github.com/weaviate/weaviate@v1.24.6/adapters/repos/db/vector/hnsw/heuristic.go (about)

     1  //                           _       _
     2  // __      _____  __ ___   ___  __ _| |_ ___
     3  // \ \ /\ / / _ \/ _` \ \ / / |/ _` | __/ _ \
     4  //  \ V  V /  __/ (_| |\ V /| | (_| | ||  __/
     5  //   \_/\_/ \___|\__,_| \_/ |_|\__,_|\__\___|
     6  //
     7  //  Copyright © 2016 - 2024 Weaviate B.V. All rights reserved.
     8  //
     9  //  CONTACT: hello@weaviate.io
    10  //
    11  
    12  package hnsw
    13  
    14  import (
    15  	"context"
    16  
    17  	"github.com/pkg/errors"
    18  	"github.com/weaviate/weaviate/adapters/repos/db/helpers"
    19  	"github.com/weaviate/weaviate/adapters/repos/db/priorityqueue"
    20  	"github.com/weaviate/weaviate/entities/storobj"
    21  )
    22  
    23  func (h *hnsw) selectNeighborsHeuristic(input *priorityqueue.Queue[any],
    24  	max int, denyList helpers.AllowList,
    25  ) error {
    26  	if input.Len() < max {
    27  		return nil
    28  	}
    29  
    30  	// TODO, if this solution stays we might need something with fewer allocs
    31  	ids := make([]uint64, input.Len())
    32  
    33  	closestFirst := h.pools.pqHeuristic.GetMin(input.Len())
    34  	i := uint64(0)
    35  	for input.Len() > 0 {
    36  		elem := input.Pop()
    37  		closestFirst.InsertWithValue(elem.ID, elem.Dist, i)
    38  		ids[i] = elem.ID
    39  		i++
    40  	}
    41  
    42  	var returnList []priorityqueue.Item[uint64]
    43  
    44  	if h.compressed.Load() {
    45  		bag := h.compressor.NewBag()
    46  		for _, id := range ids {
    47  			err := bag.Load(context.Background(), id)
    48  			if err != nil {
    49  				return err
    50  			}
    51  		}
    52  
    53  		returnList = h.pools.pqItemSlice.Get().([]priorityqueue.Item[uint64])
    54  		for closestFirst.Len() > 0 && len(returnList) < max {
    55  			curr := closestFirst.Pop()
    56  			if denyList != nil && denyList.Contains(curr.ID) {
    57  				continue
    58  			}
    59  			distToQuery := curr.Dist
    60  
    61  			good := true
    62  			for _, item := range returnList {
    63  				peerDist, err := bag.Distance(curr.ID, item.ID)
    64  				if err != nil {
    65  					return err
    66  				}
    67  
    68  				if peerDist < distToQuery {
    69  					good = false
    70  					break
    71  				}
    72  			}
    73  
    74  			if good {
    75  				returnList = append(returnList, curr)
    76  			}
    77  
    78  		}
    79  	} else {
    80  
    81  		vecs, errs := h.multiVectorForID(context.TODO(), ids)
    82  
    83  		returnList = h.pools.pqItemSlice.Get().([]priorityqueue.Item[uint64])
    84  
    85  		for closestFirst.Len() > 0 && len(returnList) < max {
    86  			curr := closestFirst.Pop()
    87  			if denyList != nil && denyList.Contains(curr.ID) {
    88  				continue
    89  			}
    90  			distToQuery := curr.Dist
    91  
    92  			currVec := vecs[curr.Value]
    93  			if err := errs[curr.Value]; err != nil {
    94  				var e storobj.ErrNotFound
    95  				if errors.As(err, &e) {
    96  					h.handleDeletedNode(e.DocID)
    97  					continue
    98  				} else {
    99  					// not a typed error, we can recover from, return with err
   100  					return errors.Wrapf(err,
   101  						"unrecoverable error for docID %d", curr.ID)
   102  				}
   103  			}
   104  			good := true
   105  			for _, item := range returnList {
   106  				peerDist, _, _ := h.distancerProvider.SingleDist(currVec,
   107  					vecs[item.Value])
   108  
   109  				if peerDist < distToQuery {
   110  					good = false
   111  					break
   112  				}
   113  			}
   114  
   115  			if good {
   116  				returnList = append(returnList, curr)
   117  			}
   118  
   119  		}
   120  	}
   121  
   122  	h.pools.pqHeuristic.Put(closestFirst)
   123  
   124  	for _, retElem := range returnList {
   125  		input.Insert(retElem.ID, retElem.Dist)
   126  	}
   127  
   128  	// rewind and return to pool
   129  	returnList = returnList[:0]
   130  
   131  	//nolint:staticcheck
   132  	h.pools.pqItemSlice.Put(returnList)
   133  
   134  	return nil
   135  }