OpenSearch/docs/reference/analysis/token-graphs.asciidoc

104 lines
3.6 KiB
Plaintext

[[token-graphs]]
=== Token graphs
When a <<analyzer-anatomy-tokenizer,tokenizer>> converts a text into a stream of
tokens, it also records the following:
* The `position` of each token in the stream
* The `positionLength`, the number of positions that a token spans
Using these, you can create a
https://en.wikipedia.org/wiki/Directed_acyclic_graph[directed acyclic graph],
called a _token graph_, for a stream. In a token graph, each position represents
a node. Each token represents an edge or arc, pointing to the next position.
image::images/analysis/token-graph-qbf-ex.svg[align="center"]
[[token-graphs-synonyms]]
==== Synonyms
Some <<analyzer-anatomy-token-filters,token filters>> can add new tokens, like
synonyms, to an existing token stream. These synonyms often span the same
positions as existing tokens.
In the following graph, `quick` and its synonym `fast` both have a position of
`0`. They span the same positions.
image::images/analysis/token-graph-qbf-synonym-ex.svg[align="center"]
[[token-graphs-multi-position-tokens]]
==== Multi-position tokens
Some token filters can add tokens that span multiple positions. These can
include tokens for multi-word synonyms, such as using "atm" as a synonym for
"automatic teller machine."
However, only some token filters, known as _graph token filters_, accurately
record the `positionLength` for multi-position tokens. This filters include:
* <<analysis-synonym-graph-tokenfilter,`synonym_graph`>>
* <<analysis-word-delimiter-graph-tokenfilter,`word_delimiter_graph`>>
In the following graph, `domain name system` and its synonym, `dns`, both have a
position of `0`. However, `dns` has a `positionLength` of `3`. Other tokens in
the graph have a default `positionLength` of `1`.
image::images/analysis/token-graph-dns-synonym-ex.svg[align="center"]
[[token-graphs-token-graphs-search]]
===== Using token graphs for search
<<analysis-index-search-time,Indexing>> ignores the `positionLength` attribute
and does not support token graphs containing multi-position tokens.
However, queries, such as the <<query-dsl-match-query,`match`>> or
<<query-dsl-match-query-phrase,`match_phrase`>> query, can use these graphs to
generate multiple sub-queries from a single query string.
.*Example*
[%collapsible]
====
A user runs a search for the following phrase using the `match_phrase` query:
`domain name system is fragile`
During <<analysis-index-search-time,search analysis>>, `dns`, a synonym for
`domain name system`, is added to the query string's token stream. The `dns`
token has a `positionLength` of `3`.
image::images/analysis/token-graph-dns-synonym-ex.svg[align="center"]
The `match_phrase` query uses this graph to generate sub-queries for the
following phrases:
[source,text]
------
dns is fragile
domain name system is fragile
------
This means the query matches documents containing either `dns is fragile` _or_
`domain name system is fragile`.
====
[[token-graphs-invalid-token-graphs]]
===== Invalid token graphs
The following token filters can add tokens that span multiple positions but
only record a default `positionLength` of `1`:
* <<analysis-synonym-tokenfilter,`synonym`>>
* <<analysis-word-delimiter-tokenfilter,`word_delimiter`>>
This means these filters will produce invalid token graphs for streams
containing such tokens.
In the following graph, `dns` is a multi-position synonym for `domain name
system`. However, `dns` has the default `positionLength` value of `1`, resulting
in an invalid graph.
image::images/analysis/token-graph-dns-invalid-ex.svg[align="center"]
Avoid using invalid token graphs for search. Invalid graphs can cause unexpected
search results.