MB-72376: Introducing New Geo-Shape Indexes

go.mod

@@ -4,7 +4,7 @@ go 1.24.0
 
 require (
 	github.com/RoaringBitmap/roaring/v2 v2.14.5
-	github.com/blevesearch/bleve_index_api v1.4.0
+	github.com/blevesearch/bleve_index_api v1.4.1-0.20260729060817-8e56340f2a7e
 )
 
 require (

go.sum

@@ -2,8 +2,8 @@ github.com/RoaringBitmap/roaring/v2 v2.14.5 h1:ckd0o545JqDPeVJDgeFoaM21eBixUnlWf
 github.com/RoaringBitmap/roaring/v2 v2.14.5/go.mod h1:eq4wdNXxtJIS/oikeCzdX1rBzek7ANzbth041hrU8Q4=
 github.com/bits-and-blooms/bitset v1.24.2 h1:M7/NzVbsytmtfHbumG+K2bremQPMJuqv1JD3vOaFxp0=
 github.com/bits-and-blooms/bitset v1.24.2/go.mod h1:7hO7Gc7Pp1vODcmWvKMRA9BNmbv6a/7QIWpPxHddWR8=
-github.com/blevesearch/bleve_index_api v1.4.0 h1:xoCC4dvTizjcsZu7yO9Ua+/259K09BrirjTwLkx2MpY=
-github.com/blevesearch/bleve_index_api v1.4.0/go.mod h1:xvd48t5XMeeioWQ5/jZvgLrV98flT2rdvEJ3l/ki4Ko=
+github.com/blevesearch/bleve_index_api v1.4.1-0.20260729060817-8e56340f2a7e h1:WJ3JnFiBuZ0m1mXaniDHpbZxKU8hjp7SIhqHopSnaM4=
+github.com/blevesearch/bleve_index_api v1.4.1-0.20260729060817-8e56340f2a7e/go.mod h1:xvd48t5XMeeioWQ5/jZvgLrV98flT2rdvEJ3l/ki4Ko=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/mschoch/smat v0.2.0 h1:8imxQsjDm8yFEAVBe7azKmKSgzSkZXDuKkSq9374khM=

segment.go

@@ -273,3 +273,50 @@ type NestedSegment interface {
 	// a parent document is deleted, all its nested child documents are also considered deleted.
 	AddNestedDocuments(deleted *roaring.Bitmap) *roaring.Bitmap
 }
+
+// GeoShapeV2Segment is an optional interface that a Segment may implement
+// to provide access to GeoShapeV2Data within the segment.
+type GeoShapeV2Segment interface {
+	Segment
+
+	// GeoShapeV2Data returns the geo shape data for the given field,
+	// excluding any documents present in the except bitmap.
+	GeoShapeV2Data(field string, except *roaring.Bitmap) (GeoShapeV2Data, error)
+}
+
+// GeoShapeV2Data provides methods to access separate parts of the GeoShapeV2 data.
+// Internally, geo docIDs are sequential from 0 to NumDocs()-1; DocNums() maps each
+// geo docID (the slice index) to its segment document number.
+type GeoShapeV2Data interface {
+	// InnerCells returns all of the shapes' inner cells in ascending order.
+	InnerCells() []uint64
+	// InnerDocIDs returns the geo docIDs parallel to InnerCells().
+	InnerDocIDs() []uint32
+	// CrossCells returns all of the shapes' cross cells in ascending order.
+	CrossCells() []uint64
+	// CrossDocIDs returns the geo docIDs parallel to CrossCells().
+	CrossDocIDs() []uint32
+	// NumDocs returns the number of documents indexed.
+	NumDocs() uint64
+	// DocNums returns the mapping from geo docID (the slice index) to
+	// segment document number.
+	DocNums() []uint32
+	// DocScores returns the precomputed inner and cross cell scores (in
+	// that order) for the documents indexed, each indexed by geo docID.
+	DocScores() (innerScores, crossScores []uint64)
+	// BoundingBox returns the bounding box bytes for the given geo docID.
+	BoundingBox(geoDocID uint32) ([]byte, error)
+	// Shape returns the shape bytes for the given geo docID.
+	Shape(geoDocID uint32) ([]byte, error)
+	// Excluded returns the bitmap of geo document IDs that are excluded
+	// from the index.
+	Excluded() *roaring.Bitmap
+	// GetScoreMap returns an empty score map, keyed by geo docID, from a
+	// segment-level pool.
+	GetScoreMap() map[uint32]uint64
+	// PutScoreMap clears the score map obtained via GetScoreMap and
+	// returns it to the segment-level pool.
+	PutScoreMap(scores map[uint32]uint64)
+	// Close closes the GeoShapeV2Data and releases any associated resources.
+	Close()
+}