improve ddb TransactWriteItems performance + small refactor

[bentsku]

Motivation

Follow up from #10415 and especially #9410 where we refactored the logic for BatchWriteItem.
TransactWriteItems had not been touched by the new changes, and should improve a lot from this.

Changes

• use the same logic than BatchWriteItem, but slightly modified to account for the differences
• add a new test for streaming with TransactWriteItems as there were none
• refactor the logic to use structural pattern matching, very well adapted to this use case with unknown input
• removed the calls to deepcopy in the event creation in favor of just creating a new event (avoid pickling/unpickling).

Benchmarks

Funnily enough, the changes are smaller than expected: Twisted, directly calling DynamoDB local without going through the Gateway, plus Kinesis batching relieving pressure on the gateway already improved the speed of TransactWriteItems by quite a lot compared to for example 3.2, so the gains are not so massive as we've been used to (still around 2x to 4x).

I've added 3.2.0 to the benchmarks to show how far we've come in around 2 weeks.

latest and with fix are done with Twisted. 3.2.0 did not have it yet, so it's still with hypercorn.

No Stream	3.2.0	latest	with fix
Throughput (higher is better)
TransactWriteItems (10) throughput item/s	365.61	1169.23	2964.81
TransactWriteItems (25) throughput item/s	286.30	1459.05	5075.70
Total time (lower is better)
TransactWriteItems (10) total seconds (5k items)	27.35	4.2763	1.6864
TransactWriteItems (25) total seconds (12k5 items)	43.66	8.5672	2.4627

With DDB Streams	3.2.0	latest	with fix
Throughput (higher is better)
TransactWriteItems (10) throughput item/s	99.41	835.92	1492.23
TransactWriteItems (25) throughput item/s	82.77	1149.21	2746.00
Total time (lower is better)
TransactWriteItems (10) total seconds (5k items)	50.29	5.9814	3.3507
TransactWriteItems (25) total seconds (12k5 items)	151.02	10.8770	4.5521

With Kinesis Streams	3.2.0	latest	with fix
Throughput (higher is better)
TransactWriteItems (10) throughput item/s	102.85	899.99	1679.19
TransactWriteItems (25) throughput item/s	85.54	1261.92	3251.62
Total time (lower is better)
TransactWriteItems (10) total seconds (5k items)	48.6138	5.5556	2.9776
TransactWriteItems (25) total seconds (12k5 items)	146.1307	9.9055	3.8442

So in the end, we can still provide a 2x to 3.5x improvement over latest.

For a batch size of 25 with no streams, we already provided a 5x improvement with latest, pushed to 17.5x.
For a batch size of 25 with streams enabled, we already provided a 15x improvement, pushed to an average of 35x.

[github-actions]

LocalStack Community integration with Pro

    2 files  ± 0      2 suites  ±0   1h 28m 17s ⏱️ - 2m 40s
2 726 tests +16  2 471 ✅ +16  255 💤 ±0  0 ❌ ±0 
2 728 runs  +16  2 471 ✅ +16  257 💤 ±0  0 ❌ ±0 

Results for commit 863cc94. ± Comparison against base commit 4be0da8.

♻️ This comment has been updated with latest results.

[thrau]

i'll let @giograno take the lead on this review!

[viren-nadkarni]

Love the data-driven approach, nice work @bentsku !