<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Rescoring on INFINI Pizza</title><link>https://pizza.rs/tags/rescoring/</link><description>Recent content in Rescoring on INFINI Pizza</description><generator>Hugo</generator><language>en</language><atom:link href="https://pizza.rs/tags/rescoring/index.xml" rel="self" type="application/rss+xml"/><item><title>Hybrid search and scoring pipeline</title><link>https://pizza.rs/docs/references/search/request/hybrid/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://pizza.rs/docs/references/search/request/hybrid/</guid><description>&lt;h1 id="hybrid-search-and-scoring-pipeline">
 Hybrid search and scoring pipeline
 &lt;a class="anchor" href="#hybrid-search-and-scoring-pipeline">#&lt;/a>
&lt;/h1>
&lt;p>Pizza runs lexical (BM25) and vector retrieval side by side and gives you
explicit, composable stages to combine and re-shape scores:&lt;/p>

&lt;p class="mermaid">
graph LR
 B("bm25 leg") --> F("fuse&lt;br/>rrf / linear / disjmax")
 V("vector leg") --> F
 F --> SC("score_functions&lt;br/>per-document function scoring")
 P["a plain query,&lt;br/>no retrieve"] --> SC
 SC --> RS("rescore&lt;br/>second pass on the top window")
 RS --> H["one ranked list"]
&lt;/p>

&lt;p>Every stage is optional and combines with a plain &lt;code>query&lt;/code>: use
&lt;code>score_functions&lt;/code> alone on a keyword search, or the full chain for
hybrid retrieval.&lt;/p></description></item></channel></rss>