public class QueryLargePassages
extends com.ibm.cloud.sdk.core.service.model.GenericModel
Modifier and Type | Class and Description |
---|---|
static class |
QueryLargePassages.Builder
Builder.
|
Modifier and Type | Method and Description |
---|---|
Long |
characters()
Gets the characters.
|
Long |
count()
Gets the count.
|
Boolean |
enabled()
Gets the enabled.
|
List<String> |
fields()
Gets the fields.
|
Boolean |
findAnswers()
Gets the findAnswers.
|
Long |
maxAnswersPerPassage()
Gets the maxAnswersPerPassage.
|
Long |
maxPerDocument()
Gets the maxPerDocument.
|
QueryLargePassages.Builder |
newBuilder()
New builder.
|
Boolean |
perDocument()
Gets the perDocument.
|
public QueryLargePassages.Builder newBuilder()
public Boolean enabled()
A passages query that returns the most relevant passages from the results.
public Boolean perDocument()
If `true`, ranks the documents by document quality, and then returns the highest-ranked passages per document in a `document_passages` field for each document entry in the results list of the response.
If `false`, ranks the passages from all of the documents by passage quality regardless of the document quality and returns them in a separate `passages` field in the response.
public Long maxPerDocument()
Maximum number of passages to return per document in the result. Ignored if **passages.per_document** is `false`.
public List<String> fields()
A list of fields to extract passages from. By default, passages are extracted from the `text` and `title` fields only. If you add this parameter and specify an empty list (`[]`) as its value, then the service searches all root-level fields for suitable passages.
public Long count()
The maximum number of passages to return. Ignored if **passages.per_document** is `true`.
public Long characters()
The approximate number of characters that any one passage will have.
public Boolean findAnswers()
When true, `answer` objects are returned as part of each passage in the query results. The primary difference between an `answer` and a `passage` is that the length of a passage is defined by the query, where the length of an `answer` is calculated by Discovery based on how much text is needed to answer the question.
This parameter is ignored if passages are not enabled for the query, or no **natural_language_query** is specified.
If the **find_answers** parameter is set to `true` and **per_document** parameter is also set to `true`, then the document search results and the passage search results within each document are reordered using the answer confidences. The goal of this reordering is to place the best answer as the first answer of the first passage of the first document. Similarly, if the **find_answers** parameter is set to `true` and **per_document** parameter is set to `false`, then the passage search results are reordered in decreasing order of the highest confidence answer for each document and passage.
The **find_answers** parameter is available only on managed instances of Discovery.
public Long maxAnswersPerPassage()
The number of `answer` objects to return per passage if the **find_answers** parmeter is specified as `true`.
Copyright © 2023 IBM Cloud. All rights reserved.