Wire createX() factory methods to the mapper's converter layer instead of builder methods

[RanVaknin]

Context

#7039 code generated new factory methods that create AttributeValues directly and avoid allocations that the Builder pattern adds. in #7073 we wired these new factory methods into the Enhanced client. This PR wires the same thing just for the Mapper

Changes

This PR does 2 things:

1. swap builder() with factory methods - one file change in StandardModelFactories.java. This is the layer that creates the AV for the converters engine (the @deprecated Marshaller layer is untouched)
2. wiring in mapper v2 into the existing DDB JMH harness.

Testing

Since we knew about this optimization ahead of time, adequate testing was added early on to catch potential regressions.

Write direction (convert):

• StandardModelFactoriesTest (68 tests)
• StandardModelFactoriesV1Test (18), StandardModelFactoriesV2CompatibleTest (20), StandardModelFactoriesV2Test (21) — same assertions across all three conversion schemas
• StandardModelFactoriesEdgeCasesTest (14) — null/empty-string/empty-set branches (createNul)

Read direction (unconvert):

• StandardModelFactoriesV2UnconvertTest (21)
• shape/ShapeResponseTest — parameterized golden-master; reconstructs POJOs from canned attribute maps against unmarshall_item_fixture.json
• shape/ShapeResponseBehaviorTest (6)

Custom converter interaction:

• ConversionToAttributeValuesTest

Per-type round-trip integration (save + load):

• mapper/V2CompatibleBooleansTest (24) — BOOL and boolean-as-N
• mapper/SimpleStringAttributesIntegrationTest (6) — S
• mapper/SimpleNumericAttributesIntegrationTest (6) — N
• mapper/BinaryAttributesIntegrationTest (5) — B
• mapper/StringSetAttributesIntegrationTest (4) — SS
• mapper/NumericSetAttributesIntegrationTest (3) — NS
• mapper/EmptyBinarySetByteBufferAttributesTest (3) — BS + empty-set edge

Every createX factory wired in this PR (createS, createN, createB, createSs, createNs, createBs, createBool, createL, createM, createNul) has at least one test asserting its output above.

Performance

Operation	Payload	Baseline ops/s	Candidate ops/s	Throughput	Baseline B/op	Candidate B/op	Allocation
DELETE	HUGE	3,414,517	3,588,195	+5.09%	640	592	-7.50%
DELETE	HUGE_FLAT	3,351,349	3,767,951	+12.43%	672	592	-11.90%
DELETE	SMALL	3,399,646	3,616,885	+6.39%	656	592	-9.76%
DELETE	TINY	3,404,190	3,936,064	+15.62%	640	592	-7.50%
GET	HUGE	141,655	144,273	+1.85%	10,096	9,968	-1.27%
GET	HUGE_FLAT	173,222	173,752	+0.31%	2,720	2,624	-3.53%
GET	SMALL	1,068,911	1,232,606	+15.31%	1,016	888	-12.60%
GET	TINY	1,713,863	2,105,325	+22.84%	752	632	-15.96%
PUT	HUGE	66,105	92,423	+39.81%	29,392	17,868	-39.21%
PUT	HUGE_FLAT	79,497	93,424	+17.52%	19,592	17,536	-10.49%
PUT	SMALL	798,939	1,056,863	+32.28%	2,432	1,856	-23.68%
PUT	TINY	1,821,874	2,077,679	+14.04%	1,172	1,088	-7.17%
QUERY	HUGE	47,015	46,985	-0.06%	30,184	30,056	-0.42%
QUERY	HUGE_FLAT	56,857	56,800	-0.10%	7,528	7,432	-1.28%
QUERY	SMALL	409,822	410,429	+0.15%	2,952	2,912	-1.36%
QUERY	TINY	706,615	732,691	+3.69%	2,232	2,184	-2.15%
SCAN	HUGE	49,117	49,172	+0.11%	28,880	28,880	-0.00%
SCAN	HUGE_FLAT	63,920	57,731	-9.68%	6,224	6,224	+0.00%
SCAN	SMALL	559,194	532,875	-4.71%	1,736	1,736	+0.00%
SCAN	TINY	1,186,133	1,183,984	-0.18%	1,120	1,120	+0.00%
UPDATE	HUGE	59,941	85,869	+43.26%	30,304	18,784	-38.01%
UPDATE	HUGE_FLAT	47,935	58,259	+21.54%	26,664	20,616	-22.68%
UPDATE	SMALL	552,477	682,228	+23.49%	3,160	2,592	-17.97%
UPDATE	TINY	1,216,015	1,322,968	+8.80%	1,576	1,448	-8.12%

[github-actions]

This pull request has been closed and the conversation has been locked. Comments on closed PRs are hard for our team to see. If you need more assistance, please open a new issue that references this one.