AnalyzerFeature uses ConstantsTrait
ArangoSearch analyzer features, used as entries of the `features` array on `POST /_api/analyzer` (and as the per-field `features` array on inverted indexes / arangosearch views).
Features determine which metadata the indexer keeps alongside each
indexed token — and therefore which AQL SEARCH operators are
available against the analyzed field:
- self::FREQUENCY is required by
BM25()/TFIDF()scoring, - self::NORM is required by
BM25()length normalisation, - self::POSITION is required by
PHRASE(), - self::OFFSET is required by snippet highlighting and is only meaningful when self::POSITION is also enabled.
Tags
Table of Contents
Constants
- FREQUENCY : string = 'frequency'
- Keeps the token frequency per field. Required by `BM25()` and `TFIDF()` scoring functions.
- NORM : string = 'norm'
- Keeps the document length norm per field. Required by `BM25()` length normalisation.
- OFFSET : string = 'offset'
- Keeps the byte offset of each token within the source field.
- POSITION : string = 'position'
- Keeps the ordinal position of each token within the source field. Required by `PHRASE()`.
Constants
FREQUENCY
Keeps the token frequency per field. Required by `BM25()` and `TFIDF()` scoring functions.
public
string
FREQUENCY
= 'frequency'
NORM
Keeps the document length norm per field. Required by `BM25()` length normalisation.
public
string
NORM
= 'norm'
OFFSET
Keeps the byte offset of each token within the source field.
public
string
OFFSET
= 'offset'
Required by highlighting / snippet extraction. Implies self::POSITION.
POSITION
Keeps the ordinal position of each token within the source field. Required by `PHRASE()`.
public
string
POSITION
= 'position'