especially the large object import—is already implemented more efficiently in the Rust codebase. Maintaining duplicate logic in both Bash and Rust adds complexity and overhead.

This task proposes extending the existing coprocessor/utils binary to handle the full initialization flow, including migrations and key insertion.