Analytics and Graph Traversal with Solr
summary
Solr now exposes a graph filter (breadth-first, non-distributed) that returns an uncored set of documents, caches results in the filter cache, and detects cycles; and a fully distributed graph streaming expression (gatherNodes, cross-shard, cross-collection) that follows edges across shards, outputs traversal metadata or GraphML, and supports shortest paths and scoreNodes (TF‑IDF style). The JSON Facet API adds domain-joins and composite ID co‑location for shard-local joins, HLL cardinalities and nested aggregations; a small recommender example yields book10 (Gridlink) as the expected result.
AI-written summary · about
chapters (26)
- 0:00Intro
- 0:29Yoniks background
- 0:54What is Solr
- 2:02What is faceted search
- 3:26faceted search versions
- 4:59Graph databases
- 5:52Properties
- 7:00Implicit relationships
- 8:34Models
- 9:37Graph Filter
- 13:02Futurama Example
- 15:07Streaming Expressions
- 16:55Stream Source
- 19:57Graph Streaming Expression
- 26:39Complete Graph Traversal
- 27:05Graph Filters
- 29:07Solr Admin
- 29:26Score nodes
- 31:15GraphML
- 31:50Graph Operators
- 32:47Composite ID Router
- 35:30Example JSON Aspect API
- 37:06Domain Changes and Join
- 39:22Additional Data
- 41:54Streaming Expressions vs JSON Aspect
- 45:58Questions
related talks
48mHow to Rewrite Faceted Search from Solr to Elastic
39mParallel SQL & Streaming Expressions in Apache Solr 6
45mHybrid Search with Apache Solr Reciprocal Rank Fusion
35mBuilding analytics application with streaming expressions
44mSearching through large graphs using Elasticsearch
01h 00mWho’s Afraid of Graphs?
21mJoining in Lucene
01h 00mGraph Mastery to AI