Built-in Suggestion Analyzer plugins
| Name | Description |
|---|---|
|
DocumentContentContainsSuggestion |
Suggests an option's value when its label appears anywhere as a substring of the document's extracted text (case-insensitive), triggering text extraction first if it hasn't run yet. Stops at the first matching option. |
|
DocumentContentTokenizerSuggestion |
Same text extraction as DocumentContentContainsSuggestion, but splits the extracted text into whitespace-separated tokens first and requires an option's label to exactly match a whole token (case-insensitive) - so a label only matches a full word, not an arbitrary substring, and every distinct matching option across the whole document is suggested, not just the first. |
|
DocumentTermsSuggestion |
Matches option labels against the document's own indexed search terms (LuceneUtils#getTerms) instead of its raw extracted text - so matching benefits from whatever the search engine's analyzer already did (tokenizing, lowercasing, and so on), rather than a naive word split. |
All three treat the select element's own options as a dictionary of possible values to look for in the document, and return the matching ones' values so they can be pre-selected as tags. Not wired to a metadata field automatically - it is referenced by canonical class name in the suggestion attribute of a select element in the metadata XML definition. See Creating your own Suggestion Analyzer for the interface and a full example.
