fix: support non-float vectors in struct array in search

[SpadeA-Tang]

issue: #3269

[sre-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: SpadeA-Tang
To complete the pull request process, please assign tedxu after the PR has been reviewed.
You can assign the PR to them by writing /assign @tedxu in a comment when ready.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.57%. Comparing base (eb7868f) to head (112242a).
⚠️ Report is 2 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3277      +/-   ##
==========================================
+ Coverage   76.36%   76.57%   +0.21%     
==========================================
  Files          63       63              
  Lines       13321    13330       +9     
==========================================
+ Hits        10173    10208      +35     
+ Misses       3148     3122      -26     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[zhuwenxing]

Verified against Milvus master-20260210-f354d39a. EmbeddingList + MAX_SIM search path works correctly for float16/int8/binary vectors. ✅

However, the single-vector element_filter search path still has issues for float16/bfloat16/int8. Raw bytes input in _prepare_placeholder_str is always tagged as BinaryVector (line 1348), causing type mismatch on the server side.

pymilvus.exceptions.MilvusException: <MilvusException: (code=65535, message=fail to search on QueryNode 1: 
worker(1) query failed: parser searchRequest failed: Assert "check_data_type(field_meta, ph)" => 
vector type must be the same, field items[embedding] - type VECTOR_ARRAY, search ph type VECTOR_BINARY 
at /workspace/source/internal/core/src/query/Plan.cpp:125)>

Reproduction script

import numpy as np
from pymilvus import MilvusClient, DataType

URI = "http://localhost:19530"
DIM = 32
NB = 100
client = MilvusClient(uri=URI)

col = "repro_ef_float16"
if col in client.list_collections():
    client.drop_collection(col)

schema = client.create_schema(auto_id=False)
schema.add_field("id", DataType.INT64, is_primary=True)
schema.add_field("top_vec", DataType.FLOAT_VECTOR, dim=DIM)

struct_schema = client.create_struct_field_schema()
struct_schema.add_field("embedding", DataType.FLOAT16_VECTOR, dim=DIM)
struct_schema.add_field("val", DataType.INT64)
schema.add_field("items", DataType.ARRAY, element_type=DataType.STRUCT,
                 struct_schema=struct_schema, max_capacity=5)

index_params = client.prepare_index_params()
index_params.add_index("top_vec", index_type="HNSW", metric_type="COSINE",
                       params={"M": 16, "efConstruction": 200})
index_params.add_index("items[embedding]", index_type="HNSW",
                       metric_type="MAX_SIM_L2",
                       params={"M": 16, "efConstruction": 200})
client.create_collection(collection_name=col, schema=schema, index_params=index_params)

data = []
for i in range(NB):
    elems = [{"embedding": np.random.rand(DIM).astype(np.float16).tobytes(),
              "val": i * 10 + j} for j in range(3)]
    data.append({"id": i, "top_vec": np.random.rand(DIM).astype(np.float32).tolist(),
                 "items": elems})
client.insert(col, data)
client.flush(col)

query_vec = data[0]["items"][0]["embedding"]  # raw bytes
client.search(
    collection_name=col, data=[query_vec],
    anns_field="items[embedding]",
    search_params={"metric_type": "L2"},
    filter="element_filter(items, $[val] >= 0)", limit=5,
)