Optimize python client performance close to Go SDK

[wenhaocs]

As titled.

This includes the performance enhancement on decoding. Due to GIL, the performance of Python SDK can be worse than other programing language.

[wenhaocs]

Can we use Free-threaded build of python?

Starting with the 3.13 release, CPython has support for a build of Python called free threading where the global interpreter lock (GIL) is disabled. Free-threaded execution allows for full utilization of the available processing power by running threads in parallel on available CPU cores. While not all software will benefit from this automatically, programs designed with threading in mind will run faster on multi-core hardware.

Starting with the 3.13 release, CPython has support for a build of Python called free threading where the global interpreter lock (GIL) is disabled. Free-threaded execution allows for full utilization of the available processing power by running threads in parallel on available CPU cores. While not all software will benefit from this automatically, programs designed with threading in mind will run faster on multi-core hardware.

[wenhaocs]

Conclusion:

(1) Python 3.14.0 is quite new, and the versions commonly used by most users are lower than this.
(2) The latency on decoding will become an overhead only when you return a large volume of data each request, while latency at server side is small. Otherwise, you won't sense it a lot.
(3) A commonly used way to walk around GIL is multiprocessing. Consider to spawn multiple processes - each one with a NebulaPool, if you observe there is bottleneck at client side.