feat(schema): add SVS-VAMANA vector indexing algorithm support (#404)

[bsbodden]

Implements SVS-VAMANA algorithm with compression support for memory-efficient vector search, porting functionality from Python redis-vl PR #404.

Changes:

• Expand VectorDataType enum with FLOAT16, BFLOAT16, INT8, UINT8
• Add CompressionType enum (LVQ4, LVQ4x4, LVQ4x8, LVQ8, LeanVec4x8, LeanVec8x8)
• Add SVS_VAMANA to Algorithm enum
• Add 7 SVS-specific parameters to VectorField (graphMaxDegree, constructionWindowSize, searchWindowSize, svsEpsilon, compression, reduce, trainingThreshold)
• Implement SVS validation (datatype, reduce, compression constraints)
• Add builder methods for all SVS parameters
• Update toJedisSchemaField() to support SVS attributes

Tests:

• Add SVSVamanaFieldTest with 19 unit tests (all passing)
• Add SVSVamanaIntegrationTest with 7 integration tests (all passing)
• Add BaseSVSIntegrationTest using Redis 8.2 container
• Test all compression types, validation rules, and constraints
• Verify index creation, data loading, and search operations

Requirements:

• Redis ≥ 8.2.0 (available as redis:8.2 Docker image)
• RediSearch ≥ 2.8.10 or SearchLight ≥ 2.8.10

Python reference: PR #404 - SVS-VAMANA support
Ported from: redisvl/schema/fields.py