Nrtsearch

A high performance gRPC server, with optional REST APIs on top of Apache Lucene version 8.x source, exposing Lucene’s core functionality over a simple gRPC based API.

• Introduction
  • Features
  • Design
  • Indexing a stream of documents
  • Near-real-time-replication
  • Build Server and Client
  • Run gRPC Server
  • Run REST Server
• Get Started
  • Docker Compose Setup
  • Launch you local NRTSearch server
  • Create your NRTSearch index
    • Create Index
    • Configure the index
    • Register Fields
    • Start Index
  • Add documents to your NRTSearch index
  • Query your NRTSearch server
  • Java client Example
• Mapping
  • Field Types
    • Boolean
    • Date Time
    • ID Field
    • Keyword
    • Location
    • Numeric
    • Object
    • Polygon
    • Text
    • Vector
    • Virtual
  • Child Fields
• Building Nrtsearch Queries
  • Boolean Query
  • Constant Score Query
  • Cross Index Query
    • Memory considerations
    • Multi-valued join fields
  • Disjunction Max Query
  • Exists Query
  • Function Filter Query
  • Function Score Query
  • Geo-Bounding Box Query
  • Geo-Point Query
  • Geo-Polygon Query
  • Geo-Radius Query
  • Match Query
  • Match All Query
  • Match Phrase Query
  • Match Phrase Prefix Query
  • Min Score Query
  • Multi Function Score Query
  • Multi-Match Query
  • Nested Query
  • Phrase Query
  • Prefix Query
  • Range Query
  • Span Multi Term Query
  • Span Near Query
  • Term Query
  • Term-In-Set Query
• Analyzers
  • Predefined Analyzers
  • Building Custom Analyzers
  • Normalizers
    • Predefined Normalizers
    • Custom Normalizers
• Highlighting
  • Requirements
    • General:
    • Fast-vector-highlighter:
  • Query Syntax
  • Example Queries
• InnerHit
  • Requirements
  • Query Syntax
  • Example Queries
    • Case 1
    • Case 2
    • Case 3
    • Case 4
    • Case 5
• Additional Collectors
  • Terms Collector
    • Bucket Order
  • Plugin Collector
  • Top Hits Collector
  • Filter Collector
  • Max Collector
  • Min Collector
  • Sum Collector
• Index Settings
  • Applying
  • Properties
    • directory
    • indexSort
    • concurrentMergeSchedulerMaxThreadCount
    • concurrentMergeSchedulerMaxMergeCount
    • indexMergeSchedulerAutoThrottle
    • nrtCachingDirectoryMaxSizeMB
    • nrtCachingDirectoryMaxMergeSizeMB
    • maxFullFlushMergeWaitMillis
• Index Live Settings
  • Applying
  • Properties
    • minRefreshSec
    • maxRefreshSec
    • maxSearcherAgeSec
    • indexRamBufferSizeMB
    • addDocumentsMaxBufferLen
    • sliceMaxDocs
    • sliceMaxSegments
    • virtualShards
    • maxMergedSegmentMB
    • segmentsPerTier
    • defaultSearchTimeoutSec
    • defaultSearchTimeoutCheckEvery
    • defaultTerminateAfter
    • maxMergePreCopyDurationSec
    • parallelFetchByField
    • parallelFetchChunkSize
• Docker Compose
  • Introduction
    • 0. PreSteps
    • 1. Starting Containers
    • 2. Primary: Start Index
    • 3. Replica: Start Index
    • 4. Replication
    • 5. Replica: Search Should Work
    • 6. Backup Index to S3
    • 7. Restart nrtSearch with Backup Index on S3
• Server Configuration
• Reserved Values
• Development
  • Adding A New Live Setting
  • Writing Documentation
• Vector Search and Embeddings in Nrtsearch
  • What is Vector/Embeddings?
  • Vector Search Applications
  • Vector Embeddings on Nrtsearch
  • Launching Cluster with Vector Search Support
    • Estimate Cluster Size
    • Configuring Cluster
    • Ingestion
  • Search
    • Vector-only search
    • Vector Search + Inline Filter
    • Vector Search + Top Level Filter
  • Optimizing Search Queries
• Multi-Retriever Search
  • Overview
  • Request Structure
  • Retrievers
    • Text Retriever
    • KNN Retriever
  • Blenders
    • Weighted RRF (Reciprocal Rank Fusion)
    • Weighted Score Order
    • Scoreless Raw Merge
    • Plugin Blender
  • Diagnostics
  • Full Example
  • See Also
• Script Rescorer
  • Overview
  • Request Structure
  • Script Variables
    • _score
    • Document Fields
    • Script Parameters
  • Built-in Language: js (Lucene Expressions)
  • Plugin Languages
  • Examples
    • Standard Query with Rescorer
    • Multi-Retriever with Rescorer
  • Chaining Rescorers
  • Per-Retriever Rescoring (L1)
  • See Also