svn commit: r1083614 [3/3] - in /sites/solr/guide: search-index.js sitemap.xml solr/9_3/query-guide/function-queries.html

[sv...@apache.org]

Modified: sites/solr/guide/solr/9_3/query-guide/function-queries.html
URL: http://svn.apache.org/viewvc/sites/solr/guide/solr/9_3/query-guide/function-queries.html?rev=1083614&r1=1083613&r2=1083614&view=diff
==============================================================================
--- sites/solr/guide/solr/9_3/query-guide/function-queries.html (original)
+++ sites/solr/guide/solr/9_3/query-guide/function-queries.html Mon Jul 10 15:22:03 2023
@@ -1339,6 +1339,42 @@ Each ValueSource must be a number.</p>
 </div>
 </div>
 <div class="sect2">
+<h3 id="vectorsimilarity-function"><a class="anchor" href="#vectorsimilarity-function"></a>vectorSimilarity Function</h3>
+<div class="paragraph">
+<p>Returns the similarity between two Knn vectors in an n-dimensional space.
+Takes in input the vector element encoding, the similarity measure plus two ValueSource instances and calculates the similarity between the two vectors.</p>
+</div>
+<div class="ulist">
+<ul>
+<li>
+<p>The encodings supported are: <code>BYTE</code>, <code>FLOAT32</code>.</p>
+</li>
+<li>
+<p>The similarities supported are: <code>EUCLIDEAN</code>, <code>COSINE</code>, <code>DOT_PRODUCT</code></p>
+</li>
+</ul>
+</div>
+<div class="paragraph">
+<p>Each ValueSource must be a knn vector (field or constant).</p>
+</div>
+<div class="paragraph">
+<p><strong>Syntax Examples</strong></p>
+</div>
+<div class="ulist">
+<ul>
+<li>
+<p><code>vectorSimilarity(FLOAT32, COSINE, [1,2,3], [4,5,6])</code>: calculates the cosine similarity between [1, 2, 3] and [4, 5, 6] for each document.</p>
+</li>
+<li>
+<p><code>vectorSimilarity(FLOAT32, DOT_PRODUCT, vectorField1, vectorField2)</code>: calculates the dot product similarity between the vector in 'vectorField1' and in 'vectorField2' for each document.</p>
+</li>
+<li>
+<p><code>vectorSimilarity(BYTE, EUCLIDEAN, [1,5,4,3], vectorField)</code>: calculates the euclidean similarity between the vector in 'vectorField' and the constant vector [1, 5, 4, 3] for each document.</p>
+</li>
+</ul>
+</div>
+</div>
+<div class="sect2">
 <h3 id="docfreqfieldval-function"><a class="anchor" href="#docfreqfieldval-function"></a>docfreq(field,val) Function</h3>
 <div class="paragraph">
 <p>Returns the number of documents that contain the term in the field.