Add hybrid search WRRF guide

modules/genai-ecosystem/pages/vector-search.adoc

@@ -137,14 +137,116 @@ ORDER BY score DESC LIMIT 10;
 =====
 ====
 
+== Hybrid full-text and vector search
+
+Vector search is useful for finding text with similar meaning, but some searches also need exact terms, names, acronyms, identifiers, or recently added vocabulary.
+Hybrid search often combines semantic vector search with keyword search, then fuses both ranked result sets into one list.
+The same rank-fusion technique can combine any number of ranked or scored result sources, for example several vector indexes, or a vector index and a graph traversal score.
+
+The example below combines:
+
+* a full-text index on `Abstract.text`
+* the vector index on `Abstract.embedding`
+* weighted reciprocal rank fusion (WRRF), which scores each result from its rank in each source list rather than comparing raw scores from different search methods directly
+
+Create a full-text index for the same text that was embedded:
+
+[source,cypher]
+----
+CREATE FULLTEXT INDEX `abstract-fulltext` IF NOT EXISTS
+FOR (abstract:Abstract)
+ON EACH [abstract.text];
+----
+
+Use a larger `sourceK` than `finalK` so each source can contribute enough candidates before fusion.
+The `rrfConstant` dampens the effect of rank position, and `sourceWeights` lets you favor keyword or semantic matches for your application.
+The query uses `abstract.id` as a deterministic tie breaker; replace it with a stable unique property from your own data model.
+
+[source,cypher]
+----
+CYPHER 25
+LET
+  query = $query,
+  queryVector = $queryVector,
+  sourceK = $sourceK,
+  finalK = $finalK,
+  rrfConstant = $rrfConstant,
+  sourceWeights = $sourceWeights
+
+CALL (query, queryVector, sourceK, rrfConstant, sourceWeights) {
+  CALL db.index.fulltext.queryNodes('abstract-fulltext', query, {limit: sourceK})
+  YIELD node AS abstract, score
+  ORDER BY score DESC, abstract.id ASC
+  WITH collect(abstract) AS abstracts, rrfConstant, sourceWeights
+  LET weight = coalesce(sourceWeights['fulltext'], 1.0)
+  UNWIND CASE WHEN size(abstracts) = 0 THEN [] ELSE range(0, size(abstracts) - 1) END AS rankIndex
+  RETURN
+    abstracts[rankIndex] AS abstract,
+    weight / (rrfConstant + rankIndex + 1) AS contribution
+
+  UNION ALL
+
+  MATCH (abstract:Abstract)
+    SEARCH abstract IN (
+      VECTOR INDEX `abstract-embeddings`
+      FOR queryVector
+      LIMIT $sourceK
+    ) SCORE AS score
+  ORDER BY score DESC, abstract.id ASC
+  WITH collect(abstract) AS abstracts, rrfConstant, sourceWeights
+  LET weight = coalesce(sourceWeights['vector'], 1.0)
+  UNWIND CASE WHEN size(abstracts) = 0 THEN [] ELSE range(0, size(abstracts) - 1) END AS rankIndex
+  RETURN
+    abstracts[rankIndex] AS abstract,
+    weight / (rrfConstant + rankIndex + 1) AS contribution
+}
+WITH abstract, finalK, sum(contribution) AS wrrf
+ORDER BY wrrf DESC, abstract.id ASC
+WITH collect({abstract: abstract, wrrf: wrrf}) AS orderedRows, finalK
+LET limitedRows = orderedRows[..finalK]
+UNWIND limitedRows AS row
+WITH row.abstract AS abstract, row.wrrf AS wrrf
+MATCH (abstract)<--(:Paper)-->(title:Title)
+RETURN title.text AS title, abstract.text AS text, wrrf
+ORDER BY wrrf DESC, abstract.id ASC;
+----
+
+For example, pass parameters like these from an application:
+
+[source,json]
+----
+{
+  "query": "hierarchical navigable small world graph",
+  "queryVector": [0.12, -0.03, 0.45],
+  "sourceK": 20,
+  "finalK": 10,
+  "rrfConstant": 60.0,
+  "sourceWeights": {
+    "fulltext": 1.0,
+    "vector": 1.0
+  }
+}
+----
+
+Each source contributes `weight / (rrfConstant + sourceRank)` for every matching abstract.
+If an abstract appears in both result sets, the query sums both contributions.
+This rewards results that rank highly in either source, and gives an additional boost to results that rank well in both.
+To add more sources, add another `UNION ALL` branch that returns the same columns, `abstract` and `contribution`, using a different weight key such as `sourceWeights['metadata']` or `sourceWeights['graph']`.
+
+Increase the full-text weight when exact terminology is critical, such as product names, codes, or domain-specific vocabulary.
+Increase the vector weight when conceptual similarity should dominate.
+For retrieval-augmented generation (RAG), a common pattern is to retrieve a larger candidate set with hybrid search, then pass the fused results to graph expansion or a reranker.
+
 == Documentation
 [cols="1,4"]
 |===
 
 | icon:book[] Documentation | https://neo4j.com/docs/cypher-manual/current/indexes/semantic-indexes/vector-indexes/[Vector indexes^]
+| icon:book[] Documentation | https://neo4j.com/docs/cypher-manual/current/indexes/semantic-indexes/full-text-indexes/[Full-text indexes^]
 | icon:book[] Documentation | https://neo4j.com/docs/cypher-manual/current/clauses/search/[`SEARCH` clause^]
 | icon:book[] Documentation | https://neo4j.com/docs/genai/plugin/current/embeddings/[Create and store embeddings^]
 | icon:book[] Documentation | https://neo4j.com/docs/cypher-manual/current/functions/vector/[Vector Similarity Functions^]
+| icon:book[] Paper | https://plg.uwaterloo.ca/~gvcormac/cormacksigir09-rrf.pdf[Reciprocal Rank Fusion^]
 |===
 
 == Relevant Links