- Defined PK constraints for tables where they were missed

- Added additional index to hive_posts (indexing single permlink_id field) to speedup insertion
- Removed not needed multi column indices from hive_account table
- Introduced function `process_hive_post_operation' covering new-post, post-edit, post-undelete scenarios in single query at SQL side
- Introduced function `delete_hive_post' to simplify post deletion at client side and improve performance.
- Improved reblog operation queries (eliminated separate queries for IDs to be next processed/updated)
- Improved follow operation queries (eliminated separate queries for IDs to be next processed/updated). Prerequisites for value caching and flush at Python side, not finished yet