align python ffi with db.h abi, fix struct overflow bugs, bump to 0.11.1

[guycipher]

fix dbstats struct missing eight trailing write amplification fields which caused tidesdb get db stats to write past the python allocated buffer and corrupt the heap on every call. fix config struct missing finish compactions on close which made tidesdb open read past the buffer. add the seven trailing write amplification counters to the cf stats struct. all six structs now match the c library sizeof byte for byte.

add missing bindings, init, finalize, raise open file limit, compression available, cancel background work, objstore s3 create and objstore s3 create config. the s3 creators are bound lazily so import still works on builds without s3 support and raise a clear error instead of crashing.

add finish compactions on close config field across the dataclass, init, open and default config. add the new write amplification fields to stats and db stats with parsing. add tdb err busy error code and message. free txn get values and column family name lists through tidesdb free instead of libc free so a custom allocator stays correct.

fix pre existing test that asserted max concurrent flushes default was non zero when the library now reports zero meaning use the library default.

add sixteen tests covering the new surface, all one hundred thirty pass. extend the python reference docs with init and process limits, fast shutdown, the s3 connector, compression available, finish compactions on close, the write amplification fields and a full error code table.