Skip to content

Commit 918de5f

Browse files
authored
Merge pull request #4656 from vespa-engine/bratseth/vector-types
Minor improvements
2 parents 715ce5b + 8ea51ec commit 918de5f

1 file changed

Lines changed: 14 additions & 17 deletions

File tree

en/reference/ranking/rank-features.html

Lines changed: 14 additions & 17 deletions
Original file line numberDiff line numberDiff line change
@@ -18,17 +18,20 @@
1818
</p>
1919
<ul>
2020
<li>
21-
Types: All rank feature values are floats. Integers are converted to exact whole value floats.
22-
String values are converted to exact whole value floats using a hash function.
21+
Types: All rank feature values are doubles or tensors. Integers are converted to exact whole value doubles.
22+
String values are converted to exact whole value doubles using a hash function.
2323
String literals in ranking expressions are converted using the same hash function,
2424
to enable equality tests on string values.
25-
</li><li>
25+
</li>
26+
<li>
2627
Features which are <em>normalized</em> are between 0 and 1, where 0 is always the minimum and 1 the maximum.
2728
Normalized features should normally be preferred because they are more easily combined by
2829
<a href="ranking-expressions.html">ranking expressions</a> into a complete normalized score.
29-
</li><li>
30+
</li>
31+
<li>
3032
A query may override <em>any</em> rank feature value by submitting that value as a feature with the query.
31-
</li><li>
33+
</li>
34+
<li>
3235
Some features have parameters. It is always allowed to quote parameters with <em>"</em>.
3336
Nested quotes are not allowed and must be escaped using <em>\</em>.
3437
Parameters that can be parsed as feature names may be left unquoted.
@@ -42,11 +45,7 @@ <h2 id="feature-list">Feature list</h2>
4245
<!--table-to-list-->
4346
<table class="table" id="rank-feature-table">
4447

45-
46-
47-
48-
49-
<tr><td colspan="3"></td></tr>
48+
<tr><td colspan="3"></td></tr>
5049
<tr><td colspan="3"><h3 id="query-features">Query features</h3></td></tr>
5150
<tr class="trx"><th class="thx">Feature name</th><th class="thx">Default</th><th class="thx">Description</th></tr>
5251

@@ -381,7 +380,7 @@ <h2 id="feature-list">Feature list</h2>
381380
<td>0</td>
382381
<td><p id="fieldMatch(name).absoluteOccurrence">
383382
Returns a normalized measure of the number of occurrence of the terms of the query:
384-
</p><p class="equation-container" class="equation-container"><!-- depends on mathjax -->
383+
</p><p class="equation-container"><!-- depends on mathjax -->
385384
$$\frac{\sum_{\text{all query terms}}(min(\text{number of occurrences of the term},maxOccurrences))}{(\text{query term count} &times; 100)}$$
386385
</p><p>
387386
This is 1 if there are many occurrences of the query terms, and 0 if there are none.
@@ -838,10 +837,8 @@ <h3 id="attribute-match-features-normalized">Attribute match features - normaliz
838837
<td>0</td>
839838
<td><p id="closeness(dimension,name)">
840839
Used with the <a href="../querying/yql.html#nearestneighbor">nearestNeighbor</a> query operator.
841-
A number which is close to 1 when a point vector in the document is close
842-
to a matching point vector in the query.
843-
The document vectors and the query vector must be the same tensor
844-
type, with one indexed dimension of size N, representing a point in an N-dimensional space.
840+
A number which is close to 1 when a vector in the document tensor is close to the vector given in the query.
841+
The indexed dimension representing a vector in document and query tensors must be identical.
845842
</p>
846843
<ul>
847844
<li>
@@ -899,12 +896,12 @@ <h3 id="attribute-match-features-normalized">Attribute match features - normaliz
899896
<td><p id="freshness(name).logscale">
900897
A logarithmic-shaped freshness; also goes from 1 to 0, but looks like <a href="#freshness">freshness plot</a>.
901898
The function is based on <code>-log(age(name) + scale)</code> and is calculated as:
902-
</p><p class="equation-container" class="equation-container"><!-- depends on mathjax -->
899+
</p><p class="equation-container"><!-- depends on mathjax -->
903900
$$\frac{log(maxAge + scale) - log(age(name) + scale)}{log(maxAge + scale) - log(scale)}$$
904901
</p><p>
905902
where scale is defined using
906903
<a href="rank-feature-configuration.html#freshness"> halfResponse and maxAge</a>:
907-
</p><p class="equation-container" class="equation-container"><!-- depends on mathjax -->
904+
</p><p class="equation-container"><!-- depends on mathjax -->
908905
$$\frac{-halfResponse^2}{2 &times; halfResponse - maxAge}$$
909906
</p><p>
910907
When <code>age(name) == halfResponse</code> the function output is 0.5.</p>

0 commit comments

Comments
 (0)