mirror of
https://gitee.com/milvus-io/milvus.git
synced 2025-12-08 10:08:42 +08:00
d6a428e880
Ref https://github.com/milvus-io/milvus/issues/42148 This PR supports create index for vector array (now, only for `DataType.FLOAT_VECTOR`) and search on it. The index type supported in this PR is `EMB_LIST_HNSW` and the metric type is `MAX_SIM` only. The way to use it: ```python milvus_client = MilvusClient("xxx:19530") schema = milvus_client.create_schema(enable_dynamic_field=True, auto_id=True) ... struct_schema = milvus_client.create_struct_array_field_schema("struct_array_field") ... struct_schema.add_field("struct_float_vec", DataType.ARRAY_OF_VECTOR, element_type=DataType.FLOAT_VECTOR, dim=128, max_capacity=1000) ... schema.add_struct_array_field(struct_schema) index_params = milvus_client.prepare_index_params() index_params.add_index(field_name="struct_float_vec", index_type="EMB_LIST_HNSW", metric_type="MAX_SIM", index_params={"nlist": 128}) ... milvus_client.create_index(COLLECTION_NAME, schema=schema, index_params=index_params) ``` Note: This PR uses `Lims` to convey offsets of the vector array to knowhere where vectors of multiple vector arrays are concatenated and we need offsets to specify which vectors belong to which vector array. --------- Signed-off-by: SpadeA <tangchenjie1210@gmail.com> Signed-off-by: SpadeA-Tang <tangchenjie1210@gmail.com>
60 lines
1.8 KiB
Go
60 lines
1.8 KiB
Go
package indexparamcheck
|
|
|
|
import (
|
|
"fmt"
|
|
"strconv"
|
|
|
|
"github.com/milvus-io/milvus-proto/go-api/v2/schemapb"
|
|
"github.com/milvus-io/milvus/pkg/v2/util/merr"
|
|
"github.com/milvus-io/milvus/pkg/v2/util/typeutil"
|
|
)
|
|
|
|
const (
|
|
MinGramKey = "min_gram"
|
|
MaxGramKey = "max_gram"
|
|
)
|
|
|
|
type NgramIndexChecker struct {
|
|
scalarIndexChecker
|
|
}
|
|
|
|
func newNgramIndexChecker() *NgramIndexChecker {
|
|
return &NgramIndexChecker{}
|
|
}
|
|
|
|
func (c *NgramIndexChecker) CheckTrain(dataType schemapb.DataType, elementType schemapb.DataType, params map[string]string) error {
|
|
if dataType != schemapb.DataType_VarChar && dataType != schemapb.DataType_JSON {
|
|
return merr.WrapErrParameterInvalidMsg("Ngram index can only be created on VARCHAR or JSON field")
|
|
}
|
|
|
|
minGramStr, minGramExist := params[MinGramKey]
|
|
maxGramStr, maxGramExist := params[MaxGramKey]
|
|
if !minGramExist || !maxGramExist {
|
|
return merr.WrapErrParameterInvalidMsg("Ngram index must specify both min_gram and max_gram")
|
|
}
|
|
|
|
minGram, err := strconv.Atoi(minGramStr)
|
|
if err != nil {
|
|
return merr.WrapErrParameterInvalidMsg("min_gram for Ngram index must be an integer, got: %s", minGramStr)
|
|
}
|
|
|
|
maxGram, err := strconv.Atoi(maxGramStr)
|
|
if err != nil {
|
|
return merr.WrapErrParameterInvalidMsg("max_gram for Ngram index must be an integer, got: %s", maxGramStr)
|
|
}
|
|
|
|
if minGram <= 0 || maxGram <= 0 || minGram > maxGram {
|
|
return merr.WrapErrParameterInvalidMsg("invalid min_gram or max_gram value for Ngram index, min_gram: %d, max_gram: %d", minGram, maxGram)
|
|
}
|
|
|
|
return c.scalarIndexChecker.CheckTrain(dataType, elementType, params)
|
|
}
|
|
|
|
func (c *NgramIndexChecker) CheckValidDataType(indexType IndexType, field *schemapb.FieldSchema) error {
|
|
dType := field.GetDataType()
|
|
if !typeutil.IsStringType(dType) && dType != schemapb.DataType_JSON {
|
|
return fmt.Errorf("ngram index can only be created on VARCHAR or JSON field")
|
|
}
|
|
return nil
|
|
}
|