perf: Refine the Model Manager code

[shaohuzhang1]

perf: Refine the Model Manager code

[shaohuzhang1]

There is no error in this code, but there is an inefficiency in obtaining keys. If a key pair already exists in the cache, it does not need to fetch one from the database each time because rsa_cache.set(cache_key, rsa_value) will always overwrite the existing value in the cache even if the key pair hasn't changed.

To optimize this function, you can modify it so that only calls to get_key_pair_by_sql() actually fetch data from the database:

def generate():
    def get_key_pair():
        rsa_value = rsa_cache.get(cache_key)
        if rsa_value is None:
            with lock:
                rsa_value = rsa_cache.get(cache_key)
                if rsa_value is not None:
                    return rsa_value
                # Fetch keys from database only if they're missing.
                rsa_value = get_key_pair_by_sql() if not has_new_keys else None
                if rsa_value is not None:
                    rsa_cache.set(cache_key, rsa_value)
                    has_new_keys = True  # Reset flag after adding new values.
        return rsa_value

    has_new_keys = False
    generate()

If we know in advance when a full update is likely (for instance, during system startup), has_new_keys could be set once at the start of application initialization phase instead of checking every access to generate. This avoids unnecessary queries while still providing quick response on demand unless a full refresh might be needed sooner due to some event.