Bug fixes and Optimizations

[CascadingRadium]

• Fix issue where vecIDsToExclude is always nil if the cache entry exists, allowing deleted documents to be returned.
• Optimize vector cache logic by checking for a cache entry after acquiring a write lock to reduce redundant work during concurrent access.
• Resolve issue of duplicate entries being added to dictLocs when constructing an in-memory segment.
• Clean up code by removing the unused method parameters that were being passed around.