Vector store support similarity search using vectors 支持使用向量进行相似度搜索

Support the direct use of vectors for similarity search instead of only conducting similarity search after embedding the query text.

使VectorStore支持直接使用向量进行相似度搜索，而非仅可使用文本嵌入进行搜索。

Expected Behavior

Like in Milvus' documentation: https://milvus.io/docs/single-vector-search.md

像Miluvs的文档中的：

FloatVec queryVector = new FloatVec(new float[]{0.3580376395471989f, -0.6023495712049978f, 0.18414012509913835f, -0.26286205330961354f, 0.9029438446296592f});

SearchReq searchReq = SearchReq.builder()
.data(Collections.singletonList(queryVector)) // use vectors
.topK(3)
.build();

Current Behavior

Only text search is available, but the text is then embedded and searched as a vector

只可使用文本搜索，而文本又经过嵌入后以向量进行搜索。

MilvusVectorStore.java#L354

Context

Start with vector-stores/spring-ai-milvus-store/src/main/java/org/springframework/ai/vectorstore/milvus/MilvusVectorStore.java at the behavior around line 354. Compare the current text-search flow with the Milvus vector-search example in the issue, then determine the API and implementation changes needed so callers can search with vectors directly while retaining text search.