fix: update check for sparse hnsw index (#33713)

issue: #29419

Signed-off-by: Buqian Zheng <zhengbuqian@gmail.com>

internal/proxy/task_index.go

@@ -51,6 +51,7 @@ const (
 
 	AutoIndexName = "AUTOINDEX"
 	DimKey        = common.DimKey
+	IsSparseKey   = common.IsSparseKey
 )
 
 type createIndexTask struct {
@@ -383,10 +384,15 @@ func checkTrain(field *schemapb.FieldSchema, indexParams map[string]string) erro
 		}
 	}
 
-	if !typeutil.IsSparseFloatVectorType(field.DataType) {
+	isSparse := typeutil.IsSparseFloatVectorType(field.DataType)
+
+	if !isSparse {
 		if err := fillDimension(field, indexParams); err != nil {
 			return err
 		}
+	} else {
+		// used only for checker, should be deleted after checking
+		indexParams[IsSparseKey] = "true"
 	}
 
 	if err := checker.CheckValidDataType(field.GetDataType()); err != nil {
@@ -399,6 +405,10 @@ func checkTrain(field *schemapb.FieldSchema, indexParams map[string]string) erro
 		return err
 	}
 
+	if isSparse {
+		delete(indexParams, IsSparseKey)
+	}
+
 	return nil
 }
 

pkg/common/common.go

@@ -120,6 +120,7 @@ const (
 	DropRatioBuildKey = "drop_ratio_build"
 
 	BitmapCardinalityLimitKey = "bitmap_cardinality_limit"
+	IsSparseKey               = "is_sparse"
 )
 
 //  Collection properties key

pkg/util/indexparamcheck/base_checker.go

@@ -19,18 +19,37 @@ package indexparamcheck
 import (
 	"fmt"
 	"math"
+	"strings"
 
 	"github.com/cockroachdb/errors"
 
 	"github.com/milvus-io/milvus-proto/go-api/v2/schemapb"
+	"github.com/milvus-io/milvus/pkg/common"
 )
 
 type baseChecker struct{}
 
 func (c baseChecker) CheckTrain(params map[string]string) error {
 	// vector dimension should be checked on collection creation. this is just some basic check
-	if !CheckIntByRange(params, DIM, 1, math.MaxInt) {
-		return fmt.Errorf("failed to check vector dimension, should be larger than 0 and smaller than math.MaxInt")
+	isSparse := false
+	if val, exist := params[common.IsSparseKey]; exist {
+		val = strings.ToLower(val)
+		if val != "true" && val != "false" {
+			return fmt.Errorf("invalid is_sparse value: %s, must be true or false", val)
+		}
+		if val == "true" {
+			isSparse = true
+		}
+	}
+	if isSparse {
+		if !CheckStrByValues(params, Metric, SparseMetrics) {
+			return fmt.Errorf("metric type not found or not supported for sparse float vectors, supported: %v", SparseMetrics)
+		}
+	} else {
+		// we do not check dim for sparse
+		if !CheckIntByRange(params, DIM, 1, math.MaxInt) {
+			return fmt.Errorf("failed to check vector dimension, should be larger than 0 and smaller than math.MaxInt")
+		}
 	}
 	return nil
 }

pkg/util/indexparamcheck/base_checker_test.go

@@ -7,6 +7,7 @@ import (
 	"github.com/stretchr/testify/assert"
 
 	"github.com/milvus-io/milvus-proto/go-api/v2/schemapb"
+	"github.com/milvus-io/milvus/pkg/common"
 	"github.com/milvus-io/milvus/pkg/util/metric"
 )
 
@@ -18,12 +19,27 @@ func Test_baseChecker_CheckTrain(t *testing.T) {
 	paramsWithoutDim := map[string]string{
 		Metric: metric.L2,
 	}
+	sparseParamsWithoutDim := map[string]string{
+		Metric:             metric.IP,
+		common.IsSparseKey: "tRue",
+	}
+	sparseParamsWrongMetric := map[string]string{
+		Metric:             metric.L2,
+		common.IsSparseKey: "True",
+	}
+	badSparseParams := map[string]string{
+		Metric:             metric.IP,
+		common.IsSparseKey: "ds",
+	}
 	cases := []struct {
 		params   map[string]string
 		errIsNil bool
 	}{
 		{validParams, true},
 		{paramsWithoutDim, false},
+		{sparseParamsWithoutDim, true},
+		{sparseParamsWrongMetric, false},
+		{badSparseParams, false},
 	}
 
 	c := newBaseChecker()