github.com/weaviate/weaviate@v1.24.6/adapters/repos/db/vector/hnsw/vector_cache_prefiller.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  	"time"
    17  
    18  	"github.com/sirupsen/logrus"
    19  	"github.com/weaviate/weaviate/adapters/repos/db/vector/cache"
    20  )
    21  
    22  type vectorCachePrefiller[T any] struct {
    23  	cache  cache.Cache[T]
    24  	index  *hnsw
    25  	logger logrus.FieldLogger
    26  }
    27  
    28  func newVectorCachePrefiller[T any](cache cache.Cache[T], index *hnsw,
    29  	logger logrus.FieldLogger,
    30  ) *vectorCachePrefiller[T] {
    31  	return &vectorCachePrefiller[T]{
    32  		cache:  cache,
    33  		index:  index,
    34  		logger: logger,
    35  	}
    36  }
    37  
    38  func (pf *vectorCachePrefiller[T]) Prefill(ctx context.Context, limit int) error {
    39  	before := time.Now()
    40  	for level := pf.maxLevel(); level >= 0; level-- {
    41  		ok, err := pf.prefillLevel(ctx, level, limit)
    42  		if err != nil {
    43  			return err
    44  		}
    45  
    46  		if !ok {
    47  			break
    48  		}
    49  	}
    50  
    51  	pf.logTotal(int(pf.cache.Len()), limit, before)
    52  	return nil
    53  }
    54  
    55  // returns false if the max has been reached, true otherwise
    56  func (pf *vectorCachePrefiller[T]) prefillLevel(ctx context.Context,
    57  	level, limit int,
    58  ) (bool, error) {
    59  	// TODO: this makes zero sense, just copy the lists, don't actually block
    60  	//  !!!!
    61  
    62  	before := time.Now()
    63  	layerCount := 0
    64  
    65  	pf.index.Lock()
    66  	nodesLen := len(pf.index.nodes)
    67  	pf.index.Unlock()
    68  
    69  	for i := 0; i < nodesLen; i++ {
    70  		if int(pf.cache.Len()) >= limit {
    71  			break
    72  		}
    73  
    74  		if err := ctx.Err(); err != nil {
    75  			return false, err
    76  		}
    77  
    78  		pf.index.shardedNodeLocks.RLock(uint64(i))
    79  		node := pf.index.nodes[i]
    80  		pf.index.shardedNodeLocks.RUnlock(uint64(i))
    81  
    82  		if node == nil {
    83  			continue
    84  		}
    85  
    86  		if levelOfNode(node) != level {
    87  			continue
    88  		}
    89  
    90  		// we are not really interested in the result, we just want to populate the
    91  		// cache
    92  		pf.index.Lock()
    93  		pf.cache.Get(ctx, uint64(i))
    94  		layerCount++
    95  		pf.index.Unlock()
    96  	}
    97  
    98  	pf.logLevel(level, layerCount, before)
    99  	return true, nil
   100  }
   101  
   102  func (pf *vectorCachePrefiller[T]) logLevel(level, count int, before time.Time) {
   103  	pf.logger.WithFields(logrus.Fields{
   104  		"action":     "hnsw_vector_cache_prefill_level",
   105  		"hnsw_level": level,
   106  		"count":      count,
   107  		"took":       time.Since(before),
   108  		"index_id":   pf.index.id,
   109  	}).Debug("prefilled level in vector cache")
   110  }
   111  
   112  func (pf *vectorCachePrefiller[T]) logTotal(count, limit int, before time.Time) {
   113  	pf.logger.WithFields(logrus.Fields{
   114  		"action":   "hnsw_vector_cache_prefill",
   115  		"limit":    limit,
   116  		"count":    count,
   117  		"took":     time.Since(before),
   118  		"index_id": pf.index.id,
   119  	}).Info("prefilled vector cache")
   120  }
   121  
   122  func levelOfNode(node *vertex) int {
   123  	node.Lock()
   124  	defer node.Unlock()
   125  
   126  	return node.level
   127  }
   128  
   129  func (pf *vectorCachePrefiller[T]) maxLevel() int {
   130  	pf.index.Lock()
   131  	defer pf.index.Unlock()
   132  
   133  	return pf.index.currentMaximumLayer
   134  }