Value:0.0 to 1.0

Defines the minimum confidence a suggestion needs to be part of the response. The number of suggestions is either limited by the ksuggest.threshold or the ksuggest.max parameter, whichever limit is reached first.



Defines the maximum number of suggestions that should be returned. The number of suggestions is either limited by the ksuggest.threshold or the ksuggest.max paramter, whichever limit is reached first.



With ksuggest.inspect you can specify how many images are involved to calculate the suggestions from. Usually, there is no need to set this param explicitly. The default value is suitable for most cases.

Example usage of the ksuggest.inspect parameter:




With param ksuggest.inspect.minterms you can set a minimum threshold to only include images in the calculation which have at least n terms (analyzed keywords). Documents with fewer terms will be ignored and removed from the result set.

This is useful if there are images in the collection with only one or two keywords assigned. Since Keyword Suggester uses a statistical approach few keywords lead to poor suggestions, since there is no significant variety or frequency of keywords.

Images having fewer terms as demanded by ksuggest.inspect.minterms will be ignored and removed from the response. Doing so, reduces the actual number of images that are involved in the suggestion calculation and the returned images. Depending on the parameter value and your collection no suggestions could be made at all, because all images have too few keywords. See ksuggest.inspect.overhead how to avoid such situations.

Only involve images in the suggestion process with at least 5 terms:




When using ksuggest.inspect.minterms the specified number of inspected images (controlled via ksuggest.inspect ) may be reduced, because images can be ignored.

To compensate ignored images, set an overhead value ksuggest.inspect.overhead, to request more images than actually needed. This acts as a buffer to take images from when other images are ignored. Those overhead images won’t be processed as long as no images are ignored.

ksuggest.inspect + ksuggest.inspect.overhead = number of internally retrieved documents. For instance, 10 documents out of 50 (ksuggest.inspect=50) have only one keyword. They will be ignored, reducing the amount of documents that are part of the statistical analysis. In this case the 10 missing documents are retrieved from the overhead.

50 images can be ignored before the overhead is used up:



Value:comma separated list of keywords

With ksuggest.boost you can specify one or more keywords describing your input image to guide the Keyword Suggester to the correct context. All given keywords will be returned as suggestions in the response.

It is usually filled with manually added or selected suggested keywords during the Keyword Suggester refinement process (see Refinement process).

The given keywords are de-duplicated, converted to lowercase and leading or trailing white spaces are removed.

If ksuggest.boost is given, pixolution flow will try to determine the field to match its value. Two possible options are tested in the following order:

  1. The field specified via df in the request parameters or the solrconfig.xml.
  2. The field specified via ksuggest.field in the request params.

If the field to use can’t be determined, an error will be thrown. The field determined must be indexed, otherwise an error will be thrown.

Example usage of the ksuggest.boost parameter during the refinement process:

?q=*:*&ksuggest.field=keywords&ksuggest=id:123&ksuggest.boost=new york,skyline