feat(query): add multi-vector query support (#402)

[bsbodden]

Add support for searching across multiple vector fields simultaneously with weighted score combination, following Python RedisVL PR #402.

New Classes:

• Vector: Data container for vector information with validation

  • Stores vector, field name, dtype, and weight
  • Validates dtype against supported Redis types
  • Ensures positive weights and non-empty vectors

• MultiVectorQuery: Enables multi-vector search with weighted scoring

  • Accepts single or multiple Vector objects
  • Builds VECTOR_RANGE queries with distance threshold 2.0
  • Combines scores using formula: w_1score_1 + w_2score_2 + ...
  • Supports filter expressions and return fields
  • Returns params map with vector_0, vector_1, etc.

Key Features:

• Simultaneous search across multiple vector fields
• Weighted score combination
• Individual score calculations: (2 - distance_i)/2
• Flexible builder API (single, varargs, list)
• Comprehensive validation
• 20 unit tests covering all scenarios

Implementation Details:

• Distance threshold hardcoded at 2.0 (includes all eligible docs)
• Vectors may differ in size and datatype
• Must use cosine distance metric
• Immutable Vector class with defensive copying
• Query string format: @field:[VECTOR_RANGE 2.0 $vec]=>{$YIELD_DISTANCE_AS: dist}

Python Reference: redis/redis-vl-python#402